Authors: Nazia Nafis, Inaki Esnaola, Alvaro Martinez-Perez, Maria-Cruz Villa-Uriol, Venet Osmani
Abstract: Generating synthetic tabular data can be challenging, however evaluation of their quality is just as challenging, if not more. This systematic review sheds light on the critical importance of rigorous evaluation of synthetic health data to ensure reliability, relevance, and their appropriate use. Based on screening of 1766 papers and a detailed review of 101 papers we identified key challenges, including lack of consensus on evaluation methods, improper use of evaluation metrics, limited input from domain experts, inadequate reporting of dataset characteristics, and limited reproducibility of results. In response, we provide several guidelines on the generation and evaluation of synthetic data, to allow the community to unlock and fully harness the transformative potential of synthetic data and accelerate innovation.
Authors: Ching-Yi Lin, Sahil Shah
Abstract: Pre-trained vision transformers have achieved remarkable performance across various visual tasks but suffer from expensive computational and memory costs. While model quantization reduces memory usage by lowering precision, these models still incur significant computational overhead due to the dequantization before matrix operations. In this work, we analyze the computation graph and propose an integerization process based on operation reordering. Specifically, the process delays dequantization until after matrix operations. This enables integerized matrix multiplication and linear module by directly processing the quantized input. To validate our approach, we synthesize the self-attention module of ViT on a systolic array-based hardware. Experimental results show that our low-bit inference reduces per-PE power consumption for linear layer and matrix multiplication, bridging the gap between quantized models and efficient inference.
Authors: Jialei Song, Xingquan Zuo, Feiyang Wang, Hai Huang, Tianle Zhang
Abstract: Deep neural networks (DNNs) are highly susceptible to adversarial samples, raising concerns about their reliability in safety-critical tasks. Currently, methods of evaluating adversarial robustness are primarily categorized into attack-based and certified robustness evaluation approaches. The former not only relies on specific attack algorithms but also is highly time-consuming, while the latter due to its analytical nature, is typically difficult to implement for large and complex models. A few studies evaluate model robustness based on the model's decision boundary, but they suffer from low evaluation accuracy. To address the aforementioned issues, we propose a novel adversarial robustness evaluation metric, Robustness Difference Index (RDI), which is based on sample clustering features. RDI draws inspiration from clustering evaluation by analyzing the intra-class and inter-class distances of feature vectors separated by the decision boundary to quantify model robustness. It is attack-independent and has high computational efficiency. Experiments show that, RDI demonstrates a stronger correlation with the gold-standard adversarial robustness metric of attack success rate (ASR). The average computation time of RDI is only 1/30 of the evaluation method based on the PGD attack. Our open-source code is available at: https://anonymous.4open.science/r/RDI-B1DA.
Authors: Ayoub Jadouli, Chaker El Amrani
Abstract: Deep learning models, especially large Transformers, carry substantial "memory" in their intermediate layers -- an \emph{internal world} that encodes a wealth of relational and contextual knowledge. This work harnesses that internal world for wildfire occurrence prediction by introducing a modular architecture built upon Gemma 3, a state-of-the-art multimodal model. Rather than relying on Gemma 3's original embedding and positional encoding stacks, we develop a custom feed-forward module that transforms tabular wildfire features into the hidden dimension required by Gemma 3's mid-layer Transformer blocks. We freeze these Gemma 3 sub-layers -- thus preserving their pretrained representation power -- while training only the smaller input and output networks. This approach minimizes the number of trainable parameters and reduces the risk of overfitting on limited wildfire data, yet retains the benefits of Gemma 3's broad knowledge. Evaluations on a Moroccan wildfire dataset demonstrate improved predictive accuracy and robustness compared to standard feed-forward and convolutional baselines. Ablation studies confirm that the frozen Transformer layers consistently contribute to better representations, underscoring the feasibility of reusing large-model mid-layers as a learned internal world. Our findings suggest that strategic modular reuse of pretrained Transformers can enable more data-efficient and interpretable solutions for critical environmental applications such as wildfire risk management.
Authors: Aviv Bick, Eric Xing, Albert Gu
Abstract: SSMs offer efficient processing of long sequences with fixed state sizes, but struggle with algorithmic tasks like retrieving past context. In this work, we examine how such in-context retrieval operates within Transformer- and SSM-based language models. We find that both architectures develop the same fundamental Gather-and-Aggregate (G&A) mechanism. A Gather Head first identifies and extracts relevant information from the context, which an Aggregate Head then integrates into a final representation. Across both model types, G&A concentrates in just a few heads, making them critical bottlenecks even for benchmarks that require a basic form of retrieval. For example, disabling a single Gather or Aggregate Head of a pruned Llama-3.1-8B degrades its ability to retrieve the correct answer letter in MMLU, reducing accuracy from 66% to 25%. This finding suggests that in-context retrieval can obscure the limited knowledge demands of certain tasks. Despite strong MMLU performance with retrieval intact, the pruned model fails on other knowledge tests. Similar G&A dependencies exist in GSM8K, BBH, and dialogue tasks. Given the significance of G&A in performance, we show that retrieval challenges in SSMs manifest in how they implement G&A, leading to smoother attention patterns rather than the sharp token transitions that effective G&A relies on. Thus, while a gap exists between Transformers and SSMs in implementing in-context retrieval, it is confined to a few heads, not the entire model. This insight suggests a unified explanation for performance differences between Transformers and SSMs while also highlighting ways to combine their strengths. For example, in pretrained hybrid models, attention components naturally take on the role of Aggregate Heads. Similarly, in a pretrained pure SSM, replacing a single G&A head with an attention-based variant significantly improves retrieval.
Authors: Orhun Vural, Bunyamin Ozaydin, Khalid Y. Aram, James Booth, Brittany F. Lindsey, Abdulaziz Ahmed
Abstract: Background: Emergency department (ED) overcrowding remains a major challenge, causing delays in care and increased operational strain. Hospital management often reacts to congestion after it occurs. Machine learning predictive modeling offers a proactive approach by forecasting patient flow metrics, such as waiting count, to improve resource planning and hospital efficiency. Objective: This study develops machine learning models to predict ED waiting room occupancy at two time scales. The hourly model forecasts the waiting count six hours ahead (e.g., a 1 PM prediction for 7 PM), while the daily model estimates the average waiting count for the next 24 hours (e.g., a 5 PM prediction for the following day's average). These tools support staffing decisions and enable earlier interventions to reduce overcrowding. Methods: Data from a partner hospital's ED in the southeastern United States were used, integrating internal metrics and external features. Eleven machine learning algorithms, including traditional and deep learning models, were trained and evaluated. Feature combinations were optimized, and performance was assessed across varying patient volumes and hours. Results: TSiTPlus achieved the best hourly prediction (MAE: 4.19, MSE: 29.32). The mean hourly waiting count was 18.11, with a standard deviation of 9.77. Accuracy varied by hour, with MAEs ranging from 2.45 (11 PM) to 5.45 (8 PM). Extreme case analysis at one, two, and three standard deviations above the mean showed MAEs of 6.16, 10.16, and 15.59, respectively. For daily predictions, XCMPlus performed best (MAE: 2.00, MSE: 6.64), with a daily mean of 18.11 and standard deviation of 4.51. Conclusions: These models accurately forecast ED waiting room occupancy and support proactive resource allocation. Their implementation has the potential to improve patient flow and reduce overcrowding in emergency care settings.
Authors: Feng Chen, Yefei He, Lequan Lin, Jing Liu, Bohan Zhuang, Qi Wu
Abstract: Sparse attention mechanisms aim to reduce computational overhead by selectively processing a subset of salient tokens while preserving model performance. Despite the effectiveness of such designs, how to actively encourage token sparsity of well-posed MLLMs remains under-explored, which fundamentally limits the achievable acceleration effect during inference. In this paper, we propose a simple RL-based post-training method named \textbf{ZipR1} that treats the token reduction ratio as the efficiency reward and answer accuracy as the performance reward. In this way, our method can jointly alleviate the computation and memory bottlenecks via directly optimizing the inference-consistent efficiency-performance tradeoff. Experimental results demonstrate that ZipR1 can reduce the token ratio of Qwen2/2.5-VL from 80\% to 25\% with a minimal accuracy reduction on 13 image and video benchmarks.
Authors: Shi Jie Yu, Sehyun Choi
Abstract: Checkpoint merging is a technique for combining multiple model snapshots into a single superior model, potentially reducing training time for large language models. This paper explores checkpoint merging in the context of parameter-efficient fine-tuning (PEFT), where only small adapter modules (e.g. LoRA) are trained. We propose Metrics-Weighted Averaging (MWA), a simple yet effective method to merge model checkpoints by weighting their parameters according to performance metrics. In particular, we investigate weighting by training loss and by training steps, under the intuition that lower-loss or later-step checkpoints are more valuable. We introduce a formula with a penalty factor to adjust weight distribution, requiring only one hyperparameter regardless of the number of checkpoints. Experiments on three fine-tuning tasks (mathematical reasoning, preference alignment, and general instruction tuning) show that MWA consistently produces merged models that outperform the naive uniform average of checkpoints. Notably, loss-weighted merging often yields the best results, delivering up to 5% higher task accuracy than the baseline uniform merge and even surpassing the final individual checkpoint's performance. These findings validate checkpoint merging for PEFT and demonstrate that a metric-driven weighting heuristic can efficiently boost model performance with minimal computational overhead.
Authors: Zihao An, Huajun Bai, Ziqiong Liu, Dong Li, Emad Barsoum
Abstract: The autoregressive nature of large language models (LLMs) limits inference speed. Each forward pass generates only a single token and is often bottlenecked by memory bandwidth. Speculative decoding alleviates this issue using a draft-then-verify approach to accelerate token generation. However, the overhead introduced during the draft phase and the training cost of the draft model limit the efficiency and adaptability of speculative decoding. In this work, we introduce PARallel Draft (PARD), a novel speculative decoding method that enables low-cost adaptation of autoregressive draft models into parallel draft models. PARD enhances inference efficiency by predicting multiple future tokens in a single forward pass of the draft phase, and incorporates a conditional drop token method to accelerate training. Its target-independence property allows a single draft model to be applied to an entire family of different models, minimizing the adaptation cost. Our proposed conditional drop token method can improves draft model training efficiency by 3x. On our optimized inference framework, PARD accelerates LLaMA3.1-8B inference by 4.08x, achieving 311.5 tokens per second.
Authors: Tianbing Xu
Abstract: Recently, foundation models such as OpenAI's O1 and O3, along with DeepSeek's R1, have demonstrated strong reasoning capacities and problem-solving skills acquired through large-scale reinforcement learning (RL), with wide applications in mathematics, coding, science, intelligent agents, and virtual assistants. In this work, we introduce an off-policy reinforcement learning algorithm, EM Policy Gradient, aimed at enhancing LLM reasoning by optimizing expected return over reasoning trajectories. We frame the reasoning task as an Expectation-Maximization (EM) optimization problem, alternating between sampling diverse rationale trajectories and performing reward-guided fine-tuning. Unlike PPO and GRPO, which rely on complex importance weights and heuristic clipping, our method provides a simpler, more principled off-policy policy gradient approach, eliminating these complexities while maintaining strong performance. We evaluate the effectiveness of EM Policy Gradient on the GSM8K and MATH (HARD) datasets, where it achieves performance comparable to or slightly surpassing the state-of-the-art GRPO, while offering additional advantages in scalability, simplicity, and reasoning conciseness. Moreover, models fine-tuned with our method exhibit cognitive behaviors, such as sub-problem decomposition, self-verification, and backtracking, highlighting its potential to enhance both the interpretability and robustness of LLM reasoning.
Authors: YongHui Xia, Lan Wang, Hao Wu
Abstract: Dynamic quality of service (QoS) data exhibit rich temporal patterns in user-service interactions, which are crucial for a comprehensive understanding of user behavior and service conditions in Web service. As the number of users and services increases, there is a large amount of unobserved QoS data, which significantly affects users'choice of services. To predict unobserved QoS data, we propose a Non-negative Snowflake Factorization of tensors model. This method designs a snowflake core tensor to enhance the model's learning capability. Additionally, it employs a single latent factor-based, nonnegative multiplication update on tensor (SLF-NMUT) for parameter learning. Empirical results demonstrate that the proposed model more accurately learns dynamic user-service interaction patterns, thereby yielding improved predictions for missing QoS data.
Authors: Guillaume Lauga (OCKHAM), Ma\"el Chaumette (OCKHAM), Edgar Desainte-Mar\'eville (OCKHAM), \'Etienne Lasalle (OCKHAM), Arthur Lebeurrier (OCKHAM)
Abstract: In this article, we investigate the potential of multilevel approaches to accelerate the training of transformer architectures. Using an ordinary differential equation (ODE) interpretation of these architectures, we propose an appropriate way of varying the discretization of these ODE Transformers in order to accelerate the training. We validate our approach experimentally by a comparison with the standard training procedure.
Authors: Giovanni Catalani, Michael Bauerheim, Fr\'ed\'eric Tost, Xavier Bertrand, Joseph Morlier
Abstract: Recent advances in Neural Fields have enabled powerful, discretization-invariant methods for learning neural operators that approximate solutions of Partial Differential Equations (PDEs) on general geometries. Building on these developments, we introduce enf2enf, an encoder--decoder methodology for predicting steady-state Partial Differential Equations with non-parameterized geometric variability, based on recently proposed Equivariant Neural Field architectures. In enf2enf, input geometries are encoded into latent point cloud embeddings that inherently preserve geometric grounding and capture local phenomena. The resulting representations are then combined with global parameters and directly decoded into continuous output fields, thus efficiently modeling the coupling between geometry and physics. By leveraging the inductive biases of locality and translation invariance, our approach is able to capture fine-scale physical features as well as complex shape variations, thereby enhancing generalization and physical compliance. Extensive experiments on a high-fidelity aerodynamic dataset, a hyper-elastic material benchmark, and multi-element airfoil geometries, demonstrate that the proposed model achieves superior or competitive performance compared to state-of-the-art graph based, operator learning, and neural field methods. Notably, our method supports real time inference and zero-shot super-resolution, enabling efficient training on low-resolution meshes while maintaining high accuracy on full-scale discretizations.
Authors: Akram Shojaei, Mehdi Delrobaei
Abstract: Chronic obstructive pulmonary disease (COPD) represents a significant global health burden, where precise severity assessment is particularly critical for effective clinical management in intensive care unit (ICU) settings. This study introduces an innovative machine learning framework for COPD severity classification utilizing the MIMIC-III critical care database, thereby expanding the applications of artificial intelligence in critical care medicine. Our research developed a robust classification model incorporating key ICU parameters such as blood gas measurements and vital signs, while implementing semi-supervised learning techniques to effectively utilize unlabeled data and enhance model performance. The random forest classifier emerged as particularly effective, demonstrating exceptional discriminative capability with 92.51% accuracy and 0.98 ROC AUC in differentiating between mild-to-moderate and severe COPD cases. This machine learning approach provides clinicians with a practical, accurate, and efficient tool for rapid COPD severity evaluation in ICU environments, with significant potential to improve both clinical decision-making processes and patient outcomes. Future research directions should prioritize external validation across diverse patient populations and integration with clinical decision support systems to optimize COPD management in critical care settings.
Authors: Tongrui Su, Qingbin Li, Shengyu Zhu, Wei Chen, Xueqi Cheng
Abstract: We study the problem of transfer-based black-box attack, where adversarial samples generated using a single surrogate model are directly applied to target models. Compared with untargeted attacks, existing methods still have lower Attack Success Rates (ASRs) in the targeted setting, i.e., the obtained adversarial examples often overfit the surrogate model but fail to mislead other models. In this paper, we hypothesize that the pixels or features in these adversarial examples collaborate in a highly dependent manner to maximize the success of an adversarial attack on the surrogate model, which we refer to as perturbation co-adaptation. Then, we propose to Mitigate perturbation Co-adaptation by DropConnect (MCD) to enhance transferability, by creating diverse variants of surrogate model at each optimization iteration. We conduct extensive experiments across various CNN- and Transformer-based models to demonstrate the effectiveness of MCD. In the challenging scenario of transferring from a CNN-based model to Transformer-based models, MCD achieves 13% higher average ASRs compared with state-of-the-art baselines. MCD boosts the performance of self-ensemble methods by bringing in more diversification across the variants while reserving sufficient semantic information for each variant. In addition, MCD attains the highest performance gain when scaling the compute of crafting adversarial examples.
Authors: Uzma, Fabien Cholet, Domenic Quinn, Cindy Smith, Siming You, William Sloan
Abstract: Environmental biotechnologies, such as drinking water biofilters, rely on complex interactions between microbial communities and their surrounding physical-chemical environments. Predicting the performance of these systems is challenging due to high-dimensional, sparse datasets that lack diversity and fail to fully capture system behaviour. Accurate predictive models require innovative, science-guided approaches. In this study, we present the first application of Buckingham Pi theory to modelling biofilter performance. This dimensionality reduction technique identifies meaningful, dimensionless variables that enhance predictive accuracy and improve model interpretability. Using these variables, we developed the Environmental Buckingham Pi Neural Network (EnviroPiNet), a physics-guided model benchmarked against traditional data-driven methods, including Principal Component Analysis (PCA) and autoencoder neural networks. Our findings demonstrate that the EnviroPiNet model achieves an R^2 value of 0.9236 on the testing dataset, significantly outperforming PCA and autoencoder methods. The Buckingham Pi variables also provide insights into the physical and chemical relationships governing biofilter behaviour, with implications for system design and optimization. This study highlights the potential of combining physical principles with AI approaches to model complex environmental systems characterized by sparse, high-dimensional datasets.
Authors: Subhadip Bandyopadhyay, Joy Bose, Sujoy Roy Chowdhury
Abstract: Data Drift is the phenomenon where the generating model behind the data changes over time. Due to data drift, any model built on the past training data becomes less relevant and inaccurate over time. Thus, detecting and controlling for data drift is critical in machine learning models. Hierarchical Temporal Memory (HTM) is a machine learning model developed by Jeff Hawkins, inspired by how the human brain processes information. It is a biologically inspired model of memory that is similar in structure to the neocortex, and whose performance is claimed to be comparable to state of the art models in detecting anomalies in time series data. Another unique benefit of HTMs is its independence from training and testing cycle; all the learning takes place online with streaming data and no separate training and testing cycle is required. In sequential learning paradigm, Sequential Probability Ratio Test (SPRT) offers some unique benefit for online learning and inference. This paper proposes a novel hybrid framework combining HTM and SPRT for real-time data drift detection and anomaly identification. Unlike existing data drift methods, our approach eliminates frequent retraining and ensures low false positive rates. HTMs currently work with one dimensional or univariate data. In a second study, we also propose an application of HTM in multidimensional supervised scenario for anomaly detection by combining the outputs of multiple HTM columns, one for each dimension of the data, through a neural network. Experimental evaluations demonstrate that the proposed method outperforms conventional drift detection techniques like the Kolmogorov-Smirnov (KS) test, Wasserstein distance, and Population Stability Index (PSI) in terms of accuracy, adaptability, and computational efficiency. Our experiments also provide insights into optimizing hyperparameters for real-time deployment in domains such as Telecom.
Authors: Bruce Collins
Abstract: The Xeno-Canto bird audio repository is an invaluable resource for those interested in vocalizations and other sounds made by birds around the world. This is particularly the case for machine learning researchers attempting to improve on the bird species recognition accuracy of classification models. However, the task of extracting labeled datasets from the recordings found in this crowd-sourced repository faces several challenges. One challenge of particular significance to machine learning practitioners is that one bird species label is applied to each audio recording, but frequently other sounds are also captured including other bird species, other animal sounds, anthropogenic and other ambient sounds. These non-target bird species sounds can result in dataset labeling discrepancies referred to as label noise. In this work we present a cleaning process consisting of audio preprocessing followed by dimensionality reduction and unsupervised outlier detection (UOD) to reduce the label noise in a dataset derived from Xeno-Canto recordings. We investigate three neural network dimensionality reduction techniques: two flavors of convolutional autoencoders and variational deep embedding (VaDE (Jiang, 2017)). While both methods show some degree of effectiveness at detecting outliers for most bird species datasets, we found significant variation in the performance of the methods from one species to the next. We believe that the results of this investigation demonstrate that the application of our cleaning process can meaningfully reduce the label noise of bird species datasets derived from Xeno-Canto audio repository but results vary across species.
Authors: Daehyeon Han, Morteza Karimzadeh
Abstract: The Ice, Cloud, and Elevation Satellite-2 (ICESat-2) provides high-resolution measurements of sea ice height. Recent studies have developed machine learning methods on ICESat-2 data, primarily focusing on surface type classification. However, the heavy reliance on manually collected labels requires significant time and effort for supervised learning, as it involves cross-referencing track measurements with overlapping background optical imagery. Additionally, the coincidence of ICESat-2 tracks with background images is relatively rare due to the different overpass patterns and atmospheric conditions. To address these limitations, this study explores the potential of unsupervised autoencoder on unlabeled data to derive latent embeddings. We develop autoencoder models based on Long Short-Term Memory (LSTM) and Convolutional Neural Networks (CNN) to reconstruct topographic sequences from ICESat-2 and derive embeddings. We then apply Uniform Manifold Approximation and Projection (UMAP) to reduce dimensions and visualize the embeddings. Our results show that embeddings from autoencoders preserve the overall structure but generate relatively more compact clusters compared to the original ICESat-2 data, indicating the potential of embeddings to lessen the number of required labels samples.
Authors: Mustafa Musab, Joseph K. Chege, Arie Yeredor, Martin Haardt
Abstract: Reliable density estimation is fundamental for numerous applications in statistics and machine learning. In many practical scenarios, data are best modeled as mixtures of component densities that capture complex and multimodal patterns. However, conventional density estimators based on uniform histograms often fail to capture local variations, especially when the underlying distribution is highly nonuniform. Furthermore, the inherent discontinuity of histograms poses challenges for tasks requiring smooth derivatives, such as gradient-based optimization, clustering, and nonparametric discriminant analysis. In this work, we present a novel non-parametric approach for multivariate probability density function (PDF) estimation that utilizes minimum description length (MDL)-based binning with quantile cuts. Our approach builds upon tensor factorization techniques, leveraging the canonical polyadic decomposition (CPD) of a joint probability tensor. We demonstrate the effectiveness of our method on synthetic data and a challenging real dry bean classification dataset.
Authors: Felix Burr, Marcel Hoffmann, Ansgar Scherp
Abstract: Despite the ample availability of graph data, obtaining vertex labels is a tedious and expensive task. Therefore, it is desirable to learn from a few labeled vertices only. Existing few-shot learners assume a class oracle, which provides labeled vertices for a desired class. However, such an oracle is not available in a real-world setting, i.e., when drawing a vertex for labeling it is unknown to which class the vertex belongs. Few-shot learners are often combined with prototypical networks, while classical semi-supervised vertex classification uses discriminative models, e.g., Graph Convolutional Networks (GCN). In this paper, we train our models by iteratively prompting a human annotator with vertices to annotate. We perform three experiments where we continually relax our assumptions. First, we assume a class oracle, i.e., the human annotator is provided with an equal number of vertices to label for each class. We denote this as "Balanced Sampling''. In the subsequent experiment, "Unbalanced Sampling,'' we replace the class oracle with $k$-medoids clustering and draw vertices to label from the clusters. In the last experiment, the "Unknown Number of Classes,'' we no longer assumed we knew the number and distribution of classes. Our results show that prototypical models outperform discriminative models in all experiments when fewer than $20$ samples per class are available. While dropping the assumption of the class oracle for the "Unbalanced Sampling'' experiment reduces the performance of the GCN by $9\%$, the prototypical network loses only $1\%$ on average. For the "Unknown Number of Classes'' experiment, the average performance for both models decreased further by $1\%$. Source code: https://github.com/Ximsa/2023-felix-ma
Authors: Patr\'icia Mu\~noz Ewald
Abstract: We fully characterize a large class of feedforward neural networks in terms of truncation maps. As an application, we show how a ReLU neural network can implement a feature map which separates concentric data.
Authors: G\'er\^ome Andry, Fran\c{c}ois Rozet, Sacha Lewin, Omer Rochman, Victor Mangeleer, Matthias Pirlet, Elise Faulx, Marilaure Gr\'egoire, Gilles Louppe
Abstract: Deep learning has transformed weather forecasting by improving both its accuracy and computational efficiency. However, before any forecast can begin, weather centers must identify the current atmospheric state from vast amounts of observational data. To address this challenging problem, we introduce Appa, a score-based data assimilation model producing global atmospheric trajectories at 0.25-degree resolution and 1-hour intervals. Powered by a 1.5B-parameter spatio-temporal latent diffusion model trained on ERA5 reanalysis data, Appa can be conditioned on any type of observations to infer the posterior distribution of plausible state trajectories, without retraining. Our unified probabilistic framework flexibly tackles multiple inference tasks -- reanalysis, filtering, and forecasting -- using the same model, eliminating the need for task-specific architectures or training procedures. Experiments demonstrate physical consistency on a global scale and good reconstructions from observations, while showing competitive forecasting skills. Our results establish latent score-based data assimilation as a promising foundation for future global atmospheric modeling systems.
Authors: Tung D. Vu, Chung Hoang, Truong-Son Hy
Abstract: The Design2Code problem, which involves converting digital designs into functional source code, is a significant challenge in software development due to its complexity and time-consuming nature. Traditional approaches often struggle with accurately interpreting the intricate visual details and structural relationships inherent in webpage designs, leading to limitations in automation and efficiency. In this paper, we propose a novel method that leverages multimodal graph representation learning to address these challenges. By integrating both visual and structural information from design sketches, our approach enhances the accuracy and efficiency of code generation, particularly in producing semantically correct and structurally sound HTML code. We present a comprehensive evaluation of our method, demonstrating significant improvements in both accuracy and efficiency compared to existing techniques. Extensive evaluation demonstrates significant improvements of multimodal graph learning over existing techniques, highlighting the potential of our method to revolutionize design-to-code automation. Code available at https://github.com/HySonLab/Design2Code
Authors: Tanvir Islam
Abstract: We propose TLoRA, a novel tri-matrix low-rank adaptation method that decomposes weight updates into three matrices: two fixed random matrices and one trainable matrix, combined with a learnable, layer-wise scaling factor. This tri-matrix design enables TLoRA to achieve highly efficient parameter adaptation while introducing minimal additional computational overhead. Through extensive experiments on the GLUE benchmark, we demonstrate that TLoRA achieves comparable performance to existing low-rank methods such as LoRA and Adapter-based techniques, while requiring significantly fewer trainable parameters. Analyzing the adaptation dynamics, we observe that TLoRA exhibits Gaussian-like weight distributions, stable parameter norms, and scaling factor variability across layers, further highlighting its expressive power and adaptability. Additionally, we show that TLoRA closely resembles LoRA in its eigenvalue distributions, parameter norms, and cosine similarity of updates, underscoring its ability to effectively approximate LoRA's adaptation behavior. Our results establish TLoRA as a highly efficient and effective fine-tuning method for LLMs, offering a significant step forward in resource-efficient model adaptation.
Authors: Zaiwei Chen
Abstract: This work presents the first finite-time analysis for the last-iterate convergence of average-reward Q-learning with an asynchronous implementation. A key feature of the algorithm we study is the use of adaptive stepsizes, which serve as local clocks for each state-action pair. We show that the iterates generated by this Q-learning algorithm converge at a rate of $O(1/k)$ (in the mean-square sense) to the optimal relative Q-function in the span seminorm. Moreover, by adding a centering step to the algorithm, we further establish pointwise mean-square convergence to a centered optimal relative Q-function, also at a rate of $O(1/k)$. To prove these results, we show that adaptive stepsizes are necessary, as without them, the algorithm fails to converge to the correct target. In addition, adaptive stepsizes can be interpreted as a form of implicit importance sampling that counteracts the effects of asynchronous updates. Technically, the use of adaptive stepsizes makes each Q-learning update depend on the entire sample history, introducing strong correlations and making the algorithm a non-Markovian stochastic approximation (SA) scheme. Our approach to overcoming this challenge involves (1) a time-inhomogeneous Markovian reformulation of non-Markovian SA, and (2) a combination of almost-sure time-varying bounds, conditioning arguments, and Markov chain concentration inequalities to break the strong correlations between the adaptive stepsizes and the iterates. The tools developed in this work are likely to be broadly applicable to the analysis of general SA algorithms with adaptive stepsizes.
Authors: Ling Wang, Minglian Han
Abstract: Link prediction is a fundamental task in dynamic graph learning (DGL), inherently shaped by the topology of the DG. Recent advancements in dynamic graph neural networks (DGNN), primarily by modeling the relationships among nodes via a message passing scheme, have significantly improved link prediction performance. However, DGNNs heavily rely on the pairwise node interactions, which neglect the common neighbor interaction in DGL. To address this limitation, we propose a High-order Graph Neural Networks with Common Neighbor Awareness (HGNN-CNA) for link prediction with two-fold ideas: a) estimating correlation score by considering multi-hop common neighbors for capturing the complex interaction between nodes; b) fusing the correlation into the message-passing process to consider common neighbor interaction directly in DGL. Experimental results on three real DGs demonstrate that the proposed HGNN-CNA acquires a significant accuracy gain over several state-of-the-art models on the link prediction task.
Authors: Wenjun Cao
Abstract: Reinforcement learning (RL) suffers from severe sample inefficiency, especially during early training, requiring extensive environmental interactions to perform competently. Existing methods tend to solve this by incorporating prior knowledge, but introduce significant architectural and implementation complexity. We propose Dynamic Action Interpolation (DAI), a universal yet straightforward framework that interpolates expert and RL actions via a time-varying weight $\alpha(t)$, integrating into any Actor-Critic algorithm with just a few lines of code and without auxiliary networks or additional losses. Our theoretical analysis shows that DAI reshapes state visitation distributions to accelerate value function learning while preserving convergence guarantees. Empirical evaluations across MuJoCo continuous control tasks demonstrate that DAI improves early-stage performance by over 160\% on average and final performance by more than 50\%, with the Humanoid task showing a 4$\times$ improvement early on and a 2$\times$ gain at convergence. These results challenge the assumption that complex architectural modifications are necessary for sample-efficient reinforcement learning.
Authors: Markus Haug, Gissel Velarde
Abstract: This work empirically evaluates machine learning models on two imbalanced public datasets (KDDCUP99 and Credit Card Fraud 2013). The method includes data preparation, model training, and evaluation, using an 80/20 (train/test) split. Models tested include eXtreme Gradient Boosting (XGB), Multi Layer Perceptron (MLP), Generative Adversarial Network (GAN), Variational Autoencoder (VAE), and Multiple-Objective Generative Adversarial Active Learning (MO-GAAL), with XGB and MLP further combined with Random-Over-Sampling (ROS) and Self-Paced-Ensemble (SPE). Evaluation involves 5-fold cross-validation and imputation techniques (mean, median, and IterativeImputer) with 10, 20, 30, and 50 % missing data. Findings show XGB and MLP outperform generative models. IterativeImputer results are comparable to mean and median, but not recommended for large datasets due to increased complexity and execution time. The code used is publicly available on GitHub (github.com/markushaug/acr-25).
Authors: Santosh Rajagopalan, Jonathan Vronsky, Songbai Yan, S. Alireza Golestaneh, Shubhra Chandra, Min Zhou
Abstract: We present ALF (Advertiser Large Foundation model), a multi-modal transformer architecture for understanding advertiser behavior and intent across text, image, video and structured data modalities. Through contrastive learning and multi-task optimization, ALF creates unified advertiser representations that capture both content and behavioral patterns. Our model achieves state-of-the-art performance on critical tasks including fraud detection, policy violation identification, and advertiser similarity matching. In production deployment, ALF reduces false positives by 90% while maintaining 99.8% precision on abuse detection tasks. The architecture's effectiveness stems from its novel combination of multi-modal transformations, inter-sample attention mechanism, spectrally normalized projections, and calibrated probabilistic outputs.
Authors: Xufei Wang, Fei Ge, Jinchen Zhu, Mingjian Zhang, Qi Wu, Jifeng Ren Shizhuang Weng
Abstract: Methods based on implicit neural representation have demonstrated remarkable capabilities in arbitrary-scale super-resolution (ASSR) tasks, but they neglect the potential value of the frequency domain, leading to sub-optimal performance. We proposes a novel network called Frequency-Integrated Transformer (FIT) to incorporate and utilize frequency information to enhance ASSR performance. FIT employs Frequency Incorporation Module (FIM) to introduce frequency information in a lossless manner and Frequency Utilization Self-Attention module (FUSAM) to efficiently leverage frequency information by exploiting spatial-frequency interrelationship and global nature of frequency. FIM enriches detail characterization by incorporating frequency information through a combination of Fast Fourier Transform (FFT) with real-imaginary mapping. In FUSAM, Interaction Implicit Self-Attention (IISA) achieves cross-domain information synergy by interacting spatial and frequency information in subspace, while Frequency Correlation Self-attention (FCSA) captures the global context by computing correlation in frequency. Experimental results demonstrate FIT yields superior performance compared to existing methods across multiple benchmark datasets. Visual feature map proves the superiority of FIM in enriching detail characterization. Frequency error map validates IISA productively improve the frequency fidelity. Local attribution map validates FCSA effectively captures global context.
Authors: Hassan Wasswa, Aziida Nanyonga, Timothy Lynar
Abstract: AI models have garnered significant research attention towards predictive task automation. However, a stationary training environment is an underlying assumption for most models and such models simply do not work on non-stationary data since a stationary relationship is learned. The existing solutions propose making data stationary prior to model training and evaluation. This leads to loss of trend and seasonal patterns which are vital components for learning temporal dependencies of the system under study. This research aims to address this limitation by proposing a method for enforcing stationary behaviour within the latent space while preserving trend and seasonal information. The method deploys techniques including Differencing, Time-series decomposition, and Latent Space Arithmetic (LSA), to learn information vital for efficient approximation of trend and seasonal information which is then stored as embeddings within the latent space of a Variational Autoencoder (VAE). The approach's ability to preserve trend and seasonal information was evaluated on two time-series non-stationary datasets. For predictive performance evaluation, four deep learning models were trained on the latent vector representations of the datasets after application of the proposed method and all models produced competitive results in comparison with state-of-the-art techniques using RMSE as the performance metric.
Authors: Mehmet Ali Ferah, Tufan Kumbasar
Abstract: System Identification (SysID) is crucial for modeling and understanding dynamical systems using experimental data. While traditional SysID methods emphasize linear models, their inability to fully capture nonlinear dynamics has driven the adoption of Deep Learning (DL) as a more powerful alternative. However, the lack of uncertainty quantification (UQ) in DL-based models poses challenges for reliability and safety, highlighting the necessity of incorporating UQ. This paper introduces a systematic framework for constructing and learning Interval Neural Networks (INNs) to perform UQ in SysID tasks. INNs are derived by transforming the learnable parameters (LPs) of pre-trained neural networks into interval-valued LPs without relying on probabilistic assumptions. By employing interval arithmetic throughout the network, INNs can generate Prediction Intervals (PIs) that capture target coverage effectively. We extend Long Short-Term Memory (LSTM) and Neural Ordinary Differential Equations (Neural ODEs) into Interval LSTM (ILSTM) and Interval NODE (INODE) architectures, providing the mathematical foundations for their application in SysID. To train INNs, we propose a DL framework that integrates a UQ loss function and parameterization tricks to handle constraints arising from interval LPs. We introduce novel concept "elasticity" for underlying uncertainty causes and validate ILSTM and INODE in SysID experiments, demonstrating their effectiveness.
Authors: Omar Naifar
Abstract: This paper introduces Tempered Fractional Gradient Descent (TFGD), a novel optimization framework that synergizes fractional calculus with exponential tempering to enhance gradient-based learning. Traditional gradient descent methods often suffer from oscillatory updates and slow convergence in high-dimensional, noisy landscapes. TFGD addresses these limitations by incorporating a tempered memory mechanism, where historical gradients are weighted by fractional coefficients $|w_j| = \binom{\alpha}{j}$ and exponentially decayed via a tempering parameter $\lambda$. Theoretical analysis establishes TFGD's convergence guarantees: in convex settings, it achieves an $\mathcal{O}(1/K)$ rate with alignment coefficient $d_{\alpha,\lambda} = (1 - e^{-\lambda})^{-\alpha}$, while stochastic variants attain $\mathcal{O}(1/k^\alpha)$ error decay. The algorithm maintains $\mathcal{O}(n)$ time complexity equivalent to SGD, with memory overhead scaling as $\mathcal{O}(d/\lambda)$ for parameter dimension $d$. Empirical validation on the Breast Cancer Wisconsin dataset demonstrates TFGD's superiority, achieving 98.25\% test accuracy (vs. 92.11\% for SGD) and 2$\times$ faster convergence. The tempered memory mechanism proves particularly effective in medical classification tasks, where feature correlations benefit from stable gradient averaging. These results position TFGD as a robust alternative to conventional optimizers in both theoretical and applied machine learning.
Authors: Robert Leppich, Michael Stenger, Daniel Grillmeyer, Vanessa Borst, Samuel Kounev
Abstract: We introduce a temporal feature encoding architecture called Time Series Representation Model (TSRM) for multivariate time series forecasting and imputation. The architecture is structured around CNN-based representation layers, each dedicated to an independent representation learning task and designed to capture diverse temporal patterns, followed by an attention-based feature extraction layer and a merge layer, designed to aggregate extracted features. The architecture is fundamentally based on a configuration that is inspired by a Transformer encoder, with self-attention mechanisms at its core. The TSRM architecture outperforms state-of-the-art approaches on most of the seven established benchmark datasets considered in our empirical evaluation for both forecasting and imputation tasks. At the same time, it significantly reduces complexity in the form of learnable parameters. The source code is available at https://github.com/RobertLeppich/TSRM.
Authors: Hangtao Zhang, Zhe Li, Kairui Zhang
Abstract: A primary challenge in ITE estimation is sample selection bias. Traditional approaches utilize treatment regularization techniques such as the Integral Probability Metrics (IPM), re-weighting, and propensity score modeling to mitigate this bias. However, these regularizations may introduce undesirable information loss and limit the performance of the model. Furthermore, treatment effects vary across different external contexts, and the existing methods are insufficient in fully interacting with and utilizing these contextual features. To address these issues, we propose a Context-Aware uplift model based on the Two-Stage training approach (TSCAN), comprising CAN-U and CAN-D sub-models. In the first stage, we train an uplift model, called CAN-U, which includes the treatment regularizations of IPM and propensity score prediction, to generate a complete dataset with counterfactual uplift labels. In the second stage, we train a model named CAN-D, which utilizes an isotonic output layer to directly model uplift effects, thereby eliminating the reliance on the regularization components. CAN-D adaptively corrects the errors estimated by CAN-U through reinforcing the factual samples, while avoiding the negative impacts associated with the aforementioned regularizations. Additionally, we introduce a Context-Aware Attention Layer throughout the two-stage process to manage the interactions between treatment, merchant, and contextual features, thereby modeling the varying treatment effect in different contexts. We conduct extensive experiments on two real-world datasets to validate the effectiveness of TSCAN. Ultimately, the deployment of our model for real-world merchant diagnosis on one of China's largest online food ordering platforms validates its practical utility and impact.
Authors: Ce Ju, Reinmar J. Kobler, Antoine Collas, Motoaki Kawanabe, Cuntai Guan, Bertrand Thirion
Abstract: Neuroimaging provides a critical framework for characterizing brain activity by quantifying connectivity patterns and functional architecture across modalities. While modern machine learning has significantly advanced our understanding of neural processing mechanisms through these datasets, decoding task-specific signatures must contend with inherent neuroimaging constraints, for example, low signal-to-noise ratios in raw electrophysiological recordings, cross-session non-stationarity, and limited sample sizes. This review focuses on machine learning approaches for covariance-based neuroimaging data, where often symmetric positive definite (SPD) matrices under full-rank conditions encode inter-channel relationships. By equipping the space of SPD matrices with Riemannian metrics (e.g., affine-invariant or log-Euclidean), their space forms a Riemannian manifold enabling geometric analysis. We unify methodologies operating on this manifold under the SPD learning framework, which systematically leverages the SPD manifold's geometry to process covariance features, thereby advancing brain imaging analytics.
Authors: Ma{\l}gorzata {\L}az\k{e}cka, Ewa Szczurek
Abstract: Integrating various data modalities brings valuable insights into underlying phenomena. Multimodal factor analysis (FA) uncovers shared axes of variation underlying different simple data modalities, where each sample is represented by a vector of features. However, FA is not suited for structured data modalities, such as text or single cell sequencing data, where multiple data points are measured per each sample and exhibit a clustering structure. To overcome this challenge, we introduce FACTM, a novel, multi-view and multi-structure Bayesian model that combines FA with correlated topic modeling and is optimized using variational inference. Additionally, we introduce a method for rotating latent factors to enhance interpretability with respect to binary features. On text and video benchmarks as well as real-world music and COVID-19 datasets, we demonstrate that FACTM outperforms other methods in identifying clusters in structured data, and integrating them with simple modalities via the inference of shared, interpretable factors.
Authors: Ruifeng Ren, Yong Liu
Abstract: Compression has been a critical lens to understand the success of Transformers. In the past, we have typically taken the target distribution as a criterion to evaluate a model's compression performance. Nevertheless,it often remains challenging to precisely assess how well the model achieves compression and to compare the information content of the learned distribution with that of the target distribution during compression,as the target distribution is typically unknown and entropy computation often incurs exponential cost. In this work, we explore these issues under a controlled experimental setup. We find that Transformers exhibit a unique inductive bias in data compression: beyond approaching the target distribution, they tend to favor learning lower-entropy distributions, with this tendency becoming more pronounced as the model size increases. This preference prevents Transformers from perfectly aligning with the target distribution, instead further compressing its information content. Furthermore, we show that the FFN module plays a critical role in driving this bias. In addition, while models remove informational redundancy from data during compression, they also exhibit redundancy within their parameters, which enables compression and can be characterized through dynamic sparsity. However, the dynamic sparsity patterns in Transformers, particularly in attention and FFN modules, demand further exploration. As for this, we show that larger Transformers show stronger preferences for bypassing attention computations via residual connections and have lower proportion of active neurons. Interestingly, we also find that training instability in larger models strongly correlates with sudden increases in dead neurons. Our work contributes to a deeper understanding of Transformers from the lens of entropy and dynamic sparsity.
Authors: Elad Sofer, Tomer Shaked, Caroline Chaux, Nir Shlezinger
Abstract: Machine learning (ML) models are often sensitive to carefully crafted yet seemingly unnoticeable perturbations. Such adversarial examples are considered to be a property of ML models, often associated with their black-box operation and sensitivity to features learned from data. This work examines the adversarial sensitivity of non-learned decision rules, and particularly of iterative optimizers. Our analysis is inspired by the recent developments in deep unfolding, which cast such optimizers as ML models. We show that non-learned iterative optimizers share the sensitivity to adversarial examples of ML models, and that attacking iterative optimizers effectively alters the optimization objective surface in a manner that modifies the minima sought. We then leverage the ability to cast iteration-limited optimizers as ML models to enhance robustness via adversarial training. For a class of proximal gradient optimizers, we rigorously prove how their learning affects adversarial sensitivity. We numerically back our findings, showing the vulnerability of various optimizers, as well as the robustness induced by unfolding and adversarial training.
Authors: Delun Lai, Yeyubei Zhang, Yunchong Liu, Chaojie Li, Huadong Mo
Abstract: This paper introduces a novel deep learning-based multimodal fusion architecture aimed at enhancing the perception capabilities of autonomous navigation robots in complex environments. By utilizing innovative feature extraction modules, adaptive fusion strategies, and time-series modeling mechanisms, the system effectively integrates RGB images and LiDAR data. The key contributions of this work are as follows: a. the design of a lightweight feature extraction network to enhance feature representation; b. the development of an adaptive weighted cross-modal fusion strategy to improve system robustness; and c. the incorporation of time-series information modeling to boost dynamic scene perception accuracy. Experimental results on the KITTI dataset demonstrate that the proposed approach increases navigation and positioning accuracy by 3.5% and 2.2%, respectively, while maintaining real-time performance. This work provides a novel solution for autonomous robot navigation in complex environments.
Authors: J\'ulia Vicens Figueres, Juliette Vanderhaeghen, Federica Bragone, Kateryna Morozovska, Khemraj Shukla
Abstract: Physics-Informed Neural Networks (PINNs) are a novel computational approach for solving partial differential equations (PDEs) with noisy and sparse initial and boundary data. Although, efficient quantification of epistemic and aleatoric uncertainties in big multi-scale problems remains challenging. We propose \$PINN a novel method of computing global uncertainty in PDEs using a Bayesian framework, by combining local Bayesian Physics-Informed Neural Networks (BPINN) with domain decomposition. The solution continuity across subdomains is obtained by imposing the flux continuity across the interface of neighboring subdomains. To demonstrate the effectiveness of \$PINN, we conduct a series of computational experiments on PDEs in 1D and 2D spatial domains. Although we have adopted conservative PINNs (cPINNs), the method can be seamlessly extended to other domain decomposition techniques. The results infer that the proposed method recovers the global uncertainty by computing the local uncertainty exactly more efficiently as the uncertainty in each subdomain can be computed concurrently. The robustness of \$PINN is verified by adding uncorrelated random noise to the training data up to 15% and testing for different domain sizes.
Authors: Sushant Vijayan
Abstract: We study the Active Simple Hypothesis Testing (ASHT) problem, a simpler variant of the Fixed Budget Best Arm Identification problem. In this work, we provide novel game theoretic formulation of the upper bounds of the ASHT problem. This formulation allows us to leverage tools of differential games and Partial Differential Equations (PDEs) to propose an approximately optimal algorithm that is computationally tractable compared to prior work. However, the optimal algorithm still suffers from a curse of dimensionality and instead we use a novel link to Blackwell Approachability to propose an algorithm that is far more efficient computationally. We show that this new algorithm, although not proven to be optimal, is always better than static algorithms in all instances of ASHT and is numerically observed to attain the optimal exponent in various instances.
Authors: Stanislav Semenov
Abstract: We propose novel smooth approximations to the classical rounding function, suitable for differentiable optimization and machine learning applications. Our constructions are based on two approaches: (1) localized sigmoid window functions centered at each integer, and (2) normalized weighted sums of sigmoid derivatives representing local densities. The first method approximates the step-like behavior of rounding through differences of shifted sigmoids, while the second method achieves smooth interpolation between integers via density-based weighting. Both methods converge pointwise to the classical rounding function as the sharpness parameter k tends to infinity, and allow controlled trade-offs between smoothness and approximation accuracy. We demonstrate that by restricting the summation to a small set of nearest integers, the computational cost remains low without sacrificing precision. These constructions provide fully differentiable alternatives to hard rounding, which are valuable in contexts where gradient-based methods are essential.
Authors: Samantha Petti, Carlos Mart\'i-G\'omez, Justin B. Kinney, Juannan Zhou, David M. McCandlish
Abstract: Mappings from biological sequences (DNA, RNA, protein) to quantitative measures of sequence functionality play an important role in contemporary biology. We are interested in the related tasks of (i) inferring predictive sequence-to-function maps and (ii) decomposing sequence-function maps to elucidate the contributions of individual subsequences. Because each sequence-function map can be written as a weighted sum over subsequences in multiple ways, meaningfully interpreting these weights requires "gauge-fixing," i.e., defining a unique representation for each map. Recent work has established that most existing gauge-fixed representations arise as the unique solutions to $L_2$-regularized regression in an overparameterized "weight space" where the choice of regularizer defines the gauge. Here, we establish the relationship between regularized regression in overparameterized weight space and Gaussian process approaches that operate in "function space," i.e. the space of all real-valued functions on a finite set of sequences. We disentangle how weight space regularizers both impose an implicit prior on the learned function and restrict the optimal weights to a particular gauge. We also show how to construct regularizers that correspond to arbitrary explicit Gaussian process priors combined with a wide variety of gauges. Next, we derive the distribution of gauge-fixed weights implied by the Gaussian process posterior and demonstrate that even for long sequences this distribution can be efficiently computed for product-kernel priors using a kernel trick. Finally, we characterize the implicit function space priors associated with the most common weight space regularizers. Overall, our framework unifies and extends our ability to infer and interpret sequence-function relationships.
Authors: Henry Herzog, Joshua Hansen, Yawen Zhang, Patrick Beukema
Abstract: Unsustainable exploitation of the oceans exacerbated by global warming is threatening coastal communities worldwide. Accurate and timely monitoring of maritime activity is an essential step to effective governance and to inform future policy. In support of this complex global-scale effort, we built Atlantes, a deep learning based system that provides the first-ever real-time view of vessel behavior at global scale. Atlantes leverages a series of bespoke transformers to distill a high volume, continuous stream of GPS messages emitted by hundreds of thousands of vessels into easily quantifiable behaviors. The combination of low latency and high performance enables operationally relevant decision-making and successful interventions on the high seas where illegal and exploitative activity is too common. Atlantes is already in use by hundreds of organizations worldwide. Here we provide an overview of the model and infrastructure that enables this system to function efficiently and cost-effectively at global-scale and in real-time.
Authors: Nandan Joshi, Erhan Guven
Abstract: The growing demand for molecules with tailored properties in fields such as drug discovery and chemical engineering has driven advancements in computational methods for molecular design. Machine learning-based approaches for de-novo molecular generation have recently garnered significant attention. This paper introduces a transformer-based vector embedding generator combined with a modified Generative Adversarial Network (GAN) to generate molecules with desired properties. The embedding generator utilizes a novel molecular descriptor, integrating Morgan fingerprints with global molecular attributes, enabling the transformer to capture local functional groups and broader molecular characteristics. Modifying the GAN generator loss function ensures the generation of molecules with specific desired properties. The transformer achieves a reconversion accuracy of 94% while translating molecular descriptors back to SMILES strings, validating the utility of the proposed embeddings for generative tasks. The approach is validated by generating novel odorant molecules using a labeled dataset of odorant and non-odorant compounds. With the modified range-loss function, the GAN exclusively generates odorant molecules. This work underscores the potential of combining novel vector embeddings with transformers and modified GAN architectures to accelerate the discovery of tailored molecules, offering a robust tool for diverse molecular design applications.
Authors: Elliot L. Epstein, Rajat Dwaraknath, Thanawat Sornwanee, John Winnicki, Jerry Weihong Liu
Abstract: We propose a novel method for density estimation that leverages an estimated score function to debias kernel density estimation (SD-KDE). In our approach, each data point is adjusted by taking a single step along the score function with a specific choice of step size, followed by standard KDE with a modified bandwidth. The step size and modified bandwidth are chosen to remove the leading order bias in the KDE. Our experiments on synthetic tasks in 1D, 2D and on MNIST, demonstrate that our proposed SD-KDE method significantly reduces the mean integrated squared error compared to the standard Silverman KDE, even with noisy estimates in the score function. These results underscore the potential of integrating score-based corrections into nonparametric density estimation.
Authors: Shunxin Guo, Jiaqi Lv, Xin Geng
Abstract: We introduce Ring-topology Decentralized Federated Learning (RDFL) for distributed model training, aiming to avoid the inherent risks of centralized failure in server-based FL. However, RDFL faces the challenge of low information-sharing efficiency due to the point-to-point communication manner when handling inherent data heterogeneity. Existing studies to mitigate data heterogeneity focus on personalized optimization of models, ignoring that the lack of shared information constraints can lead to large differences among models, weakening the benefits of collaborative learning. To tackle these challenges, we propose a Divide-and-conquer RDFL framework (DRDFL) that uses a feature generation model to extract personalized information and invariant shared knowledge from the underlying data distribution, ensuring both effective personalization and strong generalization. Specifically, we design a \textit{PersonaNet} module that encourages class-specific feature representations to follow a Gaussian mixture distribution, facilitating the learning of discriminative latent representations tailored to local data distributions. Meanwhile, the \textit{Learngene} module is introduced to encapsulate shared knowledge through an adversarial classifier to align latent representations and extract globally invariant information. Extensive experiments demonstrate that DRDFL outperforms state-of-the-art methods in various data heterogeneity settings.
Authors: Yun Qu (Cheems), Qi (Cheems), Wang, Yixiu Mao, Yiqin Lv, Xiangyang Ji
Abstract: Task robust adaptation is a long-standing pursuit in sequential decision-making. Some risk-averse strategies, e.g., the conditional value-at-risk principle, are incorporated in domain randomization or meta reinforcement learning to prioritize difficult tasks in optimization, which demand costly intensive evaluations. The efficiency issue prompts the development of robust active task sampling to train adaptive policies, where risk-predictive models are used to surrogate policy evaluation. This work characterizes the optimization pipeline of robust active task sampling as a Markov decision process, posits theoretical and practical insights, and constitutes robustness concepts in risk-averse scenarios. Importantly, we propose an easy-to-implement method, referred to as Posterior and Diversity Synergized Task Sampling (PDTS), to accommodate fast and robust sequential decision-making. Extensive experiments show that PDTS unlocks the potential of robust active task sampling, significantly improves the zero-shot and few-shot adaptation robustness in challenging tasks, and even accelerates the learning process under certain scenarios. Our project website is at https://thu-rllab.github.io/PDTS_project_page.
Authors: Panagiotis Kakosimos
Abstract: The electrification of powertrains is rising as the objective for a more viable future is intensified. To ensure continuous and reliable operation without undesirable malfunctions, it is essential to monitor the internal temperatures of machines and keep them within safe operating limits. Conventional modeling methods can be complex and usually require expert knowledge. With the amount of data collected these days, it is possible to use information models to assess thermal behaviors. This paper investigates artificial intelligence techniques for monitoring the cooling efficiency of induction machines. Experimental data was collected under specific operating conditions, and three machine-learning models have been developed. The optimal configuration for each approach was determined through rigorous hyperparameter searches, and the models were evaluated using a variety of metrics. The three solutions performed well in monitoring the condition of the machine even under transient operation, highlighting the potential of data-driven methods in improving the thermal management.
Authors: Piotr Migus
Abstract: Complex-valued neural networks (CVNNs) excel where phase matters, yet their multi-sheeted decision surfaces defy standard explainability and calibration tools. We propose a \emph{Newton-Puiseux} framework that fits a local polynomial surrogate to a high-uncertainty input and analytically decomposes this surrogate into fractional-power series. The resulting Puiseux expansions, dominant Puiseux coefficients, and phase-aligned curvature descriptors deliver closed-form estimates of robustness and over-confidence that gradient - or perturbation-based methods (saliency, LIME, SHAP) cannot provide. On a controlled $\mathbb{C}^2$ helix the surrogate attains RMSE $< 0.09$ while recovering the number of decision sheets; quartic coefficients predict adversarial flip radii within $10^{-3}$. On the real-world MIT-BIH arrhythmia corpus, Puiseux-guided, phase-aware temperature scaling lowers expected calibration error from 0.087 to 0.034, contributing to the advancement of CVNNs. Full code, pre-trained weights, and scripts are at https://github.com/piotrmgs/puiseux-cvnn.
Authors: Jianlong Chen, Chao Li, Yang Yuan, Andrew C Yao
Abstract: Large language models (LLMs) have shown promise in formal theorem proving, but their token-level processing often fails to capture the inherent hierarchical nature of mathematical proofs. We introduce \textbf{Hierarchical Attention}, a regularization method that aligns LLMs' attention mechanisms with mathematical reasoning structures. Our approach establishes a five-level hierarchy from foundational elements to high-level concepts, ensuring structured information flow in proof generation. Experiments demonstrate that our method improves proof success rates by 2.05\% on miniF2F and 1.69\% on ProofNet while reducing proof complexity by 23.81\% and 16.50\% respectively. The code is available at https://github.com/Car-pe/HAGBP.
Authors: Ming Xu, Jinrong Xiang, Zilong Xie, Xiangfu Meng
Abstract: Accurately identifying critical nodes with high spatial influence in road networks is essential for enhancing the efficiency of traffic management and urban planning. However, existing node importance ranking methods mainly rely on structural features and topological information, often overlooking critical factors such as origin-destination (OD) demand and route information. This limitation leaves considerable room for improvement in ranking accuracy. To address this issue, we propose HetGL2R, an attributed heterogeneous graph learning approach for ranking node importance in road networks. This method introduces a tripartite graph (trip graph) to model the structure of the road network, integrating OD demand, route choice, and various structural features of road segments. Based on the trip graph, we design an embedding method to learn node representations that reflect the spatial influence of road segments. The method consists of a heterogeneous random walk sampling algorithm (HetGWalk) and a Transformer encoder. HetGWalk constructs multiple attribute-guided graphs based on the trip graph to enrich the diversity of semantic associations between nodes. It then applies a joint random walk mechanism to convert both topological structures and node attributes into sequences, enabling the encoder to capture spatial dependencies more effectively among road segments. Finally, a listwise ranking strategy is employed to evaluate node importance. To validate the performance of our method, we construct two synthetic datasets using SUMO based on simulated road networks. Experimental results demonstrate that HetGL2R significantly outperforms baselines in incorporating OD demand and route choice information, achieving more accurate and robust node ranking. Furthermore, we conduct a case study using real-world taxi trajectory data from Beijing, further verifying the practicality of the proposed method.
Authors: Adwait Datar, Nihat Ay
Abstract: The Kullback-Leibler (KL) divergence plays a central role in probabilistic machine learning, where it commonly serves as the canonical loss function. Optimization in such settings is often performed over the probability simplex, where the choice of parameterization significantly impacts convergence. In this work, we study the problem of minimizing the KL divergence and analyze the behavior of gradient-based optimization algorithms under two dual coordinate systems within the framework of information geometry$-$ the exponential family ($\theta$ coordinates) and the mixture family ($\eta$ coordinates). We compare Euclidean gradient descent (GD) in these coordinates with the coordinate-invariant natural gradient descent (NGD), where the natural gradient is a Riemannian gradient that incorporates the intrinsic geometry of the parameter space. In continuous time, we prove that the convergence rates of GD in the $\theta$ and $\eta$ coordinates provide lower and upper bounds, respectively, on the convergence rate of NGD. Moreover, under affine reparameterizations of the dual coordinates, the convergence rates of GD in $\eta$ and $\theta$ coordinates can be scaled to $2c$ and $\frac{2}{c}$, respectively, for any $c>0$, while NGD maintains a fixed convergence rate of $2$, remaining invariant to such transformations and sandwiched between them. Although this suggests that NGD may not exhibit uniformly superior convergence in continuous time, we demonstrate that its advantages become pronounced in discrete time, where it achieves faster convergence and greater robustness to noise, outperforming GD. Our analysis hinges on bounding the spectrum and condition number of the Hessian of the KL divergence at the optimum, which coincides with the Fisher information matrix.
Authors: Mohammad M Maheri, Hamed Haddadi, Alex Davidson
Abstract: Verification of the integrity of deep learning inference is crucial for understanding whether a model is being applied correctly. However, such verification typically requires access to model weights and (potentially sensitive or private) training data. So-called Zero-knowledge Succinct Non-Interactive Arguments of Knowledge (ZK-SNARKs) would appear to provide the capability to verify model inference without access to such sensitive data. However, applying ZK-SNARKs to modern neural networks, such as transformers and large vision models, introduces significant computational overhead. We present TeleSparse, a ZK-friendly post-processing mechanisms to produce practical solutions to this problem. TeleSparse tackles two fundamental challenges inherent in applying ZK-SNARKs to modern neural networks: (1) Reducing circuit constraints: Over-parameterized models result in numerous constraints for ZK-SNARK verification, driving up memory and proof generation costs. We address this by applying sparsification to neural network models, enhancing proof efficiency without compromising accuracy or security. (2) Minimizing the size of lookup tables required for non-linear functions, by optimizing activation ranges through neural teleportation, a novel adaptation for narrowing activation functions' range. TeleSparse reduces prover memory usage by 67% and proof generation time by 46% on the same model, with an accuracy trade-off of approximately 1%. We implement our framework using the Halo2 proving system and demonstrate its effectiveness across multiple architectures (Vision-transformer, ResNet, MobileNet) and datasets (ImageNet,CIFAR-10,CIFAR-100). This work opens new directions for ZK-friendly model design, moving toward scalable, resource-efficient verifiable deep learning.
Authors: Yiyang Zhou, Zhaoyang Wang, Tianle Wang, Shangyu Xing, Peng Xia, Bo Li, Kaiyuan Zheng, Zijian Zhang, Zhaorun Chen, Wenhao Zheng, Xuchao Zhang, Chetan Bansal, Weitong Zhang, Ying Wei, Mohit Bansal, Huaxiu Yao
Abstract: High-quality preference data is essential for aligning foundation models with human values through preference learning. However, manual annotation of such data is often time-consuming and costly. Recent methods often adopt a self-rewarding approach, where the target model generates and annotates its own preference data, but this can lead to inaccuracies since the reward model shares weights with the target model, thereby amplifying inherent biases. To address these issues, we propose Anyprefer, a framework designed to synthesize high-quality preference data for aligning the target model. Anyprefer frames the data synthesis process as a cooperative two-player Markov Game, where the target model and the judge model collaborate together. Here, a series of external tools are introduced to assist the judge model in accurately rewarding the target model's responses, mitigating biases in the rewarding process. In addition, a feedback mechanism is introduced to optimize prompts for both models, enhancing collaboration and improving data quality. The synthesized data is compiled into a new preference dataset, Anyprefer-V1, consisting of 58K high-quality preference pairs. Extensive experiments show that Anyprefer significantly improves model alignment performance across four main applications, covering 21 datasets, achieving average improvements of 18.55% in five natural language generation datasets, 3.66% in nine vision-language understanding datasets, 30.05% in three medical image analysis datasets, and 16.00% in four visuo-motor control tasks.
Authors: Angel Mary John, Aiswarya M. U., Jerrin Thomas Panachakel
Abstract: Artificial intelligence (AI) has emerged as a ubiquitous concept in numerous domains, including the legal system. AI has the potential to revolutionize the functioning of the judiciary and the dispensation of justice. Incorporating AI into the legal system offers the prospect of enhancing decision-making for judges, lawyers, and legal professionals, while concurrently providing the public with more streamlined, efficient, and cost-effective services. The integration of AI into the legal landscape offers manifold benefits, encompassing tasks such as document review, legal research, contract analysis, case prediction, and decision-making. By automating laborious and error-prone procedures, AI has the capacity to alleviate the burden associated with these arduous tasks. Consequently, courts around the world have begun embracing AI technology as a means to enhance the administration of justice. However, alongside its potential advantages, the use of AI in the judiciary poses a range of ethical challenges. These ethical quandaries must be duly addressed to ensure the responsible and equitable deployment of AI systems. This article delineates the principal ethical challenges entailed in employing AI within the judiciary and provides recommendations to effectively address these issues.
Authors: Weitao Du, Shuning Chang, Jiasheng Tang, Yu Rong, Fan Wang, Shengchao Liu
Abstract: In this work, we propose a novel generative learning paradigm, K-Flow, an algorithm that flows along the $K$-amplitude. Here, $k$ is a scaling parameter that organizes frequency bands (or projected coefficients), and amplitude describes the norm of such projected coefficients. By incorporating the $K$-amplitude decomposition, K-Flow enables flow matching across the scaling parameter as time. We discuss three venues and six properties of K-Flow, from theoretical foundations, energy and temporal dynamics, and practical applications, respectively. Specifically, from the practical usage perspective, K-Flow allows steerable generation by controlling the information at different scales. To demonstrate the effectiveness of K-Flow, we conduct experiments on unconditional image generation, class-conditional image generation, and molecule assembly generation. Additionally, we conduct three ablation studies to demonstrate how K-Flow steers scaling parameter to effectively control the resolution of image generation.
Authors: Suping Xu, Chuyi Dai, Lin Shang, Changbin Shao, Xibei Yang, Witold Pedrycz
Abstract: Label distribution learning (LDL) is an emerging learning paradigm designed to capture the relative importance of labels for each instance. Label-specific features (LSFs), constructed by LIFT, have proven effective for learning tasks with label ambiguity by leveraging clustering-based prototypes for each label to re-characterize instances. However, directly introducing LIFT into LDL tasks can be suboptimal, as the prototypes it collects primarily reflect intra-cluster relationships while neglecting interactions among distinct clusters. Additionally, constructing LSFs using multi-perspective information, rather than relying solely on Euclidean distance, provides a more robust and comprehensive representation of instances, mitigating noise and bias that may arise from a single distance perspective. To address these limitations, we introduce Structural Anchor Points (SAPs) to capture inter-cluster interactions. This leads to a novel LSFs construction strategy, LIFT-SAP, which enhances LIFT by integrating both distance and direction information of each instance relative to SAPs. Furthermore, we propose a novel LDL algorithm, Label Distribution Learning via Label-specifIc FeaTure with SAPs (LDL-LIFT-SAP), which unifies multiple label description degrees predicted from different LSF spaces into a cohesive label distribution. Extensive experiments on 15 real-world datasets demonstrate the effectiveness of LIFT-SAP over LIFT, as well as the superiority of LDL-LIFT-SAP compared to seven other well-established algorithms.
Authors: Siddharth Chandak
Abstract: Two-time-scale stochastic approximation is an algorithm with coupled iterations which has found broad applications in reinforcement learning, optimization and game control. While several prior works have obtained a mean square error bound of $O(1/k)$ for linear two-time-scale iterations, the best known bound in the non-linear contractive setting has been $O(1/k^{2/3})$. In this work, we obtain an improved bound of $O(1/k)$ for non-linear two-time-scale stochastic approximation. Our result applies to algorithms such as gradient descent-ascent and two-time-scale Lagrangian optimization. The key step in our analysis involves rewriting the original iteration in terms of an averaged noise sequence which decays sufficiently fast. Additionally, we use an induction-based approach to show that the iterates are bounded in expectation.
Authors: Jonathan Gornet, Yiannis Kantaros, Bruno Sinopoli
Abstract: We introduce Hyperparameter Controller (HyperController), a computationally efficient algorithm for hyperparameter optimization during training of reinforcement learning neural networks. HyperController optimizes hyperparameters quickly while also maintaining improvement of the reinforcement learning neural network, resulting in faster training and deployment. It achieves this by modeling the hyperparameter optimization problem as an unknown Linear Gaussian Dynamical System, which is a system with a state that linearly changes. It then learns an efficient representation of the hyperparameter objective function using the Kalman filter, which is the optimal one-step predictor for a Linear Gaussian Dynamical System. To demonstrate the performance of HyperController, it is applied as a hyperparameter optimizer during training of reinforcement learning neural networks on a variety of OpenAI Gymnasium environments. In four out of the five Gymnasium environments, HyperController achieves highest median reward during evaluation compared to other algorithms. The results exhibit the potential of HyperController for efficient and stable training of reinforcement learning neural networks.
Authors: David Warren, Mark Dras
Abstract: Model Cascading, recently applied successfully to LLMs, is a simple but powerful technique that improves the efficiency of inference by selectively applying models of varying sizes. Models are used in sequence from smallest to largest, only deferring samples to large, costly models when smaller models are not sufficiently confident. Existing approaches to deferral use only limited small model confidence estimates because of the inaccessibility of the large model, although large model confidence is known to be important. We therefore propose a bi-directional approach to deferral that considers the confidence of small and large models in the cascade simultaneously through the use of a proxy for the large model. This requires a richer representation of model confidence to enable comparative calibration: we use an analysis of hidden states to improve post-invocation confidence of the small model, which in itself improves cascading results over prior approaches. We then combine this with a tiny proxy model to estimate pre-invocation confidence of the large model. We examine the proposed cascading system over challenging, multiple-choice datasets, finding improvements over standard cascading baselines reflected in reductions in deferrals to more costly models.
Authors: Shuo Wu, Pawan Poojary, Randall Berry
Abstract: We consider a model of Bayesian observational learning in which a sequence of agents receives a private signal about an underlying binary state of the world. Each agent makes a decision based on its own signal and its observations of previous agents. A central planner seeks to improve the accuracy of these signals by allocating a limited budget to enhance signal quality across agents. We formulate and analyze the budget allocation problem and propose two optimal allocation strategies. At least one of these strategies is shown to maximize the probability of achieving a correct information cascade.
Authors: Maitreya Sonawane, Sumit Mamtani
Abstract: Making accurate weather predictions can be particularly challenging for localized storms or events that evolve on hourly timescales, such as thunderstorms. Hence, our goal for the project was to model Weather Nowcasting for making highly localized and accurate predictions that apply to the immediate future replacing the current numerical weather models and data assimilation systems with Deep Learning approaches. A significant advantage of machine learning is that inference is computationally cheap given an already-trained model, allowing forecasts that are nearly instantaneous and in the native high resolution of the input data. In this work we developed a novel method that employs Transformer-based machine learning models to forecast precipitation. This approach works by leveraging axial attention mechanisms to learn complex patterns and dynamics from time series frames. Moreover, it is a generic framework and can be applied to univariate and multivariate time series data, as well as time series embeddings data. This paper represents an initial research on the dataset used in the domain of next frame prediciton, and hence, we demonstrate state-of-the-art results in terms of metrices (PSNR = 47.67, SSIM = 0.9943) used for the given dataset using UNet with Axial Transformer.
Authors: Zhaiming Shen, Sung Ha Kang
Abstract: Local clustering aims to identify specific substructures within a large graph without requiring full knowledge of the entire graph. These substructures are typically small compared to the overall graph, enabling the problem to be approached by finding a sparse solution to a linear system associated with the graph Laplacian. In this work, we first propose a method for identifying specific local clusters when very few labeled data is given, which we term semi-supervised local clustering. We then extend this approach to the unsupervised setting when no prior information on labels is available. The proposed methods involve randomly sampling the graph, applying diffusion through local cluster extraction, then examining the overlap among the results to find each cluster. We establish the co-membership conditions for any pair of nodes and rigorously prove the correctness of our methods. Additionally, we conduct extensive experiments to demonstrate that the proposed methods achieve state-of-the-arts results in the low-label rates regime.
Authors: Arnab Bhattacharyya, Constantinos Daskalakis, Themis Gouleakis, Yuhao Wang
Abstract: We provide efficient algorithms for the problem of distribution learning from high-dimensional Gaussian data where in each sample, some of the variable values are missing. We suppose that the variables are missing not at random (MNAR). The missingness model, denoted by $S(y)$, is the function that maps any point $y$ in $R^d$ to the subsets of its coordinates that are seen. In this work, we assume that it is known. We study the following two settings: (i) Self-censoring: An observation $x$ is generated by first sampling the true value $y$ from a $d$-dimensional Gaussian $N(\mu*, \Sigma*)$ with unknown $\mu*$ and $\Sigma*$. For each coordinate $i$, there exists a set $S_i$ subseteq $R^d$ such that $x_i = y_i$ if and only if $y_i$ in $S_i$. Otherwise, $x_i$ is missing and takes a generic value (e.g., "?"). We design an algorithm that learns $N(\mu*, \Sigma*)$ up to total variation (TV) distance epsilon, using $poly(d, 1/\epsilon)$ samples, assuming only that each pair of coordinates is observed with sufficiently high probability. (ii) Linear thresholding: An observation $x$ is generated by first sampling $y$ from a $d$-dimensional Gaussian $N(\mu*, \Sigma)$ with unknown $\mu*$ and known $\Sigma$, and then applying the missingness model $S$ where $S(y) = {i in [d] : v_i^T y <= b_i}$ for some $v_1, ..., v_d$ in $R^d$ and $b_1, ..., b_d$ in $R$. We design an efficient mean estimation algorithm, assuming that none of the possible missingness patterns is very rare conditioned on the values of the observed coordinates and that any small subset of coordinates is observed with sufficiently high probability.
Authors: Zhenyu Zhang, Zechun Liu, Yuandong Tian, Harshit Khaitan, Zhangyang Wang, Steven Li
Abstract: Large Language Models (LLMs), while demonstrating remarkable capabilities across various applications, present significant challenges during inference due to their substantial model size, especially when deployed on edge devices. Activation sparsity offers a promising solution to reduce computation and memory movement, enabling more efficient inference, particularly for small-batch on-device applications. However, current approaches face limitations with non-ReLU activation function, which are foundational to most advanced LLMs, or require heavy continual training. Additionally, the difficulty in predicting active channels and limited achievable sparsity ratios constrain the effectiveness of activation sparsity-based methods. In this paper, we introduce R-Sparse, a training-free activation sparsity approach capable of achieving high sparsity levels in advanced LLMs. We conducted two preliminary investigations into how different components contribute to the output within a single linear layer and found two key observations: (i) the non-sparse components of the input function can be regarded as a few bias terms, and (ii) The full computation can be effectively approximated by an appropriate combination of input channels and weight singular values. Building on this, we replace the linear layers in LLMs with a rank-aware sparse inference method that leverages the sparsity of input channels and singular value components, eliminating the need for active channel prediction like the output sparsity based approaches. Experiments on Llama-2/3 and Mistral models across ten diverse tasks demonstrate that R-Sparse achieves comparable performance at 50% model-level sparsity, resulting in a significant 43% end-to-end efficient improvements with customized kernels.
Authors: Qibang Liu, Vincient Zhong, Hadi Meidani, Diab Abueidda, Seid Koric, Philippe Geubelle
Abstract: Machine-learning-based surrogate models offer significant computational efficiency and faster simulations compared to traditional numerical methods, especially for problems requiring repeated evaluations of partial differential equations. This work introduces the Geometry-Informed Neural Operator Transformer (GINOT), which integrates the transformer architecture with the neural operator framework to enable forward predictions for arbitrary geometries. GINOT encodes the surface points cloud of a geometry using a sampling and grouping mechanism combined with an attention mechanism, ensuring invariance to point order and padding while maintaining robustness to variations in point density. The geometry information is seamlessly integrated with query points in the solution decoder through the attention mechanism. The performance of GINOT is validated on multiple challenging datasets, showcasing its high accuracy and strong generalization capabilities for complex and arbitrary 2D and 3D geometries.
Authors: Donghe Chen, Han Wang, Lin Cheng, Shengping Gong
Abstract: Reinforcement Learning (RL) has shown promise in control tasks but faces significant challenges in real-world applications, primarily due to the absence of safety guarantees during the learning process. Existing methods often struggle with ensuring safe exploration, leading to potential system failures and restricting applications primarily to simulated environments. Traditional approaches such as reward shaping and constrained policy optimization can fail to guarantee safety during initial learning stages, while model-based methods using Control Lyapunov Functions (CLFs) or Control Barrier Functions (CBFs) may hinder efficient exploration and performance. To address these limitations, this paper introduces Soft Actor-Critic with Control Lyapunov Function (SAC-CLF), a framework that enhances stability and safety through three key innovations: (1) a task-specific CLF design method for safe and optimal performance; (2) dynamic adjustment of constraints to maintain robustness under unmodeled dynamics; and (3) improved control input smoothness while ensuring safety. Experimental results on a classical nonlinear system and satellite attitude control demonstrate the effectiveness of SAC-CLF in overcoming the shortcomings of existing methods.
Authors: Dixiao Wei, Peng Yi, Jinlong Lei, Yiguang Hong, Yuchuan Du
Abstract: Reinforcement Learning (RL) has demonstrated excellent decision-making potential in platoon coordination problems. However, due to the variability of coordination goals, the complexity of the decision problem, and the time-consumption of trial-and-error in manual design, finding a well performance reward function to guide RL training to solve complex platoon coordination problems remains challenging. In this paper, we formally define the Platoon Coordination Reward Design Problem (PCRDP), extending the RL-based cooperative platoon coordination problem to incorporate automated reward function generation. To address PCRDP, we propose a Large Language Model (LLM)-based Platoon coordination Reward Design (PCRD) framework, which systematically automates reward function discovery through LLM-driven initialization and iterative optimization. In this method, LLM first initializes reward functions based on environment code and task requirements with an Analysis and Initial Reward (AIR) module, and then iteratively optimizes them based on training feedback with an evolutionary module. The AIR module guides LLM to deepen their understanding of code and tasks through a chain of thought, effectively mitigating hallucination risks in code generation. The evolutionary module fine-tunes and reconstructs the reward function, achieving a balance between exploration diversity and convergence stability for training. To validate our approach, we establish six challenging coordination scenarios with varying complexity levels within the Yangtze River Delta transportation network simulation. Comparative experimental results demonstrate that RL agents utilizing PCRD-generated reward functions consistently outperform human-engineered reward functions, achieving an average of 10\% higher performance metrics in all scenarios.
Authors: Bertram H{\o}jer, Oliver Jarvis, Stefan Heinrich
Abstract: Recent advancements in large language models (LLMs) have resulted in increasingly anthropomorphic language concerning the ability of LLMs to reason. Whether reasoning in LLMs should be understood to be inherently different is, however, widely debated. We propose utilizing a representation engineering approach wherein model activations are read from the residual stream of an LLM when processing a reasoning task. The activations are used to derive a control vector that is applied to the model as an inference-time intervention, modulating the representational space of the model, to improve performance on the specified task. We publish the code for deriving control vectors and analyzing model representations. The method allows us to improve performance on reasoning benchmarks and assess how control vectors influence the final logit distribution of a model via metrics such as KL divergence and entropy. We apply control vectors to Mistral-7B-Instruct and a range of Pythia models on an inductive, a deductive and mathematical reasoning task. We show that an LLM can, to a certain degree, be controlled to improve its perceived reasoning ability by modulating activations. The intervention is dependent upon the ability to reliably extract the model's typical state when correctly solving a task. Our results suggest that reasoning performance can be modulated in the same manner as other information-processing tasks performed by LLMs and demonstrate that we are capable of improving performance on specific tasks via a simple intervention on the residual stream with no additional training.
Authors: Rudy Morel, Jiequn Han, Edouard Oyallon
Abstract: We address the problem of predicting the next state of a dynamical system governed by unknown temporal partial differential equations (PDEs) using only a short trajectory. While standard transformers provide a natural black-box solution to this task, the presence of a well-structured evolution operator in the data suggests a more tailored and efficient approach. Specifically, when the PDE is fully known, classical numerical solvers can evolve the state accurately with only a few parameters. Building on this observation, we introduce DISCO, a model that uses a large hypernetwork to process a short trajectory and generate the parameters of a much smaller operator network, which then predicts the next state through time integration. Our framework decouples dynamics estimation (i.e., DISCovering an evolution operator from a short trajectory) from state prediction (i.e., evolving this operator). Experiments show that pretraining our model on diverse physics datasets achieves state-of-the-art performance while requiring significantly fewer epochs. Moreover, it generalizes well and remains competitive when fine-tuned on downstream tasks.
Authors: Qinwei Yang, Ruocheng Guo, Shasha Han, Peng Wu
Abstract: Estimating long-term treatment effects has a wide range of applications in various domains. A key feature in this context is that collecting long-term outcomes typically involves a multi-stage process and is subject to monotone missing, where individuals missing at an earlier stage remain missing at subsequent stages. Despite its prevalence, monotone missing has been rarely explored in previous studies on estimating long-term treatment effects. In this paper, we address this gap by introducing the sequential missingness assumption for identification. We propose three novel estimation methods, including inverse probability weighting, sequential regression imputation, and sequential marginal structural model (SeqMSM). Considering that the SeqMSM method may suffer from high variance due to severe data sparsity caused by monotone missing, we further propose a novel balancing-enhanced approach, BalanceNet, to improve the stability and accuracy of the estimation methods. Extensive experiments on two widely used benchmark datasets demonstrate the effectiveness of our proposed methods.
Authors: Yicheng Li, Xinghua Sun
Abstract: This paper proposes and analyzes a gradient-type algorithm based on Burer-Monteiro factorization, called the Asymmetric Projected Gradient Descent (APGD), for reconstructing the point set configuration from partial Euclidean distance measurements, known as the Euclidean Distance Matrix Completion (EDMC) problem. By paralleling the incoherence matrix completion framework, we show for the first time that global convergence guarantee with exact recovery of this routine can be established given $\mathcal{O}(\mu^2 r^3 \kappa^2 n \log n)$ Bernoulli random observations without any sample splitting. Unlike leveraging the tangent space Restricted Isometry Property (RIP) and local curvature of the low-rank embedding manifold in some very recent works, our proof provides new upper bounds to replace the random graph lemma under EDMC setting. The APGD works surprisingly well and numerical experiments demonstrate exact linear convergence behavior in rich-sample regions yet deteriorates fast when compared with the performance obtained by optimizing the s-stress function, i.e., the standard but unexplained non-convex approach for EDMC, if the sample size is limited. While virtually matching our theoretical prediction, this unusual phenomenon might indicate that: (i) the power of implicit regularization is weakened when specified in the APGD case; (ii) the stabilization of such new gradient direction requires substantially more samples than the information-theoretic limit would suggest.
Authors: Yasir Ghunaim, Andr\'es Villa, Gergo Ignacz, Gyorgy Szekely, Motasem Alfarra, Bernard Ghanem
Abstract: Advancements in machine learning for molecular property prediction have improved accuracy but at the expense of higher computational cost and longer training times. Recently, the Joint Multi-domain Pre-training (JMP) foundation model has demonstrated strong performance across various downstream tasks with reduced training time over previous models. Despite JMP's advantages, fine-tuning it on molecular datasets ranging from small-scale to large-scale requires considerable time and computational resources. In this work, we investigate strategies to enhance efficiency by reducing model size while preserving performance. To better understand the model's efficiency, we analyze the layer contributions of JMP and find that later interaction blocks provide diminishing returns, suggesting an opportunity for model compression. We explore block reduction strategies by pruning the pre-trained model and evaluating its impact on efficiency and accuracy during fine-tuning. Our analysis reveals that removing two interaction blocks results in a minimal performance drop, reducing the model size by 32% while increasing inference throughput by 1.3x. These results suggest that JMP-L is over-parameterized and that a smaller, more efficient variant can achieve comparable performance with lower computational cost. Our study provides insights for developing lighter, faster, and more scalable foundation models for molecular and materials discovery. The code is publicly available at: https://github.com/Yasir-Ghunaim/efficient-jmp.
Authors: Rom N. Parnichkun, Neehal Tumma, Armin W. Thomas, Alessandro Moro, Qi An, Taiji Suzuki, Atsushi Yamashita, Michael Poli, Stefano Massaroli
Abstract: The need to develop a general framework for architecture analysis is becoming increasingly important, given the expanding design space of sequence models. To this end, we draw insights from classical signal processing and control theory, to develop a quantitative measure of \textit{memory utilization}: the internal mechanisms through which a model stores past information to produce future outputs. This metric, which we call \textbf{\textit{effective state-size}} (ESS), is tailored to the fundamental class of systems with \textit{input-invariant} and \textit{input-varying linear operators}, encompassing a variety of computational units such as variants of attention, convolutions, and recurrences. Unlike prior work on memory utilization, which either relies on raw operator visualizations (e.g. attention maps), or simply the total \textit{memory capacity} (i.e. cache size) of a model, our metrics provide highly interpretable and actionable measurements. In particular, we show how ESS can be leveraged to improve initialization strategies, inform novel regularizers and advance the performance-efficiency frontier through model distillation. Furthermore, we demonstrate that the effect of context delimiters (such as end-of-speech tokens) on ESS highlights cross-architectural differences in how large language models utilize their available memory to recall information. Overall, we find that ESS provides valuable insights into the dynamics that dictate memory utilization, enabling the design of more efficient and effective sequence models.
Authors: Hanlu Zhang, Yumeng Ma, Shuo Wang, Guiran Liu, Binrong Zhu
Abstract: This paper proposes a parameter collaborative optimization algorithm for large language models, enhanced with graph spectral analysis. The goal is to improve both fine-tuning efficiency and structural awareness during training. In the proposed method, the parameters of a pre-trained language model are treated as nodes in a graph. A weighted graph is constructed, and Laplacian spectral decomposition is applied to enable frequency-domain modeling and structural representation of the parameter space. Based on this structure, a joint loss function is designed. It combines the task loss with a spectral regularization term to facilitate collaborative updates among parameters. In addition, a spectral filtering mechanism is introduced during the optimization phase. This mechanism adjusts gradients in a structure-aware manner, enhancing the model's training stability and convergence behavior. The method is evaluated on multiple tasks, including traditional fine-tuning comparisons, few-shot generalization tests, and convergence speed analysis. In all settings, the proposed approach demonstrates superior performance. The experimental results confirm that the spectral collaborative optimization framework effectively reduces parameter perturbations and improves fine-tuning quality while preserving overall model performance. This work contributes significantly to the field of artificial intelligence by advancing parameter-efficient training methodologies for large-scale models, reinforcing the importance of structural signal processing in deep learning optimization, and offering a robust, generalizable framework for enhancing language model adaptability and performance.
Authors: Kitsuya Azuma, Takayuki Nishio, Yuichi Kitagawa, Wakako Nakano, Takahito Tanimura
Abstract: Federated Learning (FL) enables collaborative model training across decentralized clients, enhancing privacy by keeping data local. Yet conventional FL, relying on frequent parameter-sharing, suffers from high communication overhead and limited model heterogeneity. Distillation-based FL approaches address these issues by sharing predictions (soft-labels) instead, but they often involve redundant transmissions across communication rounds, reducing efficiency. We propose SCARLET, a novel framework integrating synchronized soft-label caching and an enhanced Entropy Reduction Aggregation (Enhanced ERA) mechanism. SCARLET minimizes redundant communication by reusing cached soft-labels, achieving up to 50% reduction in communication costs compared to existing methods while maintaining accuracy. Enhanced ERA can be tuned to adapt to non-IID data variations, ensuring robust aggregation and performance in diverse client scenarios. Experimental evaluations demonstrate that SCARLET consistently outperforms state-of-the-art distillation-based FL methods in terms of accuracy and communication efficiency. The implementation of SCARLET is publicly available at https://github.com/kitsuyaazuma/SCARLET.
Authors: Haroui Ma, Francesco Quinzan, Theresa Willem, Stefan Bauer
Abstract: Machine learning (ML) systems for medical imaging have demonstrated remarkable diagnostic capabilities, but their susceptibility to biases poses significant risks, since biases may negatively impact generalization performance. In this paper, we introduce a novel statistical framework to evaluate the dependency of medical imaging ML models on sensitive attributes, such as demographics. Our method leverages the concept of counterfactual invariance, measuring the extent to which a model's predictions remain unchanged under hypothetical changes to sensitive attributes. We present a practical algorithm that combines conditional latent diffusion models with statistical hypothesis testing to identify and quantify such biases without requiring direct access to counterfactual data. Through experiments on synthetic datasets and large-scale real-world medical imaging datasets, including \textsc{cheXpert} and MIMIC-CXR, we demonstrate that our approach aligns closely with counterfactual fairness principles and outperforms standard baselines. This work provides a robust tool to ensure that ML diagnostic systems generalize well, e.g., across demographic groups, offering a critical step towards AI safety in healthcare. Code: https://github.com/Neferpitou3871/AI-Alignment-Medical-Imaging.
URLs: https://github.com/Neferpitou3871/AI-Alignment-Medical-Imaging.
Authors: Biqing Duan, Qing Wang, Di Liu, Wei Zhou, Zhenli He, Shengfa Miao
Abstract: Incremental learning that learns new classes over time after the model's deployment is becoming increasingly crucial, particularly for industrial edge systems, where it is difficult to communicate with a remote server to conduct computation-intensive learning. As more classes are expected to learn after their execution for edge devices. In this paper, we propose LODAP, a new on-device incremental learning framework for edge systems. The key part of LODAP is a new module, namely Efficient Incremental Module (EIM). EIM is composed of normal convolutions and lightweight operations. During incremental learning, EIM exploits some lightweight operations, called adapters, to effectively and efficiently learn features for new classes so that it can improve the accuracy of incremental learning while reducing model complexity as well as training overhead. The efficiency of LODAP is further enhanced by a data pruning strategy that significantly reduces the training data, thereby lowering the training overhead. We conducted extensive experiments on the CIFAR-100 and Tiny- ImageNet datasets. Experimental results show that LODAP improves the accuracy by up to 4.32\% over existing methods while reducing around 50\% of model complexity. In addition, evaluations on real edge systems demonstrate its applicability for on-device machine learning. The code is available at https://github.com/duanbiqing/LODAP.
Authors: Youngjoon Lee, Jinu Gong, Joonhyuk Kang
Abstract: Federated Learning (FL) enables model training across decentralized devices without sharing raw data, thereby preserving privacy in sensitive domains like healthcare. In this paper, we evaluate Kolmogorov-Arnold Networks (KAN) architectures against traditional MLP across six state-of-the-art FL algorithms on a blood cell classification dataset. Notably, our experiments demonstrate that KAN can effectively replace MLP in federated environments, achieving superior performance with simpler architectures. Furthermore, we analyze the impact of key hyperparameters-grid size and network architecture-on KAN performance under varying degrees of Non-IID data distribution. Additionally, our ablation studies reveal that optimizing KAN width while maintaining minimal depth yields the best performance in federated settings. As a result, these findings establish KAN as a promising alternative for privacy-preserving medical imaging applications in distributed healthcare. To the best of our knowledge, this is the first comprehensive benchmark of KAN in FL settings for medical imaging task.
Authors: Lei Xu, Shanshan Wang, Emmanuel Casseau, Chenglong Xiao
Abstract: High-level synthesis (HLS) design space exploration (DSE) is an optimization process in electronic design automation (EDA) that systematically explores high-level design configurations to achieve Pareto-optimal hardware implementations balancing performance, area, and power (PPA). To optimize this process, HLS prediction tasks often employ message-passing neural networks (MPNNs), leveraging complex architectures to achieve high accuracy. These predictors serve as evaluators in the DSE process, effectively bypassing the time-consuming estimations traditionally required by HLS tools. However, existing models often prioritize structural complexity and minimization of training loss, overlooking task-specific characteristics. Additionally, while evolutionary algorithms are widely used in DSE, they typically require extensive domain-specific knowledge to design effective crossover and mutation operators. To address these limitations, we propose CoGNNs-LLMEA, a framework that integrates a graph neural network with task-adaptive message passing and a large language model-enhanced evolutionary algorithm. As a predictive model, CoGNNs directly leverages intermediate representations generated from source code after compiler front-end processing, enabling prediction of quality of results (QoR) without invoking HLS tools. Due to its strong adaptability to tasks, CoGNNs can be tuned to predict post-HLS and post-implementation outcomes, effectively bridging the gap between high-level abstractions and physical implementation characteristics. CoGNNs achieves state-of-the-art prediction accuracy in post-HLS QoR prediction, reducing mean prediction errors by 2.8$\times$ for latency and 3.4$\times$ for resource utilization compared to baseline models.
Authors: Muhammad Sabih, Abrarul Karim, Jakob Wittmann, Frank Hannig, J\"urgen Teich
Abstract: The customizability of RISC-V makes it an attractive choice for accelerating deep neural networks (DNNs). It can be achieved through instruction set extensions and corresponding custom functional units. Yet, efficiently exploiting these opportunities requires a hardware/software co-design approach in which the DNN model, software, and hardware are designed together. In this paper, we propose novel RISC-V extensions for accelerating DNN models containing semi-structured and unstructured sparsity. While the idea of accelerating structured and unstructured pruning is not new, our novel design offers various advantages over other designs. To exploit semi-structured sparsity, we take advantage of the fine-grained (bit-level) configurability of FPGAs and suggest reserving a few bits in a block of DNN weights to encode the information about sparsity in the succeeding blocks. The proposed custom functional unit utilizes this information to skip computations. To exploit unstructured sparsity, we propose a variable cycle sequential multiply-and-accumulate unit that performs only as many multiplications as the non-zero weights. Our implementation of unstructured and semi-structured pruning accelerators can provide speedups of up to a factor of 3 and 4, respectively. We then propose a combined design that can accelerate both types of sparsities, providing speedups of up to a factor of 5. Our designs consume a small amount of additional FPGA resources such that the resulting co-designs enable the acceleration of DNNs even on small FPGAs. We benchmark our designs on standard TinyML applications such as keyword spotting, image classification, and person detection.
Authors: Michael Banf, Johannes Kuhn
Abstract: Large Language Models (LLMs) have shown remarkable capabilities across various domains, yet they struggle with knowledge-intensive tasks in areas that demand factual accuracy, e.g. industrial automation and healthcare. Key limitations include their tendency to hallucinate, lack of source traceability (provenance), and challenges in timely knowledge updates. Combining language models with knowledge graphs (GraphRAG) offers promising avenues for overcoming these deficits. However, a major challenge lies in creating such a knowledge graph in the first place. Here, we propose a novel approach that combines LLMs with a tripartite knowledge graph representation, which is constructed by connecting complex, domain-specific objects via a curated ontology of corresponding, domain-specific concepts to relevant sections within chunks of text through a concept-anchored pre-analysis of source documents starting from an initial lexical graph. As a consequence, our Tripartite-GraphRAG approach implements: i) a concept-specific, information-preserving pre-compression of textual chunks; ii) allows for the formation of a concept-specific relevance estimation of embedding similarities grounded in statistics; and iii) avoids common challenges w.r.t. continuous extendability, such as the need for entity resolution and deduplication. By applying a transformation to the knowledge graph, we formulate LLM prompt creation as an unsupervised node classification problem, drawing on ideas from Markov Random Fields. We evaluate our approach on a healthcare use case, involving multi-faceted analyses of patient anamneses given a set of medical concepts as well as clinical literature. Experiments indicate that it can optimize information density, coverage, and arrangement of LLM prompts while reducing their lengths, which may lead to reduced costs and more consistent and reliable LLM outputs.
Authors: Yonghui Zhai, Yang Zhang, Minghao Shang, Lihua Pang, Yaxin Ren
Abstract: Graph Transformers (GTs) have shown advantages in numerous graph structure tasks but their self-attention mechanism ignores the generalization bias of graphs, with existing methods mainly compensating for this bias from aspects like position encoding, attention bias and relative distance yet still having sub-optimal performance and being insufficient by only considering the structural perspective of generalization bias. To address this, this paper proposes Grafourierformer, which innovatively combines GT with inductive bias containing Frequency-Structure information by applying Graph Fourier Transform to the Attention Matrix: specifically, eigenvalues from the Graph Laplacian matrix are used to construct an Eigenvalue matrix mask (reflecting node positions and structural relationships with neighboring nodes to enable consideration of node range structural characteristics and focus on local graph details), and inverse Fourier transform is employed to extract node high-frequency and low-frequency features, calculate low-frequency and high-frequency energy, and construct a node frequency-energy matrix to filter the eigenvalue matrix mask, allowing attention heads to incorporate both graph structural information and node frequency information optimization, adaptively distinguish global trends from local details, and effectively suppress redundant information interference. Extensive experiments on various benchmarks show Grafourierformer consistently outperforms GNN and GT-based models in graph classification and node classification tasks, with ablation experiments further validating the effectiveness and necessity of the method. Codes are available at https://github.com/Arichibald/Grafourierformer.git
Authors: Xilong Xie, Liang Wang, Limin Xiao, Meng Han, Lin Sun, Shuai Zheng, Xiangrong Xu
Abstract: Large language models (LLMs) have significantly advanced the natural language processing paradigm but impose substantial demands on memory and computational resources. Quantization is one of the most effective ways to reduce memory consumption of LLMs. However, advanced single-precision quantization methods experience significant accuracy degradation when quantizing to ultra-low bits. Existing mixed-precision quantization methods are quantized by groups with coarse granularity. Employing high precision for group data leads to substantial memory overhead, whereas low precision severely impacts model accuracy. To address this issue, we propose FineQ, software-hardware co-design for low-bit fine-grained mixed-precision quantization of LLMs. First, FineQ partitions the weights into finer-grained clusters and considers the distribution of outliers within these clusters, thus achieving a balance between model accuracy and memory overhead. Then, we propose an outlier protection mechanism within clusters that uses 3 bits to represent outliers and introduce an encoding scheme for index and data concatenation to enable aligned memory access. Finally, we introduce an accelerator utilizing temporal coding that effectively supports the quantization algorithm while simplifying the multipliers in the systolic array. FineQ achieves higher model accuracy compared to the SOTA mixed-precision quantization algorithm at a close average bit-width. Meanwhile, the accelerator achieves up to 1.79x energy efficiency and reduces the area of the systolic array by 61.2%.
Authors: Nicola Debole, Pietro Barbiero, Francesco Giannini, Andrea Passeggini, Stefano Teso, Emanuele Marconato
Abstract: Concept Bottleneck Models (CBMs) are neural networks designed to conjoin high performance with ante-hoc interpretability. CBMs work by first mapping inputs (e.g., images) to high-level concepts (e.g., visible objects and their properties) and then use these to solve a downstream task (e.g., tagging or scoring an image) in an interpretable manner. Their performance and interpretability, however, hinge on the quality of the concepts they learn. The go-to strategy for ensuring good quality concepts is to leverage expert annotations, which are expensive to collect and seldom available in applications. Researchers have recently addressed this issue by introducing "VLM-CBM" architectures that replace manual annotations with weak supervision from foundation models. It is however unclear what is the impact of doing so on the quality of the learned concepts. To answer this question, we put state-of-the-art VLM-CBMs to the test, analyzing their learned concepts empirically using a selection of significant metrics. Our results show that, depending on the task, VLM supervision can sensibly differ from expert annotations, and that concept accuracy and quality are not strongly correlated. Our code is available at https://github.com/debryu/CQA.
Authors: Claudia Drygala, Hanno Gottschalk, Thomas Kruse, S\'egol\`ene Martin, Annika M\"utze
Abstract: Brenier proved that under certain conditions on a source and a target probability measure there exists a strictly convex function such that its gradient is a transport map from the source to the target distribution. This function is called the Brenier potential. Furthermore, detailed information on the H\"older regularity of the Brenier potential is available. In this work we develop the statistical learning theory of generative adversarial neural networks that learn the Brenier potential. As by the transformation of densities formula, the density of the generated measure depends on the second derivative of the Brenier potential, we develop the universal approximation theory of ReCU networks with cubic activation $\mathtt{ReCU}(x)=\max\{0,x\}^3$ that combines the favorable approximation properties of H\"older functions with a Lipschitz continuous density. In order to assure the convexity of such general networks, we introduce an adversarial training procedure for a potential function represented by the ReCU networks that combines the classical discriminator cross entropy loss with a penalty term that enforces (strict) convexity. We give a detailed decomposition of learning errors and show that for a suitable high penalty parameter all networks chosen in the adversarial min-max optimization problem are strictly convex. This is further exploited to prove the consistency of the learning procedure for (slowly) expanding network capacity. We also implement the described learning algorithm and apply it to a number of standard test cases from Gaussian mixture to image data as target distributions. As predicted in theory, we observe that the convexity loss becomes inactive during the training process and the potentials represented by the neural networks have learned convexity.
Authors: Haishan Wang, Arno Solin, Vikas Garg
Abstract: Graph neural networks (GNNs) are known to be vulnerable to oversmoothing due to their implicit homophily assumption. We mitigate this problem with a novel scheme that regulates the aggregation of messages, modulating the type and extent of message passing locally thereby preserving both the low and high-frequency components of information. Our approach relies solely on learnt embeddings, obviating the need for auxiliary labels, thus extending the benefits of heterophily-aware embeddings to broader applications, e.g., generative modelling. Our experiments, conducted across various data sets and GNN architectures, demonstrate performance enhancements and reveal heterophily patterns across standard classification benchmarks. Furthermore, application to molecular generation showcases notable performance improvements on chemoinformatics benchmarks.
Authors: Runtian Zhai
Abstract: This dissertation establishes the contexture theory to mathematically characterize the mechanism of representation learning, or pretraining. Despite the remarkable empirical success of foundation models, it is not very clear what representations they learn, and why these representations are useful for various downstream tasks. A scientific understanding of representation learning is critical, especially at this point when scaling up the model size is producing diminishing returns, and designing new pretraining methods is imperative for further progress. Prior work treated different representation learning methods quite differently, whereas the contexture theory provides a unified framework for analyzing these methods. The central argument is that a representation is learned from the association between the input X and a context variable A. We prove that if an encoder captures the maximum information of this association, in which case we say that the encoder learns the contexture, then it will be optimal on the class of tasks that are compatible with the context. We also show that a context is the most useful when the association between X and A is neither too strong nor too weak. The important implication of the contexture theory is that increasing the model size alone will achieve diminishing returns, and further advancements require better contexts. We demonstrate that many pretraining objectives can learn the contexture, including supervised learning, self-supervised learning, generative models, etc. Then, we introduce two general objectives -- SVME and KISE, for learning the contexture. We also show how to mix multiple contexts together, an effortless way to create better contexts from existing ones. Then, we prove statistical learning bounds for representation learning. Finally, we discuss the effect of the data distribution shift from pretraining to the downstream task.
Authors: Yoonhyuk Choi, Chong-Kwon Kim
Abstract: Recent research on graph neural networks (GNNs) has explored mechanisms for capturing local uncertainty and exploiting graph hierarchies to mitigate data sparsity and leverage structural properties. However, the synergistic integration of these two approaches remains underexplored. In this work, we introduce a novel architecture, the Hierarchical Uncertainty-Aware Graph Neural Network (HU-GNN), which unifies multi-scale representation learning, principled uncertainty estimation, and self-supervised embedding diversity within a single end-to-end framework. Specifically, HU-GNN adaptively forms node clusters and estimates uncertainty at multiple structural scales from individual nodes to higher levels. These uncertainty estimates guide a robust message-passing mechanism and attention weighting, effectively mitigating noise and adversarial perturbations while preserving predictive accuracy on both node- and graph-level tasks. We also offer key theoretical contributions, including a probabilistic formulation, rigorous uncertainty-calibration guarantees, and formal robustness bounds. Finally, by incorporating recent advances in graph contrastive learning, HU-GNN maintains diverse, structurally faithful embeddings. Extensive experiments on standard benchmarks demonstrate that our model achieves state-of-the-art robustness and interpretability.
Authors: Minjong Cheon
Abstract: Recent advances in AI-based weather forecasting models, such as FourCastNet, Pangu-Weather, and GraphCast, have demonstrated the remarkable ability of deep learning to emulate complex atmospheric dynamics. Building on this momentum, we propose Mj\"olnir, a novel deep learning-based framework for global lightning flash density parameterization. Trained on ERA5 atmospheric predictors and World Wide Lightning Location Network (WWLLN) observations at a daily temporal resolution and 1 degree spatial resolution, Mj\"olnir captures the nonlinear mapping between large-scale environmental conditions and lightning activity. The model architecture is based on the InceptionNeXt backbone with SENet, and a multi-task learning strategy to simultaneously predict lightning occurrence and magnitude. Extensive evaluations yield that Mollnir accurately reproduces the global distribution, seasonal variability, and regional characteristics of lightning activity, achieving a global Pearson correlation coefficient of 0.96 for annual mean fields. These results suggest that Mj\"olnir serves not only as an effective data-driven global lightning parameterization but also as a promising AI-based scheme for next-generation Earth system models (AI-ESMs).
Authors: Amir Zandieh, Majid Daliri, Majid Hadian, Vahab Mirrokni
Abstract: Vector quantization, a problem rooted in Shannon's source coding theory, aims to quantize high-dimensional Euclidean vectors while minimizing distortion in their geometric structure. We propose TurboQuant to address both mean-squared error (MSE) and inner product distortion, overcoming limitations of existing methods that fail to achieve optimal distortion rates. Our data-oblivious algorithms, suitable for online applications, achieve near-optimal distortion rates (within a small constant factor) across all bit-widths and dimensions. TurboQuant achieves this by randomly rotating input vectors, inducing a concentrated Beta distribution on coordinates, and leveraging the near-independence property of distinct coordinates in high dimensions to simply apply optimal scalar quantizers per each coordinate. Recognizing that MSE-optimal quantizers introduce bias in inner product estimation, we propose a two-stage approach: applying an MSE quantizer followed by a 1-bit Quantized JL (QJL) transform on the residual, resulting in an unbiased inner product quantizer. We also provide a formal proof of the information-theoretic lower bounds on best achievable distortion rate by any vector quantizer, demonstrating that TurboQuant closely matches these bounds, differing only by a small constant ($\approx 2.7$) factor. Experimental results validate our theoretical findings, showing that for KV cache quantization, we achieve absolute quality neutrality with 3.5 bits per channel and marginal quality degradation with 2.5 bits per channel. Furthermore, in nearest neighbor search tasks, our method outperforms existing product quantization techniques in recall while reducing indexing time to virtually zero.
Authors: Hude Liu, Jerry Yao-Chieh Hu, Zhao Song, Han Liu
Abstract: We establish the universal approximation capability of single-layer, single-head self- and cross-attention mechanisms with minimal attached structures. Our key insight is to interpret single-head attention as an input domain-partition mechanism that assigns distinct values to subregions. This allows us to engineer the attention weights such that this assignment imitates the target function. Building on this, we prove that a single self-attention layer, preceded by sum-of-linear transformations, is capable of approximating any continuous function on a compact domain under the $L_\infty$-norm. Furthermore, we extend this construction to approximate any Lebesgue integrable function under $L_p$-norm for $1\leq p <\infty$. Lastly, we also extend our techniques and show that, for the first time, single-head cross-attention achieves the same universal approximation guarantees.
Authors: Diying Yang, Yingwei Hou, Danyang Xiao, Weigang Wu
Abstract: Gradient compression is an effective technique for reducing communication costs in federated learning (FL), and error feedback (EF) is usually adopted to remedy the compression errors. However, there remains a lack of systematic study on these techniques in asynchronous FL. In this paper, we fill this gap by analyzing the convergence behaviors of FL under different frameworks. We firstly consider a basic asynchronous FL framework AsynFL, and provide an improved convergence analysis that relies on fewer assumptions and yields a superior convergence rate than prior studies. Then, we consider a variant framework with gradient compression, AsynFLC. We show sufficient conditions for its convergence to the optimum, indicating the interaction between asynchronous delay and compression rate. Our analysis also demonstrates that asynchronous delay amplifies the variance caused by compression, thereby hindering convergence, and such an impact is exacerbated by high data heterogeneity. Furthermore, we study the convergence of AsynFLC-EF, the framework that further integrates EF. We prove that EF can effectively reduce the variance of gradient estimation despite asynchronous delay, which enables AsynFLC-EF to match the convergence rate of AsynFL. We also show that the impact of asynchronous delay on EF is limited to slowing down the higher-order convergence term. Experimental results substantiate our analytical findings very well.
Authors: Malhar A. Managoli, Vinod M. Prabhakaran, Suhas Diggavi
Abstract: Federated learning with heterogeneous data and personalization has received significant recent attention. Separately, robustness to corrupted data in the context of federated learning has also been studied. In this paper we explore combining personalization for heterogeneous data with robustness, where a constant fraction of the clients are corrupted. Motivated by this broad problem, we formulate a simple instantiation which captures some of its difficulty. We focus on the specific problem of personalized mean estimation where the data is drawn from a Gaussian mixture model. We give an algorithm whose error depends almost linearly on the ratio of corrupted to uncorrupted samples, and show a lower bound with the same behavior, albeit with a gap of a constant factor.
Authors: Liyuan Wang, Jiachen Chen, Kathryn L. Lunetta, Danyang Huang, Huimin Cheng, Debarghya Mukherjee
Abstract: Transfer learning enhances model performance by utilizing knowledge from related domains, particularly when labeled data is scarce. While existing research addresses transfer learning under various distribution shifts in independent settings, handling dependencies in networked data remains challenging. To address this challenge, we propose a high-dimensional transfer learning framework based on network convolutional regression (NCR), inspired by the success of graph convolutional networks (GCNs). The NCR model incorporates random network structure by allowing each node's response to depend on its features and the aggregated features of its neighbors, capturing local dependencies effectively. Our methodology includes a two-step transfer learning algorithm that addresses domain shift between source and target networks, along with a source detection mechanism to identify informative domains. Theoretically, we analyze the lasso estimator in the context of a random graph based on the Erdos-Renyi model assumption, demonstrating that transfer learning improves convergence rates when informative sources are present. Empirical evaluations, including simulations and a real-world application using Sina Weibo data, demonstrate substantial improvements in prediction accuracy, particularly when labeled data in the target domain is limited.
Authors: Adam Younsi, Abdalgader Abubaker, Mohamed El Amine Seddik, Hakim Hacid, Salem Lahlou
Abstract: Achieving both accuracy and diverse reasoning remains challenging for Large Language Models (LLMs) in complex domains like mathematics. A key bottleneck is evaluating intermediate reasoning steps to guide generation without costly human annotations. To address this, we first introduce a novel Process Reward Model (PRM) trained automatically using Monte Carlo Tree Search coupled with a similarity-based data augmentation technique, effectively capturing step-level reasoning quality. Leveraging this PRM, we then adapt Generative Flow Networks (GFlowNets) to operate at the reasoning step level. Unlike traditional reinforcement learning focused on maximizing a single reward, GFlowNets naturally sample diverse, high-quality solutions proportional to their rewards, as measured by our PRM. Empirical evaluation shows strong improvements in both accuracy and solution diversity on challenging mathematical benchmarks (e.g., +2.59% absolute accuracy on MATH Level 5 for Llama3.2-3B), with effective generalization to unseen datasets (+9.4% absolute on SAT MATH). Our work demonstrates the potential of PRM-guided, step-level GFlowNets for developing more robust and versatile mathematical reasoning in LLMs.
Authors: Yunwei Ren, Eshaan Nichani, Denny Wu, Jason D. Lee
Abstract: We study the complexity of online stochastic gradient descent (SGD) for learning a two-layer neural network with $P$ neurons on isotropic Gaussian data: $f_*(\boldsymbol{x}) = \sum_{p=1}^P a_p\cdot \sigma(\langle\boldsymbol{x},\boldsymbol{v}_p^*\rangle)$, $\boldsymbol{x} \sim \mathcal{N}(0,\boldsymbol{I}_d)$, where the activation $\sigma:\mathbb{R}\to\mathbb{R}$ is an even function with information exponent $k_*>2$ (defined as the lowest degree in the Hermite expansion), $\{\boldsymbol{v}^*_p\}_{p\in[P]}\subset \mathbb{R}^d$ are orthonormal signal directions, and the non-negative second-layer coefficients satisfy $\sum_{p} a_p^2=1$. We focus on the challenging ``extensive-width'' regime $P\gg 1$ and permit diverging condition number in the second-layer, covering as a special case the power-law scaling $a_p\asymp p^{-\beta}$ where $\beta\in\mathbb{R}_{\ge 0}$. We provide a precise analysis of SGD dynamics for the training of a student two-layer network to minimize the mean squared error (MSE) objective, and explicitly identify sharp transition times to recover each signal direction. In the power-law setting, we characterize scaling law exponents for the MSE loss with respect to the number of training samples and SGD steps, as well as the number of parameters in the student neural network. Our analysis entails that while the learning of individual teacher neurons exhibits abrupt transitions, the juxtaposition of $P\gg 1$ emergent learning curves at different timescales leads to a smooth scaling law in the cumulative objective.
Authors: Abdelhakim Amer, David Felsager, Yury Brodskiy, Andriy Sarabakha
Abstract: Physics-informed neural networks (PINNs) integrate physical laws with data-driven models to improve generalization and sample efficiency. This work introduces an open-source implementation of the Physics-Informed Neural Network with Control (PINC) framework, designed to model the dynamics of an underwater vehicle. Using initial states, control actions, and time inputs, PINC extends PINNs to enable physically consistent transitions beyond the training domain. Various PINC configurations are tested, including differing loss functions, gradient-weighting schemes, and hyperparameters. Validation on a simulated underwater vehicle demonstrates more accurate long-horizon predictions compared to a non-physics-informed baseline
Authors: Xin Wang, Haoyang Li, Zeyang Zhang, Haibo Chen, Wenwu Zhu
Abstract: Large language models (LLMs) have dramatically advanced machine learning research including natural language processing, computer vision, data mining, etc., yet they still exhibit critical limitations in reasoning, factual consistency, and interpretability. In this paper, we introduce a novel learning paradigm -- Modular Machine Learning (MML) -- as an essential approach toward new-generation LLMs. MML decomposes the complex structure of LLMs into three interdependent components: modular representation, modular model, and modular reasoning, aiming to enhance LLMs' capability of counterfactual reasoning, mitigating hallucinations, as well as promoting fairness, safety, and transparency. Specifically, the proposed MML paradigm can: i) clarify the internal working mechanism of LLMs through the disentanglement of semantic components; ii) allow for flexible and task-adaptive model design; iii) enable interpretable and logic-driven decision-making process. We present a feasible implementation of MML-based LLMs via leveraging advanced techniques such as disentangled representation learning, neural architecture search and neuro-symbolic learning. We critically identify key challenges, such as the integration of continuous neural and discrete symbolic processes, joint optimization, and computational scalability, present promising future research directions that deserve further exploration. Ultimately, the integration of the MML paradigm with LLMs has the potential to bridge the gap between statistical (deep) learning and formal (logical) reasoning, thereby paving the way for robust, adaptable, and trustworthy AI systems across a wide range of real-world applications.
Authors: Sungnyun Kim, Sungwoo Cho, Sangmin Bae, Kangwook Jang, Se-Young Yun
Abstract: Audio-visual speech recognition (AVSR) incorporates auditory and visual modalities to improve recognition accuracy, particularly in noisy environments where audio-only speech systems are insufficient. While previous research has largely addressed audio disruptions, few studies have dealt with visual corruptions, e.g., lip occlusions or blurred videos, which are also detrimental. To address this real-world challenge, we propose CAV2vec, a novel self-supervised speech representation learning framework particularly designed to handle audio-visual joint corruption. CAV2vec employs a self-distillation approach with a corrupted prediction task, where the student model learns to predict clean targets, generated by the teacher model, with corrupted input frames. Specifically, we suggest a unimodal multi-task learning, which distills cross-modal knowledge and aligns the corrupted modalities, by predicting clean audio targets with corrupted videos, and clean video targets with corrupted audios. This strategy mitigates the dispersion in the representation space caused by corrupted modalities, leading to more reliable and robust audio-visual fusion. Our experiments on robust AVSR benchmarks demonstrate that the corrupted representation learning method significantly enhances recognition accuracy across generalized environments involving various types of corruption.
Authors: Geethu Joy, Christian Huyck, Xin-She Yang
Abstract: There are many different nature-inspired algorithms in the literature, and almost all such algorithms have algorithm-dependent parameters that need to be tuned. The proper setting and parameter tuning should be carried out to maximize the performance of the algorithm under consideration. This work is the extension of the recent work on parameter tuning by Joy et al. (2024) presented at the International Conference on Computational Science (ICCS 2024), and the Firefly Algorithm (FA) is tuned using three different methods: the Monte Carlo method, the Quasi-Monte Carlo method and the Latin Hypercube Sampling. The FA with the tuned parameters is then used to solve a set of six different optimization problems, and the possible effect of parameter setting on the quality of the optimal solutions is analyzed. Rigorous statistical hypothesis tests have been carried out, including Student's t-tests, F-tests, non-parametric Friedman tests and ANOVA. Results show that the performance of the FA is not influenced by the tuning methods used. In addition, the tuned parameter values are largely independent of the tuning methods used. This indicates that the FA can be flexible and equally effective in solving optimization problems, and any of the three tuning methods can be used to tune its parameters effectively.
Authors: Harsh Patel
Abstract: Distributed Denial of Service (DDoS) attacks represent a persistent and evolving threat to modern networked systems, capable of causing large-scale service disruptions. The complexity of such attacks, often hidden within high-dimensional and redundant network traffic data, necessitates robust and intelligent feature selection techniques for effective detection. Traditional methods such as filter-based, wrapper-based, and embedded approaches, each offer strengths but struggle with scalability or adaptability in complex attack environments. In this study, we explore these existing techniques through a detailed comparative analysis and highlight their limitations when applied to large-scale DDoS detection tasks. Building upon these insights, we introduce a novel Generative Adversarial Network-based Feature Selection (GANFS) method that leverages adversarial learning dynamics to identify the most informative features. By training a GAN exclusively on attack traffic and employing a perturbation-based sensitivity analysis on the Discriminator, GANFS effectively ranks feature importance without relying on full supervision. Experimental evaluations using the CIC-DDoS2019 dataset demonstrate that GANFS not only improves the accuracy of downstream classifiers but also enhances computational efficiency by significantly reducing feature dimensionality. These results point to the potential of integrating generative learning models into cybersecurity pipelines to build more adaptive and scalable detection systems.
Authors: Guanchen Wu, Linzhi Zheng, Han Xie, Zhen Xiang, Jiaying Lu, Darren Liu, Delgersuren Bold, Bo Li, Xiao Hu, Carl Yang
Abstract: The de-identification of private information in medical data is a crucial process to mitigate the risk of confidentiality breaches, particularly when patient personal details are not adequately removed before the release of medical records. Although rule-based and learning-based methods have been proposed, they often struggle with limited generalizability and require substantial amounts of annotated data for effective performance. Recent advancements in large language models (LLMs) have shown significant promise in addressing these issues due to their superior language comprehension capabilities. However, LLMs present challenges, including potential privacy risks when using commercial LLM APIs and high computational costs for deploying open-source LLMs locally. In this work, we introduce LPPA, an LLM-empowered Privacy-Protected PHI Annotation framework for clinical notes, targeting the English language. By fine-tuning LLMs locally with synthetic notes, LPPA ensures strong privacy protection and high PHI annotation accuracy. Extensive experiments demonstrate LPPA's effectiveness in accurately de-identifying private information, offering a scalable and efficient solution for enhancing patient privacy protection.
Authors: Sabrina Bruckmeier, Huadong Mo, James Qin
Abstract: This paper presents two attack strategies designed to evade detection in ADMM-based systems by preventing significant changes to the residual during the attacked iteration. While many detection algorithms focus on identifying false data injection through residual changes, we show that our attacks remain undetected by keeping the residual largely unchanged. The first strategy uses a random starting point combined with Gram-Schmidt orthogonalization to ensure stealth, with potential for refinement by enhancing the orthogonal component to increase system disruption. The second strategy builds on the first, targeting financial gains by manipulating reactive power and pushing the system to its upper voltage limit, exploiting operational constraints. The effectiveness of the proposed attack-resilient mechanism is demonstrated through case studies on the IEEE 14-bus system. A comparison of the two strategies, along with commonly used naive attacks, reveals trade-offs between simplicity, detectability, and effectiveness, providing insights into ADMM system vulnerabilities. These findings underscore the need for more robust monitoring algorithms to protect against advanced attack strategies.
Authors: Fabio Palmese, Anna Maria Mandalari, Hamed Haddadi, Alessandro Enrico Cesare Redondi
Abstract: The rapid expansion of Internet of Things (IoT) devices, particularly in smart home environments, has introduced considerable security and privacy concerns due to their persistent connectivity and interaction with cloud services. Despite advancements in IoT security, effective privacy measures remain uncovered, with existing solutions often relying on cloud-based threat detection that exposes sensitive data or outdated allow-lists that inadequately restrict non-essential network traffic. This work presents ML-IoTrim, a system for detecting and mitigating non-essential IoT traffic (i.e., not influencing the device operations) by analyzing network behavior at the edge, leveraging Machine Learning to classify network destinations. Our approach includes building a labeled dataset based on IoT device behavior and employing a feature-extraction pipeline to enable a binary classification of essential vs. non-essential network destinations. We test our framework in a consumer smart home setup with IoT devices from five categories, demonstrating that the model can accurately identify and block non-essential traffic, including previously unseen destinations, without relying on traditional allow-lists. We implement our solution on a home access point, showing the framework has strong potential for scalable deployment, supporting near-real-time traffic classification in large-scale IoT environments with hundreds of devices. This research advances privacy-aware traffic control in smart homes, paving the way for future developments in IoT device privacy.
Authors: Anantha Sharma, Swetha Devabhaktuni, Eklove Mohan
Abstract: This paper explores the strategic use of modern synthetic data generation and advanced data perturbation techniques to enhance security, maintain analytical utility, and improve operational efficiency when managing large datasets, with a particular focus on the Banking, Financial Services, and Insurance (BFSI) sector. We contrast these advanced methods encompassing generative models like GANs, sophisticated context-aware PII transformation, configurable statistical perturbation, and differential privacy with traditional anonymization approaches. The goal is to create realistic, privacy-preserving datasets that retain high utility for complex machine learning tasks and analytics, a critical need in the data-sensitive industries like BFSI, Healthcare, Retail, and Telecommunications. We discuss how these modern techniques potentially offer significant improvements in balancing privacy preservation while maintaining data utility compared to older methods. Furthermore, we examine the potential for operational gains, such as reduced overhead and accelerated analytics, by using these privacy-enhanced datasets. We also explore key use cases where these methods can mitigate regulatory risks and enable scalable, data-driven innovation without compromising sensitive customer information.
Authors: Baimam Boukar Jean Jacques
Abstract: Classifying potentially hazardous asteroids (PHAs) is crucial for planetary defense and deep space navigation, yet traditional methods often overlook the dynamical relationships among asteroids. We introduce a Graph Neural Network (GNN) approach that models asteroids as nodes with orbital and physical features, connected by edges representing their similarities, using a NASA dataset of 958,524 records. Despite an extreme class imbalance with only 0.22% of the dataset with the hazardous label, our model achieves an overall accuracy of 99% and an AUC of 0.99, with a recall of 78% and an F1-score of 37% for hazardous asteroids after applying the Synthetic Minority Oversampling Technique. Feature importance analysis highlights albedo, perihelion distance, and semi-major axis as main predictors. This framework supports planetary defense missions and confirms AI's potential in enabling autonomous navigation for future missions such as NASA's NEO Surveyor and ESA's Ramses, offering an interpretable and scalable solution for asteroid hazard assessment.
Authors: Ali SaraerToosi, Avery Broderick
Abstract: The Event Horizon Telescope (EHT) enables the exploration of black hole accretion flows at event-horizon scales. Fitting ray-traced physical models to EHT observations requires the generation of synthetic images, a task that is computationally demanding. This study leverages \alinet, a generative machine learning model, to efficiently produce radiatively inefficient accretion flow (RIAF) images as a function of the specified physical parameters. \alinet has previously been shown to be able to interpolate black hole images and their associated physical parameters after training on a computationally tractable set of library images. We utilize this model to estimate the uncertainty introduced by a number of anticipated unmodeled physical effects, including interstellar scattering and intrinsic source variability. We then use this to calibrate physical parameter estimates and their associated uncertainties from RIAF model fits to mock EHT data via a library of general relativistic magnetohydrodynamics models.
Authors: Christodoulos Peltekis, Chrysostomos Nicopoulos, Giorgos Dimitrakopoulos
Abstract: Modern Machine Learning (ML) applications often benefit from structured sparsity, a technique that efficiently reduces model complexity and simplifies handling of sparse data in hardware. Sparse systolic tensor arrays - specifically designed to accelerate these structured-sparse ML models - play a pivotal role in enabling efficient computations. As ML is increasingly integrated into safety-critical systems, it is of paramount importance to ensure the reliability of these systems. This paper introduces an online error-checking technique capable of detecting and locating permanent faults within sparse systolic tensor arrays before computation begins. The new technique relies on merely four test vectors and exploits the weight values already loaded within the systolic array to comprehensively test the system. Fault-injection campaigns within the gate-level netlist, while executing three well-established Convolutional Neural Networks (CNN), validate the efficiency of the proposed approach, which is shown to achieve very high fault coverage, while incurring minimal performance and area overheads.
Authors: Nguyen Thi Minh Phu, Duong Tan Loc, Vo Nguyen Le Duy
Abstract: Unsupervised anomaly detection (AD) is a fundamental problem in machine learning and statistics. A popular approach to unsupervised AD is clustering-based detection. However, this method lacks the ability to guarantee the reliability of the detected anomalies. In this paper, we propose SI-CLAD (Statistical Inference for CLustering-based Anomaly Detection), a novel statistical framework for testing the clustering-based AD results. The key strength of SI-CLAD lies in its ability to rigorously control the probability of falsely identifying anomalies, maintaining it below a pre-specified significance level $\alpha$ (e.g., $\alpha = 0.05$). By analyzing the selection mechanism inherent in clustering-based AD and leveraging the Selective Inference (SI) framework, we prove that false detection control is attainable. Moreover, we introduce a strategy to boost the true detection rate, enhancing the overall performance of SI-CLAD. Extensive experiments on synthetic and real-world datasets provide strong empirical support for our theoretical findings, showcasing the superior performance of the proposed method.
Authors: Benjamin Schiffer, Lucas Janson
Abstract: Understanding how to efficiently learn while adhering to safety constraints is essential for using online reinforcement learning in practical applications. However, proving rigorous regret bounds for safety-constrained reinforcement learning is difficult due to the complex interaction between safety, exploration, and exploitation. In this work, we seek to establish foundations for safety-constrained reinforcement learning by studying the canonical problem of controlling a one-dimensional linear dynamical system with unknown dynamics. We study the safety-constrained version of this problem, where the state must with high probability stay within a safe region, and we provide the first safe algorithm that achieves regret of $\tilde{O}_T(\sqrt{T})$. Furthermore, the regret is with respect to the baseline of truncated linear controllers, a natural baseline of non-linear controllers that are well-suited for safety-constrained linear systems. In addition to introducing this new baseline, we also prove several desirable continuity properties of the optimal controller in this baseline. In showing our main result, we prove that whenever the constraints impact the optimal controller, the non-linearity of our controller class leads to a faster rate of learning than in the unconstrained setting.
Authors: Siddharth Singh, Mahua Singh, Abhinav Bhatele
Abstract: We evaluate the current state of collective communication on GPU-based supercomputers for large language model (LLM) training at scale. Existing libraries such as RCCL and Cray-MPICH exhibit critical limitations on systems such as Frontier -- Cray-MPICH underutilizes network and compute resources, while RCCL suffers from severe scalability issues. To address these challenges, we introduce PCCL, a communication library with highly optimized implementations of all-gather and reduce-scatter operations tailored for distributed deep learning workloads. PCCL is designed to maximally utilize all available network and compute resources and to scale efficiently to thousands of GPUs. It achieves substantial performance improvements, delivering 6-33x speedups over RCCL and 28-70x over Cray-MPICH for all-gather on 2048 GCDs of Frontier. These gains translate directly to end-to-end performance: in large-scale GPT-3-style training, PCCL provides up to 60% and 40% speedups over RCCL for 7B and 13B parameter models, respectively.
Authors: Apoorva Beedu, Irfan Essa
Abstract: Video summarization creates an abridged version (i.e., a summary) that provides a quick overview of the video while retaining pertinent information. In this work, we focus on summarizing instructional videos and propose a method for breaking down a video into meaningful segments, each corresponding to essential steps in the video. We propose \textbf{HierSum}, a hierarchical approach that integrates fine-grained local cues from subtitles with global contextual information provided by video-level instructions. Our approach utilizes the ``most replayed" statistic as a supervisory signal to identify critical segments, thereby improving the effectiveness of the summary. We evaluate on benchmark datasets such as TVSum, BLiSS, Mr.HiSum, and the WikiHow test set, and show that HierSum consistently outperforms existing methods in key metrics such as F1-score and rank correlation. We also curate a new multi-modal dataset using WikiHow and EHow videos and associated articles containing step-by-step instructions. Through extensive ablation studies, we demonstrate that training on this dataset significantly enhances summarization on the target datasets.
Authors: Ladan Tazik (Dept. of Computer Science, Mathematics, Physics and Statistics, University of British Columbia, Okanagan campus), James Stafford (Dept. of Statistical Sciences, University of Toronto), John Braun (Dept. of Computer Science, Mathematics, Physics and Statistics, University of British Columbia, Okanagan campus)
Abstract: The local least squares estimator for a regression curve cannot provide optimal results when non-Gaussian noise is present. Both theoretical and empirical evidence suggests that residuals often exhibit distributional properties different from those of a normal distribution, making it worthwhile to consider estimation based on other norms. It is suggested that $L_p$-norm estimators be used to minimize the residuals when these exhibit non-normal kurtosis. In this paper, we propose a local polynomial $L_p$-norm regression that replaces weighted least squares estimation with weighted $L_p$-norm estimation for fitting the polynomial locally. We also introduce a new method for estimating the parameter $p$ from the residuals, enhancing the adaptability of the approach. Through numerical and theoretical investigation, we demonstrate our method's superiority over local least squares in one-dimensional data and show promising outcomes for higher dimensions, specifically in 2D.
Authors: Ojasw Upadhyay, Abishek Saravankumar, Ayman Ismail
Abstract: Large Language Models (LLMs) are powerful but often require extensive fine-tuning and large datasets for specialized domains like law. General-purpose pre-training may not capture legal nuances, and acquiring sufficient legal data is challenging. We introduce SynLexLM, a novel approach to efficiently pre-train a legal LLM. Our method employs curriculum learning, progressing from simple to complex legal texts and queries, combined with synthetic data augmentation using models like Gemini Pro to address data scarcity. We aim to achieve improved performance on legal benchmarks (BigLaw-Bench, EUR-Lex-Sum) compared to traditional models and fine-tuned versions. Preliminary work involves generating synthetic QA pairs reflecting legal reasoning. This work aims to enhance legal document analysis and research tools, potentially democratizing access to advanced legal AI.
Authors: Manuel Weber, Carly Beneke
Abstract: We propose PyViT-FUSE, a foundation model for earth observation data explicitly designed to handle multi-modal imagery by learning to fuse an arbitrary number of mixed-resolution input bands into a single representation through an attention mechanism. The learned patch tokens are further processed by a stack of vision transformers with a novel pyramidal structure. We train the model on a globally sampled dataset in a self-supervised manner, leveraging core concepts of the SwAV algorithm. We show the interpretability of the fusion mechanism by visualization of the attention scores and the models applicability to downstream tasks.
Authors: Uday Kiran Reddy Tadipatri, Benjamin D. Haeffele, Joshua Agterberg, Ingvar Ziemann, Ren\'e Vidal
Abstract: Low-order linear System IDentification (SysID) addresses the challenge of estimating the parameters of a linear dynamical system from finite samples of observations and control inputs with minimal state representation. Traditional approaches often utilize Hankel-rank minimization, which relies on convex relaxations that can require numerous, costly singular value decompositions (SVDs) to optimize. In this work, we propose two nonconvex reformulations to tackle low-order SysID (i) Burer-Monterio (BM) factorization of the Hankel matrix for efficient nuclear norm minimization, and (ii) optimizing directly over system parameters for real, diagonalizable systems with an atomic norm style decomposition. These reformulations circumvent the need for repeated heavy SVD computations, significantly improving computational efficiency. Moreover, we prove that optimizing directly over the system parameters yields lower statistical error rates, and lower sample complexities that do not scale linearly with trajectory length like in Hankel-nuclear norm minimization. Additionally, while our proposed formulations are nonconvex, we provide theoretical guarantees of achieving global optimality in polynomial time. Finally, we demonstrate algorithms that solve these nonconvex programs and validate our theoretical claims on synthetic data.
Authors: Xiren Zhou, Shikang Liu, Xinyu Yan, Yizhan Fan, Xiangyu Wang, Yu Kang, Jian Cheng, Huanhuan Chen
Abstract: Urban roads and infrastructure, vital to city operations, face growing threats from subsurface anomalies like cracks and cavities. Ground Penetrating Radar (GPR) effectively visualizes underground conditions employing electromagnetic (EM) waves; however, accurate anomaly detection via GPR remains challenging due to limited labeled data, varying subsurface conditions, and indistinct target boundaries. Although visually image-like, GPR data fundamentally represent EM waves, with variations within and between waves critical for identifying anomalies. Addressing these, we propose the Reservoir-enhanced Segment Anything Model (Res-SAM), an innovative framework exploiting both visual discernibility and wave-changing properties of GPR data. Res-SAM initially identifies apparent candidate anomaly regions given minimal prompts, and further refines them by analyzing anomaly-induced changing information within and between EM waves in local GPR data, enabling precise and complete anomaly region extraction and category determination. Real-world experiments demonstrate that Res-SAM achieves high detection accuracy (>85%) and outperforms state-of-the-art. Notably, Res-SAM requires only minimal accessible non-target data, avoids intensive training, and incorporates simple human interaction to enhance reliability. Our research provides a scalable, resource-efficient solution for rapid subsurface anomaly detection across diverse environments, improving urban safety monitoring while reducing manual effort and computational cost.
Authors: Jong Inn Park, Maanas Taneja, Qianwen Wang, Dongyeop Kang
Abstract: Generating engaging, accurate short-form videos from scientific papers is challenging due to content complexity and the gap between expert authors and readers. Existing end-to-end methods often suffer from factual inaccuracies and visual artifacts, limiting their utility for scientific dissemination. To address these issues, we propose SciTalk, a novel multi-LLM agentic framework, grounding videos in various sources, such as text, figures, visual styles, and avatars. Inspired by content creators' workflows, SciTalk uses specialized agents for content summarization, visual scene planning, and text and layout editing, and incorporates an iterative feedback mechanism where video agents simulate user roles to give feedback on generated videos from previous iterations and refine generation prompts. Experimental evaluations show that SciTalk outperforms simple prompting methods in generating scientifically accurate and engaging content over the refined loop of video generation. Although preliminary results are still not yet matching human creators' quality, our framework provides valuable insights into the challenges and benefits of feedback-driven video generation. Our code, data, and generated videos will be publicly available.
Authors: Fran\c{c}ois-Xavier Briol, Alexandra Gessner, Toni Karvonen, Maren Mahsereci
Abstract: Kernel mean embeddings -- integrals of a kernel with respect to a probability distribution -- are essential in Bayesian quadrature, but also widely used in other computational tools for numerical integration or for statistical inference based on the maximum mean discrepancy. These methods often require, or are enhanced by, the availability of a closed-form expression for the kernel mean embedding. However, deriving such expressions can be challenging, limiting the applicability of kernel-based techniques when practitioners do not have access to a closed-form embedding. This paper addresses this limitation by providing a comprehensive dictionary of known kernel mean embeddings, along with practical tools for deriving new embeddings from known ones. We also provide a Python library that includes minimal implementations of the embeddings.
Authors: Tengfei Xing, Xiaodan Ren, Jie Li
Abstract: Stress analysis is an important part of material design. For materials with complex microstructures, such as two-phase random materials (TRMs), material failure is often accompanied by stress concentration. Phase interfaces in two-phase materials are critical for stress concentration. Therefore, the prediction error of stress at phase boundaries is crucial. In practical engineering, the pixels of the obtained material microstructure images are limited, which limits the resolution of stress images generated by deep learning methods, making it difficult to observe stress concentration regions. Existing Image Super-Resolution (ISR) technologies are all based on data-driven supervised learning. However, stress images have natural physical constraints, which provide new ideas for new ISR technologies. In this study, we constructed a stress prediction framework for TRMs. First, the framework uses a proposed Multiple Compositions U-net (MC U-net) to predict stress in low-resolution material microstructures. By considering the phase interface information of the microstructure, the MC U-net effectively reduces the problem of excessive prediction errors at phase boundaries. Secondly, a Mixed Physics-Informed Neural Network (MPINN) based method for stress ISR (SRPINN) was proposed. By introducing the constraints of physical information, the new method does not require paired stress images for training and can increase the resolution of stress images to any multiple. This enables a multiscale analysis of the stress concentration regions at phase boundaries. Finally, we performed stress analysis on TRMs with different phase volume fractions and loading states through transfer learning. The results show the proposed stress prediction framework has satisfactory accuracy and generalization ability.
Authors: Ken-Joel Simmoteit, Philipp Schillinger, Leonel Rozo
Abstract: Ensuring safety and robustness of robot skills is becoming crucial as robots are required to perform increasingly complex and dynamic tasks. The former is essential when performing tasks in cluttered environments, while the latter is relevant to overcome unseen task situations. This paper addresses the challenge of ensuring both safety and robustness in dynamic robot skills learned from demonstrations. Specifically, we build on neural contractive dynamical systems to provide robust extrapolation of the learned skills, while designing a full-body obstacle avoidance strategy that preserves contraction stability via diffeomorphic transforms. This is particularly crucial in complex environments where implicit scene representations, such as Signed Distance Fields (SDFs), are necessary. To this end, our framework called Signed Distance Field Diffeomorphic Transform, leverages SDFs and flow-based diffeomorphisms to achieve contraction-preserving obstacle avoidance. We thoroughly evaluate our framework on synthetic datasets and several real-world robotic tasks in a kitchen environment. Our results show that our approach locally adapts the learned contractive vector field while staying close to the learned dynamics and without introducing highly-curved motion paths, thus outperforming several state-of-the-art methods.
Authors: David Sychrovsk\'y, Christopher Solinas, Revan MacQueen, Kevin Wang, James R. Wright, Nathan R. Sturtevant, Michael Bowling
Abstract: Nash equilibrium is perhaps the best-known solution concept in game theory. Such a solution assigns a strategy to each player which offers no incentive to unilaterally deviate. While a Nash equilibrium is guaranteed to always exist, the problem of finding one in general-sum games is PPAD-complete, generally considered intractable. Regret minimization is an efficient framework for approximating Nash equilibria in two-player zero-sum games. However, in general-sum games, such algorithms are only guaranteed to converge to a coarse-correlated equilibrium (CCE), a solution concept where players can correlate their strategies. In this work, we use meta-learning to minimize the correlations in strategies produced by a regret minimizer. This encourages the regret minimizer to find strategies that are closer to a Nash equilibrium. The meta-learned regret minimizer is still guaranteed to converge to a CCE, but we give a bound on the distance to Nash equilibrium in terms of our meta-loss. We evaluate our approach in general-sum imperfect information games. Our algorithms provide significantly better approximations of Nash equilibria than state-of-the-art regret minimization techniques.
Authors: Alexandra Abbas, Nora Petrova, Helios Ael Lyons, Natalia Perez-Campanero
Abstract: Recent work has shown that language models' refusal behavior is primarily encoded in a single direction in their latent space, making it vulnerable to targeted attacks. Although Latent Adversarial Training (LAT) attempts to improve robustness by introducing noise during training, a key question remains: How does this noise-based training affect the underlying representation of refusal behavior? Understanding this encoding is crucial for evaluating LAT's effectiveness and limitations, just as the discovery of linear refusal directions revealed vulnerabilities in traditional supervised safety fine-tuning (SSFT). Through the analysis of Llama 2 7B, we examine how LAT reorganizes the refusal behavior in the model's latent space compared to SSFT and embedding space adversarial training (AT). By computing activation differences between harmful and harmless instruction pairs and applying Singular Value Decomposition (SVD), we find that LAT significantly alters the refusal representation, concentrating it in the first two SVD components which explain approximately 75 percent of the activation differences variance - significantly higher than in reference models. This concentrated representation leads to more effective and transferable refusal vectors for ablation attacks: LAT models show improved robustness when attacked with vectors from reference models but become more vulnerable to self-generated vectors compared to SSFT and AT. Our findings suggest that LAT's training perturbations enable a more comprehensive representation of refusal behavior, highlighting both its potential strengths and vulnerabilities for improving model safety.
Authors: Yuha Park, Kunwoong Kim, Insung Kong, Yongdai Kim
Abstract: We propose a parametric integral probability metric (IPM) to measure the discrepancy between two probability measures. The proposed IPM leverages a specific parametric family of discriminators, such as single-node neural networks with ReLU activation, to effectively distinguish between distributions, making it applicable in high-dimensional settings. By optimizing over the parameters of the chosen discriminator class, the proposed IPM demonstrates that its estimators have good convergence rates and can serve as a surrogate for other IPMs that use smooth nonparametric discriminator classes. We present an efficient algorithm for practical computation, offering a simple implementation and requiring fewer hyperparameters. Furthermore, we explore its applications in various tasks, such as covariate balancing for causal inference and fair representation learning. Across such diverse applications, we demonstrate that the proposed IPM provides strong theoretical guarantees, and empirical experiments show that it achieves comparable or even superior performance to other methods.
Authors: Cyril Shih-Huan Hsu, Anestis Dalgkitsis, Chrysa Papagianni, Paola Grosso
Abstract: In the forthcoming era of 6G networks, characterized by unprecedented data rates, ultra-low latency, and extensive connectivity, effective management of Virtualized Network Functions (VNFs) is essential. VNFs are software-based counterparts of traditional hardware devices that facilitate flexible and scalable service provisioning. Service Function Chains (SFCs), structured as ordered sequences of VNFs, are pivotal in orchestrating complex network services. Nevertheless, partitioning SFCs across multi-domain network infrastructures presents substantial challenges due to stringent latency constraints and limited resource availability. Conventional optimization-based methods typically exhibit low scalability, whereas existing data-driven approaches often fail to adequately balance computational efficiency with the capability to effectively account for dependencies inherent in SFCs. To overcome these limitations, we introduce a Transformer-empowered actor-critic framework specifically designed for sequence-aware SFC partitioning. By utilizing the self-attention mechanism, our approach effectively models complex inter-dependencies among VNFs, facilitating coordinated and parallelized decision-making processes. Additionally, we enhance training stability and convergence using $\epsilon$-LoPe exploration strategy as well as Asymptotic Return Normalization. Comprehensive simulation results demonstrate that the proposed methodology outperforms existing state-of-the-art solutions in terms of long-term acceptance rates, resource utilization efficiency, and scalability, while achieving rapid inference. This study not only advances intelligent network orchestration by delivering a scalable and robust solution for SFC partitioning within emerging 6G environments, but also bridging recent advancements in Large Language Models (LLMs) with the optimization of next-generation networks.
Authors: Benedict Leimkuhler, Ren\'e Lohmann, Peter Whalley
Abstract: We present a framework for adaptive-stepsize MCMC sampling based on time-rescaled Langevin dynamics, in which the stepsize variation is dynamically driven by an additional degree of freedom. Our approach augments the phase space by an additional variable which in turn defines a time reparameterization. The use of an auxiliary relaxation equation allows accumulation of a moving average of a local monitor function and provides for precise control of the timestep while circumventing the need to modify the drift term in the physical system. Our algorithm is straightforward to implement and can be readily combined with any off-the-peg fixed-stepsize Langevin integrator. As a particular example, we consider control of the stepsize by monitoring the norm of the log-posterior gradient, which takes inspiration from the Adam optimizer, the stepsize being automatically reduced in regions of steep change of the log posterior and increased on plateaus, improving numerical stability and convergence speed. As in Adam, the stepsize variation depends on the recent history of the gradient norm, which enhances stability and improves accuracy compared to more immediate control approaches. We demonstrate the potential benefit of this method--both in accuracy and in stability--in numerical experiments including Neal's funnel and a Bayesian neural network for classification of MNIST data.
Authors: David Sychrovsk\'y, Martin Schmid, Michal \v{S}ustr, Michael Bowling
Abstract: Regret minimization is a general approach to online optimization which plays a crucial role in many algorithms for approximating Nash equilibria in two-player zero-sum games. The literature mainly focuses on solving individual games in isolation. However, in practice, players often encounter a distribution of similar but distinct games. For example, when trading correlated assets on the stock market, or when refining the strategy in subgames of a much larger game. Recently, offline meta-learning was used to accelerate one-sided equilibrium finding on such distributions. We build upon this, extending the framework to the more challenging self-play setting, which is the basis for most state-of-the-art equilibrium approximation algorithms for domains at scale. When selecting the strategy, our method uniquely integrates information across all decision states, promoting global communication as opposed to the traditional local regret decomposition. Empirical evaluation on normal-form games and river poker subgames shows our meta-learned algorithms considerably outperform other state-of-the-art regret minimization algorithms.
Authors: Debarati Das, Khanh Chi Le, Ritik Sachin Parkar, Karin De Langis, Brendan Madson, Chad M. Berryman, Robin M. Willis, Daniel H. Moses, Brett McDonnell, Daniel Schwarcz, Dongyeop Kang
Abstract: Legal practitioners, particularly those early in their careers, face complex, high-stakes tasks that require adaptive, context-sensitive reasoning. While AI holds promise in supporting legal work, current datasets and models are narrowly focused on isolated subtasks and fail to capture the end-to-end decision-making required in real-world practice. To address this gap, we introduce LawFlow, a dataset of complete end-to-end legal workflows collected from trained law students, grounded in real-world business entity formation scenarios. Unlike prior datasets focused on input-output pairs or linear chains of thought, LawFlow captures dynamic, modular, and iterative reasoning processes that reflect the ambiguity, revision, and client-adaptive strategies of legal practice. Using LawFlow, we compare human and LLM-generated workflows, revealing systematic differences in structure, reasoning flexibility, and plan execution. Human workflows tend to be modular and adaptive, while LLM workflows are more sequential, exhaustive, and less sensitive to downstream implications. Our findings also suggest that legal professionals prefer AI to carry out supportive roles, such as brainstorming, identifying blind spots, and surfacing alternatives, rather than executing complex workflows end-to-end. Building on these findings, we propose a set of design suggestions, rooted in empirical observations, that align AI assistance with human goals of clarity, completeness, creativity, and efficiency, through hybrid planning, adaptive execution, and decision-point support. Our results highlight both the current limitations of LLMs in supporting complex legal workflows and opportunities for developing more collaborative, reasoning-aware legal AI systems. All data and code are available on our project page (https://minnesotanlp.github.io/LawFlow-website/).
Authors: Erfan Loweimi, Mengjie Qian, Kate Knill, Mark Gales
Abstract: There is a growing abundance of publicly available or company-owned audio/video archives, highlighting the increasing importance of efficient access to desired content and information retrieval from these archives. This paper investigates the challenges, solutions, effectiveness, and robustness of speaker retrieval systems developed "in the wild" which involves addressing two primary challenges: extraction of task-relevant labels from limited metadata for system development and evaluation, as well as the unconstrained acoustic conditions encountered in the archive, ranging from quiet studios to adverse noisy environments. While we focus on the publicly-available BBC Rewind archive (spanning 1948 to 1979), our framework addresses the broader issue of speaker retrieval on extensive and possibly aged archives with no control over the content and acoustic conditions. Typically, these archives offer a brief and general file description, mostly inadequate for specific applications like speaker retrieval, and manual annotation of such large-scale archives is unfeasible. We explore various aspects of system development (e.g., speaker diarisation, embedding extraction, query selection) and analyse the challenges, possible solutions, and their functionality. To evaluate the performance, we conduct systematic experiments in both clean setup and against various distortions simulating real-world applications. Our findings demonstrate the effectiveness and robustness of the developed speaker retrieval systems, establishing the versatility and scalability of the proposed framework for a wide range of applications beyond the BBC Rewind corpus.
Authors: Masoud Ataei
Abstract: This paper investigates the structural dynamics of stock market volatility through the Financial Chaos Index, a tensor- and eigenvalue-based measure designed to capture realized volatility via mutual fluctuations among asset prices. Motivated by empirical evidence of regime-dependent volatility behavior and perceptual time dilation during financial crises, we develop a regime-switching framework based on the Modified Lognormal Power-Law distribution. Analysis of the FCIX from January 1990 to December 2023 identifies three distinct market regimes, low-chaos, intermediate-chaos, and high-chaos, each characterized by differing levels of systemic stress, statistical dispersion and persistence characteristics. Building upon the segmented regime structure, we further examine the informational forces that shape forward-looking market expectations. Using sentiment-based predictors derived from the Equity Market Volatility tracker, we employ an elastic net regression model to forecast implied volatility, as proxied by the VIX index. Our findings indicate that shifts in macroeconomic, financial, policy, and geopolitical uncertainty exhibit strong predictive power for volatility dynamics across regimes. Together, these results offer a unified empirical perspective on how systemic uncertainty governs both the realized evolution of financial markets and the anticipatory behavior embedded in implied volatility measures.
Authors: Gal Almog, Ariel Shamir, Ohad Fried
Abstract: While latent diffusion models achieve impressive image editing results, their application to iterative editing of the same image is severely restricted. When trying to apply consecutive edit operations using current models, they accumulate artifacts and noise due to repeated transitions between pixel and latent spaces. Some methods have attempted to address this limitation by performing the entire edit chain within the latent space, sacrificing flexibility by supporting only a limited, predetermined set of diffusion editing operations. We present a RE-encode decode (REED) training scheme for variational autoencoders (VAEs), which promotes image quality preservation even after many iterations. Our work enables multi-method iterative image editing: users can perform a variety of iterative edit operations, with each operation building on the output of the previous one using both diffusion-based operations and conventional editing techniques. We demonstrate the advantage of REED-VAE across a range of image editing scenarios, including text-based and mask-based editing frameworks. In addition, we show how REED-VAE enhances the overall editability of images, increasing the likelihood of successful and precise edit operations. We hope that this work will serve as a benchmark for the newly introduced task of multi-method image editing. Our code and models will be available at https://github.com/galmog/REED-VAE
Authors: Jinghao Lyu, Kyle J. Ray, James P. Crutchfield
Abstract: Markedly increased computational power and data acquisition have led to growing interest in data-driven inverse dynamics problems. These seek to answer a fundamental question: What can we learn from time series measurements of a complex dynamical system? For small systems interacting with external environments, the effective dynamics are inherently stochastic, making it crucial to properly manage noise in data. Here, we explore this for systems obeying Langevin dynamics and, using currents, we construct a learning framework for stochastic modeling. Currents have recently gained increased attention for their role in bounding entropy production (EP) from thermodynamic uncertainty relations (TURs). We introduce a fundamental relationship between the cumulant currents there and standard machine-learning loss functions. Using this, we derive loss functions for several key thermodynamic functions directly from the system dynamics without the (common) intermediate step of deriving a TUR. These loss functions reproduce results derived both from TURs and other methods. More significantly, they open a path to discover new loss functions for previously inaccessible quantities. Notably, this includes access to per-trajectory entropy production, even if the observed system is driven far from its steady-state. We also consider higher order estimation. Our method is straightforward and unifies dynamic inference with recent approaches to entropy production estimation. Taken altogether, this reveals a deep connection between diffusion models in machine learning and entropy production estimation in stochastic thermodynamics.
Authors: Xizhuo (Cici), Zhang, Bing Yao
Abstract: Rapid developments in advanced sensing and imaging have significantly enhanced information visibility, opening opportunities for predictive modeling of complex dynamic systems. However, sensing signals acquired from such complex systems are often distributed across 3D geometries and rapidly evolving over time, posing significant challenges in spatiotemporal predictive modeling. This paper proposes a geometry-aware active learning framework for modeling spatiotemporal dynamic systems. Specifically, we propose a geometry-aware spatiotemporal Gaussian Process (G-ST-GP) to effectively integrate the temporal correlations and geometric manifold features for reliable prediction of high-dimensional dynamic behaviors. In addition, we develop an adaptive active learning strategy to strategically identify informative spatial locations for data collection and further maximize the prediction accuracy. This strategy achieves the adaptive trade-off between the prediction uncertainty in the G-ST-GP model and the space-filling design guided by the geodesic distance across the 3D geometry. We implement the proposed framework to model the spatiotemporal electrodynamics in a 3D heart geometry. Numerical experiments show that our framework outperforms traditional methods lacking the mechanism of geometric information incorporation or effective data collection.
Authors: Alireza Ghafarollahi, Markus J. Buehler
Abstract: Advances in artificial intelligence (AI) promise autonomous discovery, yet most systems still resurface knowledge latent in their training data. We present Sparks, a multi-modal multi-agent AI model that executes the entire discovery cycle that includes hypothesis generation, experiment design and iterative refinement to develop generalizable principles and a report without human intervention. Applied to protein science, Sparks uncovered two previously unknown phenomena: (i) a length-dependent mechanical crossover whereby beta-sheet-biased peptides surpass alpha-helical ones in unfolding force beyond ~80 residues, establishing a new design principle for peptide mechanics; and (ii) a chain-length/secondary-structure stability map revealing unexpectedly robust beta-sheet-rich architectures and a "frustration zone" of high variance in mixed alpha/beta folds. These findings emerged from fully self-directed reasoning cycles that combined generative sequence design, high-accuracy structure prediction and physics-aware property models, with paired generation-and-reflection agents enforcing self-correction and reproducibility. The key result is that Sparks can independently conduct rigorous scientific inquiry and identify previously unknown scientific principles.
Authors: Justin M\"ucke, Ansgar Scherp
Abstract: Semantic reasoning aims to infer new knowledge from existing knowledge, with OWL ontologies serving as a standardized framework for organizing information. A key challenge in semantic reasoning is verifying ontology consistency. However, state-of-the-art reasoners are computationally expensive, and their efficiency decreases as ontology sizes grow. While classical machine learning models have been explored for consistency checking, they struggle to capture complex relationships within ontologies. Large language models (LLMs) have shown promising results for simple reasoning tasks but perform poorly on structured reasoning. The recently introduced Graph Language Model (GLM) offers a way to simultaneously process graph-structured data and text. This paper proposes GLaMoR (Graph Language Model for Reasoning), a reasoning pipeline that transforms OWL ontologies into graph-structured data and adapts the GLM architecture for consistency checking. We evaluate GLaMoR on ontologies from the NCBO BioPortal repository, converting them into triples suitable for model input. Our results show that the GLM outperforms all baseline models, achieving $95\%$ accuracy while being 20 times faster than classical reasoners. The Code is accessible under: https://github.com/JustinMuecke/GLaMoR
Authors: Volkan Bakir, Polat Goktas, Sureyya Akyuz
Abstract: Explainable artificial intelligence (XAI) has become increasingly important in decision-critical domains such as healthcare, finance, and law. Counterfactual (CF) explanations, a key approach in XAI, provide users with actionable insights by suggesting minimal modifications to input features that lead to different model outcomes. Despite significant advancements, existing CF generation methods often struggle to balance proximity, diversity, and robustness, limiting their real-world applicability. A widely adopted framework, Diverse Counterfactual Explanations (DiCE), emphasizes diversity but lacks robustness, making CF explanations sensitive to perturbations and domain constraints. To address these challenges, we introduce DiCE-Extended, an enhanced CF explanation framework that integrates multi-objective optimization techniques to improve robustness while maintaining interpretability. Our approach introduces a novel robustness metric based on the Dice-Sorensen coefficient, ensuring stability under small input variations. Additionally, we refine CF generation using weighted loss components (lambda_p, lambda_d, lambda_r) to balance proximity, diversity, and robustness. We empirically validate DiCE-Extended on benchmark datasets (COMPAS, Lending Club, German Credit, Adult Income) across multiple ML backends (Scikit-learn, PyTorch, TensorFlow). Results demonstrate improved CF validity, stability, and alignment with decision boundaries compared to standard DiCE-generated explanations. Our findings highlight the potential of DiCE-Extended in generating more reliable and interpretable CFs for high-stakes applications. Future work will explore adaptive optimization techniques and domain-specific constraints to further enhance CF generation in real-world scenarios.
Authors: James Giroux, Michael Martinez, Cristiano Fanelli
Abstract: The integration of Deep Learning (DL) into experimental nuclear and particle physics has driven significant progress in simulation and reconstruction workflows. However, traditional simulation frameworks such as Geant4 remain computationally intensive, especially for Cherenkov detectors, where simulating optical photon transport through complex geometries and reflective surfaces introduces a major bottleneck. To address this, we present an open, standalone fast simulation tool for Detection of Internally Reflected Cherenkov Light (DIRC) detectors, with a focus on the High-Performance DIRC (hpDIRC) at the future Electron-Ion Collider (EIC). Our framework incorporates a suite of generative models tailored to accelerate particle identification (PID) tasks by offering a scalable, GPU-accelerated alternative to full Geant4-based simulations. Designed with accessibility in mind, our simulation package enables both DL researchers and physicists to efficiently generate high-fidelity large-scale datasets on demand, without relying on complex traditional simulation stacks. This flexibility supports the development and benchmarking of novel DL-driven PID methods. Moreover, this fast simulation pipeline represents a critical step toward enabling EIC-wide PID strategies that depend on virtually unlimited simulated samples, spanning the full acceptance of the hpDIRC.
Authors: Hongni Jin, Gurinder Singh, Kenneth M. Merz Jr
Abstract: Implicit neural representations have shown potential in various applications. However, accurately reconstructing the image or providing clear details via image super-resolution remains challenging. This paper introduces Quantum Fourier Gaussian Network (QFGN), a quantum-based machine learning model for better signal representations. The frequency spectrum is well balanced by penalizing the low-frequency components, leading to the improved expressivity of quantum circuits. The results demonstrate that with minimal parameters, QFGN outperforms the current state-of-the-art (SOTA) models. Despite noise on hardware, the model achieves accuracy comparable to that of SIREN, highlighting the potential applications of quantum machine learning in this field.
Authors: Mohammad Mahdi Abootorabi, Omid Ghahroodi, Pardis Sadat Zahraei, Hossein Behzadasl, Alireza Mirrokni, Mobina Salimipanah, Arash Rasouli, Bahar Behzadipour, Sara Azarnoush, Benyamin Maleki, Erfan Sadraiye, Kiarash Kiani Feriz, Mahdi Teymouri Nahad, Ali Moghadasi, Abolfazl Eshagh Abianeh, Nizi Nazar, Hamid R. Rabiee, Mahdieh Soleymani Baghshah, Meisam Ahmadi, Ehsaneddin Asgari
Abstract: Generative AI is reshaping art, gaming, and most notably animation. Recent breakthroughs in foundation and diffusion models have reduced the time and cost of producing animated content. Characters are central animation components, involving motion, emotions, gestures, and facial expressions. The pace and breadth of advances in recent months make it difficult to maintain a coherent view of the field, motivating the need for an integrative review. Unlike earlier overviews that treat avatars, gestures, or facial animation in isolation, this survey offers a single, comprehensive perspective on all the main generative AI applications for character animation. We begin by examining the state-of-the-art in facial animation, expression rendering, image synthesis, avatar creation, gesture modeling, motion synthesis, object generation, and texture synthesis. We highlight leading research, practical deployments, commonly used datasets, and emerging trends for each area. To support newcomers, we also provide a comprehensive background section that introduces foundational models and evaluation metrics, equipping readers with the knowledge needed to enter the field. We discuss open challenges and map future research directions, providing a roadmap to advance AI-driven character-animation technologies. This survey is intended as a resource for researchers and developers entering the field of generative AI animation or adjacent fields. Resources are available at: https://github.com/llm-lab-org/Generative-AI-for-Character-Animation-Survey.
URLs: https://github.com/llm-lab-org/Generative-AI-for-Character-Animation-Survey.
Authors: Deeksha Varshney, Keane Ong, Rui Mao, Erik Cambria, Gianmarco Mengaldo
Abstract: Accurate assessments of extreme weather events are vital for research and policy, yet localized and granular data remain scarce in many parts of the world. This data gap limits our ability to analyze potential outcomes and implications of extreme weather events, hindering effective decision-making. Large Language Models (LLMs) can process vast amounts of unstructured text data, extract meaningful insights, and generate detailed assessments by synthesizing information from multiple sources. Furthermore, LLMs can seamlessly transfer their general language understanding to smaller models, enabling these models to retain key knowledge while being fine-tuned for specific tasks. In this paper, we propose Extreme Weather Reasoning-Aware Alignment (EWRA), a method that enhances small language models (SLMs) by incorporating structured reasoning paths derived from LLMs, and ExtremeWeatherNews, a large dataset of extreme weather event-related news articles. EWRA and ExtremeWeatherNews together form the overall framework, ClimaEmpact, that focuses on addressing three critical extreme-weather tasks: categorization of tangible vulnerabilities/impacts, topic labeling, and emotion analysis. By aligning SLMs with advanced reasoning strategies on ExtremeWeatherNews (and its derived dataset ExtremeAlign used specifically for SLM alignment), EWRA improves the SLMs' ability to generate well-grounded and domain-specific responses for extreme weather analytics. Our results show that the approach proposed guides SLMs to output domain-aligned responses, surpassing the performance of task-specific models and offering enhanced real-world applicability for extreme weather analytics.
Authors: Anyong Qin, Chaoqi Yuan, Qiang Li, Feng Yang, Tiecheng Song, Chenqiang Gao
Abstract: Convolutional neural networks (CNNs) are effective for hyperspectral image (HSI) classification, but their 3D convolutional structures introduce high computational costs and limited generalization in few-shot scenarios. Domain shifts caused by sensor differences and environmental variations further hinder cross-dataset adaptability. Metric-based few-shot learning (FSL) prototype networks mitigate this problem, yet their performance is sensitive to prototype quality, especially with limited samples. To overcome these challenges, a dual-branch residual network that integrates spatial and spectral features via parallel branches is proposed in this letter. Additionally, more robust refined prototypes are obtained through a regulation term. Furthermore, a kernel probability matching strategy aligns source and target domain features, alleviating domain shift. Experiments on four publicly available HSI datasets illustrate that the proposal achieves superior performance compared to other methods.
Authors: Satwik Kundu, Swaroop Ghosh
Abstract: Circuit compilation, a crucial process for adapting quantum algorithms to hardware constraints, often operates as a ``black box,'' with limited visibility into the optimization techniques used by proprietary systems or advanced open-source frameworks. Due to fundamental differences in qubit technologies, efficient compiler design is an expensive process, further exposing these systems to various security threats. In this work, we take a first step toward evaluating one such challenge affecting compiler confidentiality, specifically, reverse-engineering compilation methodologies. We propose a simple ML-based framework to infer underlying optimization techniques by leveraging structural differences observed between original and compiled circuits. The motivation is twofold: (1) enhancing transparency in circuit optimization for improved cross-platform debugging and performance tuning, and (2) identifying potential intellectual property (IP)-protected optimizations employed by commercial systems. Our extensive evaluation across thousands of quantum circuits shows that a neural network performs the best in detecting optimization passes, with individual pass F1-scores reaching as high as 0.96. Thus, our initial study demonstrates the viability of this threat to compiler confidentiality and underscores the need for active research in this area.
Authors: Abhishek Pasula, Deepak N. Subramani
Abstract: Climate change affects ocean temperature, salinity and sea level, impacting monsoons and ocean productivity. Future projections by Global Climate Models based on shared socioeconomic pathways from the Coupled Model Intercomparison Project (CMIP) are widely used to understand the effects of climate change. However, CMIP models have significant bias compared to reanalysis in the Bay of Bengal for the time period when both projections and reanalysis are available. For example, there is a 1.5C root mean square error (RMSE) in the sea surface temperature (SST) projections of the climate model CNRM-CM6 compared to the Ocean Reanalysis System (ORAS5). We develop a suite of data-driven deep learning models for bias correction of climate model projections and apply it to correct SST projections of the Bay of Bengal. We propose the use of three different deep neural network architectures: convolutional encoder-decoder UNet, Bidirectional LSTM and ConvLSTM. We also use a baseline linear regression model and the Equi-Distant Cumulative Density Function (EDCDF) bias correction method for comparison and evaluating the impact of the new deep learning models. All bias correction models are trained using pairs of monthly CMIP6 projections and the corresponding month's ORAS5 as input and output. Historical data (1950-2014) and future projection data (2015-2020) of CNRM-CM6 are used for training and validation, including hyperparameter tuning. Testing is performed on future projection data from 2021 to 2024. Detailed analysis of the three deep neural models has been completed. We found that the UNet architecture trained using a climatology-removed CNRM-CM6 projection as input and climatology-removed ORAS5 as output gives the best bias-corrected projections. Our novel deep learning-based method for correcting CNRM-CM6 data has a 15% reduction in RMSE compared EDCDF.
Authors: Jiaqi Chen, Bang Zhang, Ruotian Ma, Peisong Wang, Xiaodan Liang, Zhaopeng Tu, Xiaolong Li, Kwan-Yee K. Wong
Abstract: Evaluating the step-by-step reliability of large language model (LLM) reasoning, such as Chain-of-Thought, remains challenging due to the difficulty and cost of obtaining high-quality step-level supervision. In this paper, we introduce Self-Play Critic (SPC), a novel approach where a critic model evolves its ability to assess reasoning steps through adversarial self-play games, eliminating the need for manual step-level annotation. SPC involves fine-tuning two copies of a base model to play two roles, namely a "sneaky generator" that deliberately produces erroneous steps designed to be difficult to detect, and a "critic" that analyzes the correctness of reasoning steps. These two models engage in an adversarial game in which the generator aims to fool the critic, while the critic model seeks to identify the generator's errors. Using reinforcement learning based on the game outcomes, the models iteratively improve; the winner of each confrontation receives a positive reward and the loser receives a negative reward, driving continuous self-evolution. Experiments on three reasoning process benchmarks (ProcessBench, PRM800K, DeltaBench) demonstrate that our SPC progressively enhances its error detection capabilities (e.g., accuracy increases from 70.8% to 77.7% on ProcessBench) and surpasses strong baselines, including distilled R1 model. Furthermore, applying SPC to guide the test-time search of diverse LLMs significantly improves their mathematical reasoning performance on MATH500 and AIME2024, outperforming state-of-the-art process reward models.
Authors: Elisabeth Fittschen, Bella Xia, Leib Celnik, Paul Dilley, Tom Lippincott
Abstract: We measure the effects of several implementation choices for the Dynamic Embedded Topic Model, as applied to five distinct diachronic corpora, with the goal of isolating important decisions for its use and further development. We identify priorities that will maximize utility in applied scholarship, including the practical scalability of vocabulary size to best exploit the strengths of embedded representations, and more flexible modeling of intervals to accommodate the uneven temporal distributions of historical writing. Of similar importance, we find performance is not significantly or consistently affected by several aspects that otherwise limit the model's application or might consume the resources of a grid search.
Authors: Alexander Baumann, Leonardo Ayala, Silvia Seidlitz, Jan Sellner, Alexander Studier-Fischer, Berkin \"Ozdemir, Lena Maier-Hein, Slobodan Ilic
Abstract: Spectral imaging offers promising applications across diverse domains, including medicine and urban scene understanding, and is already established as a critical modality in remote sensing. However, variability in channel dimensionality and captured wavelengths among spectral cameras impede the development of AI-driven methodologies, leading to camera-specific models with limited generalizability and inadequate cross-camera applicability. To address this bottleneck, we introduce $\textbf{CARL}$, a model for $\textbf{C}$amera-$\textbf{A}$gnostic $\textbf{R}$epresentation $\textbf{L}$earning across RGB, multispectral, and hyperspectral imaging modalities. To enable the conversion of a spectral image with any channel dimensionality to a camera-agnostic embedding, we introduce wavelength positional encoding and a self-attention-cross-attention mechanism to compress spectral information into learned query representations. Spectral-spatial pre-training is achieved with a novel spectral self-supervised JEPA-inspired strategy tailored to CARL. Large-scale experiments across the domains of medical imaging, autonomous driving, and satellite imaging demonstrate our model's unique robustness to spectral heterogeneity, outperforming on datasets with simulated and real-world cross-camera spectral variations. The scalability and versatility of the proposed approach position our model as a backbone for future spectral foundation models.
Authors: Alexander Dubbs
Abstract: We derive the ideal train/test split for the ridge regression to high accuracy in the limit that the number of training rows m becomes large. The split must depend on the ridge tuning parameter, alpha, but we find that the dependence is weak and can asymptotically be ignored; all parameters vanish except for m and the number of features, n. This is the first time that such a split is calculated mathematically for a machine learning model in the large data limit. The goal of the calculations is to maximize "integrity," so that the measured error in the trained model is as close as possible to what it theoretically should be. This paper's result for the ridge regression split matches prior art for the plain vanilla linear regression split to the first two terms asymptotically, and it appears that practically there is no difference.
Authors: Yoshiaki Kawase
Abstract: Quantum neural networks hold promise for tackling computationally challenging tasks that are intractable for classical computers. However, their practical application is hindered by significant optimization challenges, arising from complex loss landscapes characterized by barren plateaus and numerous local minima. These problems become more severe as the number of parameters or qubits increases, hampering effective training. To mitigate these optimization challenges, particularly for quantum machine learning applied to classical data, we employ an approach of distributing overlapping local patches across multiple quantum neural networks, processing each patch with an independent quantum neural network, and aggregating their outputs for prediction. In this study, we investigate how the number of parameters and patches affects the loss landscape geometry of this distributed quantum neural network architecture via Hessian analysis and loss landscape visualization. Our results confirm that increasing the number of parameters tends to lead to deeper and sharper loss landscapes. Crucially, we demonstrate that increasing the number of patches significantly reduces the largest Hessian eigenvalue at minima. This finding suggests that our distributed patch approach acts as a form of implicit regularization, promoting optimization stability and potentially enhancing generalization. Our study provides valuable insights into optimization challenges and highlights that the distributed patch approach is a promising strategy for developing more trainable and practical quantum machine learning models for classical data tasks.
Authors: Dylan Bouchard, Mohit Singh Chauhan
Abstract: Hallucinations are a persistent problem with Large Language Models (LLMs). As these models become increasingly used in high-stakes domains, such as healthcare and finance, the need for effective hallucination detection is crucial. To this end, we propose a versatile framework for zero-resource hallucination detection that practitioners can apply to real-world use cases. To achieve this, we adapt a variety of existing uncertainty quantification (UQ) techniques, including black-box UQ, white-box UQ, and LLM-as-a-Judge, transforming them as necessary into standardized response-level confidence scores ranging from 0 to 1. To enhance flexibility, we introduce a tunable ensemble approach that incorporates any combination of the individual confidence scores. This approach enables practitioners to optimize the ensemble for a specific use case for improved performance. To streamline implementation, the full suite of scorers is offered in this paper's companion Python toolkit, UQLM. To evaluate the performance of the various scorers, we conduct an extensive set of experiments using several LLM question-answering benchmarks. We find that our tunable ensemble typically surpasses its individual components and outperforms existing hallucination detection methods. Our results demonstrate the benefits of customized hallucination detection strategies for improving the accuracy and reliability of LLMs.
Authors: Angel Mary John, Jerrin Thomas Panachakel, Anusha S. P
Abstract: This paper explores the integration of human rights considerations into AI regulatory frameworks across different IEEE regions - specifically the United States (Region 1-6), Europe (Region 8), China (part of Region 10), and Singapore (part of Region 10). While all acknowledge the transformative potential of AI and the necessity of ethical guidelines, their regulatory approaches significantly differ. Europe exhibits a rigorous framework with stringent protections for individual rights, while the U.S. promotes innovation with less restrictive regulations. China emphasizes state control and societal order in its AI strategies. In contrast, Singapore's advisory framework encourages self-regulation and aligns closely with international norms. This comparative analysis underlines the need for ongoing global dialogue to harmonize AI regulations that safeguard human rights while promoting technological advancement, reflecting the diverse perspectives and priorities of each region.
Authors: Mohamed Gado, Towhid Taliee, Muhammad Memon, Dmitry Ignatov, Radu Timofte
Abstract: Visual storytelling is an interdisciplinary field combining computer vision and natural language processing to generate cohesive narratives from sequences of images. This paper presents a novel approach that leverages recent advancements in multimodal models, specifically adapting transformer-based architectures and large multimodal models, for the visual storytelling task. Leveraging the large-scale Visual Storytelling (VIST) dataset, our VIST-GPT model produces visually grounded, contextually appropriate narratives. We address the limitations of traditional evaluation metrics, such as BLEU, METEOR, ROUGE, and CIDEr, which are not suitable for this task. Instead, we utilize RoViST and GROOVIST, novel reference-free metrics designed to assess visual storytelling, focusing on visual grounding, coherence, and non-redundancy. These metrics provide a more nuanced evaluation of narrative quality, aligning closely with human judgment.
Authors: Hanchen Yang, Zishen Wan, Ritik Raj, Joongun Park, Ziwei Li, Ananda Samajdar, Arijit Raychowdhury, Tushar Krishna
Abstract: Neuro-Symbolic AI (NSAI) is an emerging paradigm that integrates neural networks with symbolic reasoning to enhance the transparency, reasoning capabilities, and data efficiency of AI systems. Recent NSAI systems have gained traction due to their exceptional performance in reasoning tasks and human-AI collaborative scenarios. Despite these algorithmic advancements, executing NSAI tasks on existing hardware (e.g., CPUs, GPUs, TPUs) remains challenging, due to their heterogeneous computing kernels, high memory intensity, and unique memory access patterns. Moreover, current NSAI algorithms exhibit significant variation in operation types and scales, making them incompatible with existing ML accelerators. These challenges highlight the need for a versatile and flexible acceleration framework tailored to NSAI workloads. In this paper, we propose NSFlow, an FPGA-based acceleration framework designed to achieve high efficiency, scalability, and versatility across NSAI systems. NSFlow features a design architecture generator that identifies workload data dependencies and creates optimized dataflow architectures, as well as a reconfigurable array with flexible compute units, re-organizable memory, and mixed-precision capabilities. Evaluating across NSAI workloads, NSFlow achieves 31x speedup over Jetson TX2, more than 2x over GPU, 8x speedup over TPU-like systolic array, and more than 3x over Xilinx DPU. NSFlow also demonstrates enhanced scalability, with only 4x runtime increase when symbolic workloads scale by 150x. To the best of our knowledge, NSFlow is the first framework to enable real-time generalizable NSAI algorithms acceleration, demonstrating a promising solution for next-generation cognitive systems.
Authors: James O' Neill, Santhosh Subramanian, Eric Lin, Vaikkunth Mugunthan
Abstract: The trend towards large language models (LLMs) for guardrailing against undesired behaviors is increasing and has shown promise for censoring user inputs. However, increased latency, memory consumption, hosting expenses and non-structured outputs can make their use prohibitive. In this work, we show that task-specific data generation can lead to fine-tuned classifiers that significantly outperform current state of the art (SoTA) while being orders of magnitude smaller. Secondly, we show that using a single model, \texttt{MultiTaskGuard}, that is pretrained on a large synthetically generated dataset with unique task instructions further improves generalization. Thirdly, our most performant models, \texttt{UniGuard}, are found using our proposed search-based model merging approach that finds an optimal set of parameters to combine single-policy models and multi-policy guardrail models. % On 7 public datasets and 4 guardrail benchmarks we created, our efficient guardrail classifiers improve over the best performing SoTA publicly available LLMs and 3$^{\text{rd}}$ party guardrail APIs in detecting unsafe and safe behaviors by an average F1 score improvement of \textbf{29.92} points over Aegis-LlamaGuard and \textbf{21.62} over \texttt{gpt-4o}, respectively. Lastly, our guardrail synthetic data generation process that uses custom task-specific guardrail poli
Authors: Dongqi Liu, Xi Yu, Vera Demberg, Mirella Lapata
Abstract: Lay summaries for scientific documents typically include explanations to help readers grasp sophisticated concepts or arguments. However, current automatic summarization methods do not explicitly model explanations, which makes it difficult to align the proportion of explanatory content with human-written summaries. In this paper, we present a plan-based approach that leverages discourse frameworks to organize summary generation and guide explanatory sentences by prompting responses to the plan. Specifically, we propose two discourse-driven planning strategies, where the plan is conditioned as part of the input or part of the output prefix, respectively. Empirical experiments on three lay summarization datasets show that our approach outperforms existing state-of-the-art methods in terms of summary quality, and it enhances model robustness, controllability, and mitigates hallucination.
Authors: Nan Lu, Ethan X. Fang, Junwei Lu
Abstract: Reinforcement Learning from Human Feedback (RLHF) has become a pivotal paradigm in artificial intelligence to align large models with human preferences. In this paper, we propose a novel statistical framework to simultaneously conduct the online decision-making and statistical inference on the optimal model using human preference data based on dynamic contextual information. Our approach introduces an efficient decision strategy that achieves both the optimal regret bound and the asymptotic distribution of the estimators. A key challenge in RLHF is handling the dependent online human preference outcomes with dynamic contexts. To address this, in the methodological aspect, we propose a two-stage algorithm starting with $\epsilon$-greedy followed by exploitations; in the theoretical aspect, we tailor anti-concentration inequalities and matrix martingale concentration techniques to derive the uniform estimation rate and asymptotic normality of the estimators using dependent samples from both stages. Extensive simulation results demonstrate that our method outperforms state-of-the-art strategies. We apply the proposed framework to analyze the human preference data for ranking large language models on the Massive Multitask Language Understanding dataset, yielding insightful results on the performance of different large language models for medical anatomy knowledge.
Authors: Chathurika S Abeykoon, Aleksandr Beknazaryan, Hailin Sang
Abstract: Recent studies observed a surprising concept on model test error called the double descent phenomenon, where the increasing model complexity decreases the test error first and then the error increases and decreases again. To observe this, we work on a two layer neural network model with a ReLU activation function designed for binary classification under supervised learning. Our aim is to observe and investigate the mathematical theory behind the double descent behavior of model test error for varying model sizes. We quantify the model size by the ratio of number of training samples to the dimension of the model. Due to the complexity of the empirical risk minimization procedure, we use the Convex Gaussian Min Max Theorem to find a suitable candidate for the global training loss.
Authors: Erkan Karabulut, Paul Groth, Victoria Degeler
Abstract: Association Rule Mining (ARM) is the task of mining patterns among data features in the form of logical rules, with applications across a myriad of domains. However, high-dimensional datasets often result in an excessive number of rules, increasing execution time and negatively impacting downstream task performance. Managing this rule explosion remains a central challenge in ARM research. To address this, we introduce Aerial+, a novel neurosymbolic ARM method. Aerial+ leverages an under-complete autoencoder to create a neural representation of the data, capturing associations between features. It extracts rules from this neural representation by exploiting the model's reconstruction mechanism. Extensive evaluations on five datasets against seven baselines demonstrate that Aerial+ achieves state-of-the-art results by learning more concise, high-quality rule sets with full data coverage. When integrated into rule-based interpretable machine learning models, Aerial+ significantly reduces execution time while maintaining or improving accuracy.
Authors: Gionni Marchetti
Abstract: We present a machine learning analysis of circular dichroism spectra of globular proteins from the SP175 database, using the optimal transport-based $1$-Wasserstein distance $\mathcal{W}_1$ (with order $p=1$) and the manifold learning algorithm $t$-SNE. Our results demonstrate that $\mathcal{W}_1$ is consistent with both Euclidean and Manhattan metrics while exhibiting robustness to noise. On the other hand, $t$-SNE uncovers meaningful structure in the high-dimensional data. The clustering in the $t$-SNE embedding is primarily determined by proteins with distinct secondary structure compositions: one cluster predominantly contains $\beta$-rich proteins, while the other consists mainly of proteins with mixed $\alpha/\beta$ and $\alpha$-helical content.
Authors: Jean-R\'emy Conti, St\'ephan Cl\'emen\c{c}on
Abstract: The urging societal demand for fair AI systems has put pressure on the research community to develop predictive models that are not only globally accurate but also meet new fairness criteria, reflecting the lack of disparate mistreatment with respect to sensitive attributes ($\textit{e.g.}$ gender, ethnicity, age). In particular, the variability of the errors made by certain Facial Recognition (FR) systems across specific segments of the population compromises the deployment of the latter, and was judged unacceptable by regulatory authorities. Designing fair FR systems is a very challenging problem, mainly due to the complex and functional nature of the performance measure used in this domain ($\textit{i.e.}$ ROC curves) and because of the huge heterogeneity of the face image datasets usually available for training. In this paper, we propose a novel post-processing approach to improve the fairness of pre-trained FR models by optimizing a regression loss which acts on centroid-based scores. Beyond the computational advantages of the method, we present numerical experiments providing strong empirical evidence of the gain in fairness and of the ability to preserve global accuracy.
Authors: Weishi Wang, Mark K. Transtrum, Vincenzo Lordi, Vasily V. Bulatov, Amit Samanta
Abstract: An adaptive physics-informed model design strategy for machine-learning interatomic potentials (MLIPs) is proposed. This strategy follows an iterative reconfiguration of composite models from single-term models, followed by a unified training procedure. A model evaluation method based on the Fisher information matrix (FIM) and multiple-property error metrics is proposed to guide model reconfiguration and hyperparameter optimization. Combining the model reconfiguration and the model evaluation subroutines, we provide an adaptive MLIP design strategy that balances flexibility and extensibility. In a case study of designing models against a structurally diverse niobium dataset, we managed to obtain an optimal configuration with 75 parameters generated by our framework that achieved a force RMSE of 0.172 eV/{\AA} and an energy RMSE of 0.013 eV/atom.
Authors: Jacky He, Guiran Liu, Binrong Zhu, Hanlu Zhang, Hongye Zheng, Xiaokai Wang
Abstract: This paper focuses on the dynamic optimization of the Retrieval-Augmented Generation (RAG) architecture. It proposes a state-aware dynamic knowledge retrieval mechanism to enhance semantic understanding and knowledge scheduling efficiency in large language models for open-domain question answering and complex generation tasks. The method introduces a multi-level perceptive retrieval vector construction strategy and a differentiable document matching path. These components enable end-to-end joint training and collaborative optimization of the retrieval and generation modules. This effectively addresses the limitations of static RAG structures in context adaptation and knowledge access. Experiments are conducted on the Natural Questions dataset. The proposed structure is thoroughly evaluated across different large models, including GPT-4, GPT-4o, and DeepSeek. Comparative and ablation experiments from multiple perspectives confirm the significant improvements in BLEU and ROUGE-L scores. The approach also demonstrates stronger robustness and generation consistency in tasks involving semantic ambiguity and multi-document fusion. These results highlight its broad application potential and practical value in building high-quality language generation systems.
Authors: Junting Ren, Armin Schwartzman
Abstract: When implementing prediction models for high-stakes real-world applications such as medicine, finance, and autonomous systems, quantifying prediction uncertainty is critical for effective risk management. Traditional approaches to uncertainty quantification, such as confidence and prediction intervals, provide probability coverage guarantees for the expected outcomes $f(\boldsymbol{x})$ or the realized outcomes $f(\boldsymbol{x})+\epsilon$. Instead, this paper introduces a novel, model-agnostic framework for quantifying uncertainty in continuous and binary outcomes using confidence sets for outcome excursions, where the goal is to identify a subset of the feature space where the expected or realized outcome exceeds a specific value. The proposed method constructs data-dependent inner and outer confidence sets that aim to contain the true feature subset for which the expected or realized outcomes of these features exceed a specified threshold. We establish theoretical guarantees for the probability that these confidence sets contain the true feature subset, both asymptotically and for finite sample sizes. The framework is validated through simulations and applied to real-world datasets, demonstrating its utility in contexts such as housing price prediction and time to sepsis diagnosis in healthcare. This approach provides a unified method for uncertainty quantification that is broadly applicable across various continuous and binary prediction models.
Authors: Sonia Joseph, Praneet Suresh, Lorenz Hufe, Edward Stevinson, Robert Graham, Yash Vadi, Danilo Bzdok, Sebastian Lapuschkin, Lee Sharkey, Blake Aaron Richards
Abstract: Robust tooling and publicly available pre-trained models have helped drive recent advances in mechanistic interpretability for language models. However, similar progress in vision mechanistic interpretability has been hindered by the lack of accessible frameworks and pre-trained weights. We present Prisma (Access the codebase here: https://github.com/Prisma-Multimodal/ViT-Prisma), an open-source framework designed to accelerate vision mechanistic interpretability research, providing a unified toolkit for accessing 75+ vision and video transformers; support for sparse autoencoder (SAE), transcoder, and crosscoder training; a suite of 80+ pre-trained SAE weights; activation caching, circuit analysis tools, and visualization tools; and educational resources. Our analysis reveals surprising findings, including that effective vision SAEs can exhibit substantially lower sparsity patterns than language SAEs, and that in some instances, SAE reconstructions can decrease model loss. Prisma enables new research directions for understanding vision model internals while lowering barriers to entry in this emerging field.
Authors: Mina Karzand, Guy Bresler
Abstract: We consider an online model for recommendation systems, with each user being recommended an item at each time-step and providing 'like' or 'dislike' feedback. A latent variable model specifies the user preferences: both users and items are clustered into types. The model captures structure in both the item and user spaces, as used by item-item and user-user collaborative filtering algorithms. We study the situation in which the type preference matrix has i.i.d. entries. Our main contribution is an algorithm that simultaneously uses both item and user structures, proved to be near-optimal via corresponding information-theoretic lower bounds. In particular, our analysis highlights the sub-optimality of using only one of item or user structure (as is done in most collaborative filtering algorithms).
Authors: Vijay Prakash S
Abstract: Straight line equation $y=mx$ with slope $m$, when singularly perturbed as $ay^3+y=mx$ with a positive parameter $a$, results in S-shaped curves or S-curves on a real plane. As $a\rightarrow 0$, we get back $y=mx$ which is a cumulative distribution function of a continuous uniform distribution that describes the occurrence of every event in an interval to be equally probable. As $a\rightarrow\infty$, the derivative of $y$ has finite support only at $y=0$ resembling a degenerate distribution. Based on these arguments, in this work, we propose that these S-curves can represent maximum entropy uniform distribution to a zero entropy single value. We also argue that these S-curves are superposable as they are only parametrically nonlinear but fundamentally linear. So far, the superposed forms have been used to capture the patterns of natural systems such as nonlinear dynamics of biological growth and kinetics of enzyme reactions. Here, we attempt to use the S-curve and its superposed form as a statistical model. We fit the models on a classical dataset containing flower measurements of iris plants and analyze their usefulness in pattern recognition. Based on these models, we claim that any non-uniform pattern can be represented as a singular perturbation to uniform distribution. However, our parametric estimation procedure have some limitations such as sensitivity to initial conditions depending on the data at hand.
Authors: Kanghong Shi, Ruigang Wang, Ian R. Manchester
Abstract: We propose a neural control method to provide guaranteed stabilization for mechanical systems using a novel negative imaginary neural ordinary differential equation (NINODE) controller. Specifically, we employ neural networks with desired properties as state-space function matrices within a Hamiltonian framework to ensure the system possesses the NI property. This NINODE system can serve as a controller that asymptotically stabilizes an NI plant under certain conditions. For mechanical plants with colocated force actuators and position sensors, we demonstrate that all the conditions required for stability can be translated into regularity constraints on the neural networks used in the controller. We illustrate the utility, effectiveness, and stability guarantees of the NINODE controller through an example involving a nonlinear mass-spring system.
Authors: Omid Semiari, Hosein Nikopour, Shilpa Talwar
Abstract: Next-generation wireless cellular networks are expected to provide unparalleled Quality-of-Service (QoS) for emerging wireless applications, necessitating strict performance guarantees, e.g., in terms of link-level data rates. A critical challenge in meeting these QoS requirements is the prevention of cell congestion, which involves balancing the load to ensure sufficient radio resources are available for each cell to serve its designated User Equipments (UEs). In this work, a novel QoS-aware Load Balancing (LB) approach is developed to optimize the performance of Guaranteed Bit Rate (GBR) and Best Effort (BE) traffic in a multi-band Open Radio Access Network (O-RAN) under QoS and resource constraints. The proposed solution builds on Graph Reinforcement Learning (GRL), a powerful framework at the intersection of Graph Neural Network (GNN) and RL. The QoS-aware LB is modeled as a Markov Decision Process, with states represented as graphs. QoS consideration are integrated into both state representations and reward signal design. The LB agent is then trained using an off-policy dueling Deep Q Network (DQN) that leverages a GNN-based architecture. This design ensures the LB policy is invariant to the ordering of nodes (UE or cell), flexible in handling various network sizes, and capable of accounting for spatial node dependencies in LB decisions. Performance of the GRL-based solution is compared with two baseline methods. Results show substantial performance gains, including a $53\%$ reduction in QoS violations and a fourfold increase in the 5th percentile rate for BE traffic.
Authors: Ke Hong, Xiuhong Li, Minxu Liu, Qiuli Mao, Tianqi Wu, Zixiao Huang, Lufang Chen, Zhong Wang, Yichong Zhang, Zhenhua Zhu, Guohao Dai, Yu Wang
Abstract: Generative models have achieved remarkable success across various applications, driving the demand for multi-GPU computing. Inter-GPU communication becomes a bottleneck in multi-GPU computing systems, particularly on consumer-grade GPUs. By exploiting concurrent hardware execution, overlapping computation and communication latency is an effective technique for mitigating the communication overhead. We identify that an efficient and adaptable overlapping design should satisfy (1) tile-wise overlapping to maximize the overlapping opportunity, (2) interference-free computation to maintain the original computational performance, and (3) communication agnosticism to reduce the development burden against varying communication primitives. Nevertheless, current designs fail to simultaneously optimize for all of those features. To address the issue, we propose FlashOverlap, a lightweight design characterized by tile-wise overlapping, interference-free computation, and communication agnosticism. FlashOverlap utilizes a novel signaling mechanism to identify tile-wise data dependency without interrupting the computation process, and reorders data to contiguous addresses, enabling communication by simply calling NCCL APIs. Experiments show that such a lightweight design achieves up to 1.65x speedup, outperforming existing works in most cases.
Authors: Xi Fu, Wei-Bang Jiang, Yi Ding, Cuntai Guan
Abstract: Multimodal physiological signals, such as EEG, ECG, EOG, and EMG, are crucial for healthcare and brain-computer interfaces. While existing methods rely on specialized architectures and dataset-specific fusion strategies, they struggle to learn universal representations that generalize across datasets and handle missing modalities at inference time. To address these issues, we propose PhysioOmni, a foundation model for multimodal physiological signal analysis that models both homogeneous and heterogeneous features to decouple multimodal signals and extract generic representations while maintaining compatibility with arbitrary missing modalities. PhysioOmni trains a decoupled multimodal tokenizer, enabling masked signal pre-training via modality-invariant and modality-specific objectives. To ensure adaptability to diverse and incomplete modality combinations, the pre-trained encoders undergo resilient fine-tuning with prototype alignment on downstream datasets. Extensive experiments on four downstream tasks, emotion recognition, sleep stage classification, motor prediction, and mental workload detection, demonstrate that PhysioOmni achieves state-of-the-art performance while maintaining strong robustness to missing modalities. Our code and model weights will be released.
Authors: Kaichen Zhang, Yuzhong Hong, Junwei Bao, Hongfei Jiang, Yang Song, Dingqian Hong, Hui Xiong
Abstract: Post-training plays a crucial role in refining and aligning large language models to meet specific tasks and human preferences. While recent advancements in post-training techniques, such as Group Relative Policy Optimization (GRPO), leverage increased sampling with relative reward scoring to achieve superior performance, these methods often suffer from training instability that limits their practical adoption. To address this challenge, we present Group Variance Policy Optimization (GVPO). GVPO incorporates the analytical solution to KL-constrained reward maximization directly into its gradient weights, ensuring alignment with the optimal policy. The method provides intuitive physical interpretations: its gradient mirrors the mean squared error between the central distance of implicit rewards and that of actual rewards. GVPO offers two key advantages: (1) it guarantees a unique optimal solution, exactly the KL-constrained reward maximization objective, (2) it supports flexible sampling distributions that avoids on-policy and importance sampling limitations. By unifying theoretical guarantees with practical adaptability, GVPO establishes a new paradigm for reliable and versatile LLM post-training.
Authors: Massimo Fioravanti, Samuele Pasini, Giovanni Agosta
Abstract: Reinforcement learning (RL) algorithms, due to their reliance on external systems to learn from, require digital environments (e.g., simulators) with very simple interfaces, which in turn constrain significantly the implementation of such environments. In particular, these environments are implemented either as separate processes or as state machines, leading to synchronization and communication overheads in the first case, and to unstructured programming in the second. We propose a new domain-specific, co-routine-based, compiled language, called Rulebook, designed to automatically generate the state machine required to interact with machine learning (ML) algorithms and similar applications, with no performance overhead. Rulebook allows users to express programs without needing to be aware of the specific interface required by the ML components. By decoupling the execution model of the program from the syntactical encoding of the program, and thus without the need for manual state management, Rulebook allows to create larger and more sophisticated environments at a lower development cost.
Authors: Run Luo, Renke Shan, Longze Chen, Ziqiang Liu, Lu Wang, Min Yang, Xiaobo Xia
Abstract: Large Vision-Language Models (LVLMs) are pivotal for real-world AI tasks like embodied intelligence due to their strong vision-language reasoning abilities. However, current LVLMs process entire images at the token level, which is inefficient compared to humans who analyze information and generate content at the conceptual level, extracting relevant visual concepts with minimal effort. This inefficiency, stemming from the lack of a visual concept model, limits LVLMs' usability in real-world applications. To address this, we propose VCM, an end-to-end self-supervised visual concept modeling framework. VCM leverages implicit contrastive learning across multiple sampled instances and vision-language fine-tuning to construct a visual concept model without requiring costly concept-level annotations. Our results show that VCM significantly reduces computational costs (e.g., 85\% fewer FLOPs for LLaVA-1.5-7B) while maintaining strong performance across diverse image understanding tasks. Moreover, VCM enhances visual encoders' capabilities in classic visual concept perception tasks. Extensive quantitative and qualitative experiments validate the effectiveness and efficiency of VCM.
Authors: Subham Das, Ashtakala Meghanath, Bikash K. Behera, Shahid Mumtaz, Saif Al-Kuwari, Ahmed Farouk
Abstract: Social financial technology focuses on trust, sustainability, and social responsibility, which require advanced technologies to address complex financial tasks in the digital era. With the rapid growth in online transactions, automating credit card fraud detection and loan eligibility prediction has become increasingly challenging. Classical machine learning (ML) models have been used to solve these challenges; however, these approaches often encounter scalability, overfitting, and high computational costs due to complexity and high-dimensional financial data. Quantum computing (QC) and quantum machine learning (QML) provide a promising solution to efficiently processing high-dimensional datasets and enabling real-time identification of subtle fraud patterns. However, existing quantum algorithms lack robustness in noisy environments and fail to optimize performance with reduced feature sets. To address these limitations, we propose a quantum feature deep neural network (QFDNN), a novel, resource efficient, and noise-resilient quantum model that optimizes feature representation while requiring fewer qubits and simpler variational circuits. The model is evaluated using credit card fraud detection and loan eligibility prediction datasets, achieving competitive accuracies of 82.2% and 74.4%, respectively, with reduced computational overhead. Furthermore, we test QFDNN against six noise models, demonstrating its robustness across various error conditions. Our findings highlight QFDNN potential to enhance trust and security in social financial technology by accurately detecting fraudulent transactions while supporting sustainability through its resource-efficient design and minimal computational overhead.
Authors: Yike Zhao, Haoyuan Cai, Ali H. Sayed
Abstract: This paper studies a stochastic dynamic game between two competing teams, each consisting of a network of collaborating agents. Unlike fully cooperative settings, where all agents share a common objective, each team in this game aims to minimize its own distinct objective. In the adversarial setting, their objectives could be conflicting as in zero-sum games. Throughout the competition, agents share strategic information within their own team while simultaneously inferring and adapting to the strategies of the opposing team. We propose diffusion learning algorithms to address two important classes of this network game: i) a zero-sum game characterized by weak cross-team subgraph interactions, and ii) a general non-zero-sum game exhibiting strong cross-team subgraph interactions. We analyze the stability performance of the proposed algorithms under reasonable assumptions and illustrate the theoretical results through experiments on Cournot team competition and decentralized GAN training.
Authors: Leon Davies, Baihua Li, Mohamad Saada, Simon S{\o}lvsten, Qinggang Meng
Abstract: SLAM is a fundamental component of modern autonomous systems, providing robots and their operators with a deeper understanding of their environment. SLAM systems often encounter challenges due to the dynamic nature of robotic motion, leading to inaccuracies in mapping quality, particularly in 2D representations such as Occupancy Grid Maps. These errors can significantly degrade map quality, hindering the effectiveness of specific downstream tasks such as floor plan creation. To address this challenge, we introduce our novel 'GAN-SLAM', a new SLAM approach that leverages Generative Adversarial Networks to clean and complete occupancy grids during the SLAM process, reducing the impact of noise and inaccuracies introduced on the output map. We adapt and integrate accurate pose estimation techniques typically used for 3D SLAM into a 2D form. This enables the quality improvement 3D LiDAR-odometry has seen in recent years to be effective for 2D representations. Our results demonstrate substantial improvements in map fidelity and quality, with minimal noise and errors, affirming the effectiveness of GAN-SLAM for real-world mapping applications within large-scale complex environments. We validate our approach on real-world data operating in real-time, and on famous examples of 2D maps. The improved quality of the output map enables new downstream tasks, such as floor plan drafting, further enhancing the capabilities of autonomous systems. Our novel approach to SLAM offers a significant step forward in the field, improving the usability for SLAM in mapping-based tasks, and offers insight into the usage of GANs for OGM error correction.
Authors: Leon Davies, Baihua Li, Mohamad Saada, Simon S{\o}lvsten, Qinggang Meng
Abstract: SLAM (Simultaneous Localisation and Mapping) is a crucial component for robotic systems, providing a map of an environment, the current location and previous trajectory of a robot. While 3D LiDAR SLAM has received notable improvements in recent years, 2D SLAM lags behind. Gradual drifts in odometry and pose estimation inaccuracies hinder modern 2D LiDAR-odometry algorithms in large complex environments. Dynamic robotic motion coupled with inherent estimation based SLAM processes introduce noise and errors, degrading map quality. Occupancy Grid Mapping (OGM) produces results that are often noisy and unclear. This is due to the fact that evidence based mapping represents maps according to uncertain observations. This is why OGMs are so popular in exploration or navigation tasks. However, this also limits OGMs' effectiveness for specific mapping based tasks such as floor plan creation in complex scenes. To address this, we propose our novel Transformation and Translation Occupancy Grid Mapping (TT-OGM). We adapt and enable accurate and robust pose estimation techniques from 3D SLAM to the world of 2D and mitigate errors to improve map quality using Generative Adversarial Networks (GANs). We introduce a novel data generation method via deep reinforcement learning (DRL) to build datasets large enough for training a GAN for SLAM error correction. We demonstrate our SLAM in real-time on data collected at Loughborough University. We also prove its generalisability on a variety of large complex environments on a collection of large scale well-known 2D occupancy maps. Our novel approach enables the creation of high quality OGMs in complex scenes, far surpassing the capabilities of current SLAM algorithms in terms of quality, accuracy and reliability.
Authors: Shotaro Takasu, Toshio Aoyagi
Abstract: Reservoir computing is a powerful framework for real-time information processing, characterized by its high computational ability and quick learning, with applications ranging from machine learning to biological systems. In this paper, we demonstrate that the memory capacity of a reservoir recurrent neural network scales sublinearly with the number of readout neurons. To elucidate this phenomenon, we develop a theoretical framework for analytically deriving memory capacity, attributing the decaying growth of memory capacity to neuronal correlations. In addition, numerical simulations reveal that once memory capacity becomes sublinear, increasing the number of readout neurons successively enables nonlinear processing at progressively higher polynomial orders. Furthermore, our theoretical framework suggests that neuronal correlations govern not only memory capacity but also the sequential growth of nonlinear computational capabilities. Our findings establish a foundation for designing scalable and cost-effective reservoir computing, providing novel insights into the interplay among neuronal correlations, linear memory, and nonlinear processing.
Authors: Osma Suominen, Juho Inkinen, Mona Lehtinen
Abstract: This paper presents the Annif system in SemEval-2025 Task 5 (LLMs4Subjects), which focussed on subject indexing using large language models (LLMs). The task required creating subject predictions for bibliographic records from the bilingual TIBKAT database using the GND subject vocabulary. Our approach combines traditional natural language processing and machine learning techniques implemented in the Annif toolkit with innovative LLM-based methods for translation and synthetic data generation, and merging predictions from monolingual models. The system ranked first in the all-subjects category and second in the tib-core-subjects category in the quantitative evaluation, and fourth in qualitative evaluations. These findings demonstrate the potential of combining traditional XMTC algorithms with modern LLM techniques to improve the accuracy and efficiency of subject indexing in multilingual contexts.
Authors: Mohamed Amine Ferrag, Norbert Tihanyi, Merouane Debbah
Abstract: Large language models and autonomous AI agents have evolved rapidly, resulting in a diverse array of evaluation benchmarks, frameworks, and collaboration protocols. However, the landscape remains fragmented and lacks a unified taxonomy or comprehensive survey. Therefore, we present a side-by-side comparison of benchmarks developed between 2019 and 2025 that evaluate these models and agents across multiple domains. In addition, we propose a taxonomy of approximately 60 benchmarks that cover general and academic knowledge reasoning, mathematical problem-solving, code generation and software engineering, factual grounding and retrieval, domain-specific evaluations, multimodal and embodied tasks, task orchestration, and interactive assessments. Furthermore, we review AI-agent frameworks introduced between 2023 and 2025 that integrate large language models with modular toolkits to enable autonomous decision-making and multi-step reasoning. Moreover, we present real-world applications of autonomous AI agents in materials science, biomedical research, academic ideation, software engineering, synthetic data generation, chemical reasoning, mathematical problem-solving, geographic information systems, multimedia, healthcare, and finance. We then survey key agent-to-agent collaboration protocols, namely the Agent Communication Protocol (ACP), the Model Context Protocol (MCP), and the Agent-to-Agent Protocol (A2A). Finally, we discuss recommendations for future research, focusing on advanced reasoning strategies, failure modes in multi-agent LLM systems, automated scientific discovery, dynamic tool integration via reinforcement learning, integrated search capabilities, and security vulnerabilities in agent protocols.
Authors: Nikolaos Chaidos, Angeliki Dimitriou, Nikolaos Spanos, Athanasios Voulodimos, Giorgos Stamou
Abstract: Graph Neural Networks (GNNs) have emerged as an efficient alternative to convolutional approaches for vision tasks such as image classification, leveraging patch-based representations instead of raw pixels. These methods construct graphs where image patches serve as nodes, and edges are established based on patch similarity or classification relevance. Despite their efficiency, the explainability of GNN-based vision models remains underexplored, even though graphs are naturally interpretable. In this work, we analyze the semantic consistency of the graphs formed at different layers of GNN-based image classifiers, focusing on how well they preserve object structures and meaningful relationships. A comprehensive analysis is presented by quantifying the extent to which inter-layer graph connections reflect semantic similarity and spatial coherence. Explanations from standard and adversarial settings are also compared to assess whether they reflect the classifiers' robustness. Additionally, we visualize the flow of information across layers through heatmap-based visualization techniques, thereby highlighting the models' explainability. Our findings demonstrate that the decision-making processes of these models can be effectively explained, while also revealing that their reasoning does not necessarily align with human perception, especially in deeper layers.
Authors: Anush Lakshman Sivaraman, Kojo Adu-Gyamfi, Ibne Farabi Shihab, Anuj Sharma
Abstract: Accurate weather classification from low-quality traffic camera imagery remains a challenging task, particularly under adverse nighttime conditions. In this study, we propose a scalable framework that combines generative domain adaptation with efficient contrastive learning to enhance classification performance. Using CycleGAN-based domain translation, we improve the quality of nighttime images, enabling better feature extraction by downstream models. While the baseline EVA-02 model employing CLIP-based contrastive loss achieves an overall accuracy of 96.55\%, it exhibits a significant performance gap between daytime (97.21\%) and nighttime conditions (63.40\%). Replacing CLIP with the lightweight SigLIP-2 (Sigmoid contrastive loss) achieves a competitive overall accuracy of 94.00\%, with substantial improvements in nighttime performance (85.90\% accuracy). The combination of Vision-SigLIP-2, Text-SigLIP-2, CycleGAN, and contrastive training achieves the best nighttime accuracy (85.90\%) among all models tested, while EVA-02 with CycleGAN maintains the highest overall accuracy (97.01\%) and per-class accuracies. These findings demonstrate the potential of combining domain adaptation and efficient contrastive learning to build practical, resource-efficient weather classification systems for intelligent transportation infrastructure.
Authors: Heisei Yonezawa, Ansei Yonezawa, Itsuro Kajiwara
Abstract: Complex mechanical systems such as vehicle powertrains are inherently subject to multiple nonlinearities and uncertainties arising from parametric variations. Modeling and calibration errors are therefore unavoidable, making the transfer of control systems from simulation to real-world systems a critical challenge. Traditional robust controls have limitations in handling certain types of nonlinearities and uncertainties, requiring a more practical approach capable of comprehensively compensating for these various constraints. This study proposes a new robust control approach using the framework of deep reinforcement learning (DRL). The key strategy lies in the synergy among domain randomization-based DRL, long short-term memory (LSTM)-based actor and critic networks, and model-based control (MBC). The problem setup is modeled via the latent Markov decision process (LMDP), a set of vanilla MDPs, for a controlled system subject to uncertainties and nonlinearities. In LMDP, the dynamics of an environment simulator is randomized during training to improve the robustness of the control system to real testing environments. The randomization increases training difficulties as well as conservativeness of the resultant control system; therefore, progress is assisted by concurrent use of a model-based controller based on a nominal system model. Compared to traditional DRL-based controls, the proposed controller design is smarter in that we can achieve a high level of generalization ability with a more compact neural network architecture and a smaller amount of training data. The proposed approach is verified via practical application to active damping for a complex powertrain system with nonlinearities and parametric variations. Comparative tests demonstrate the high robustness of the proposed approach.
Authors: Ranran Zhen, Juntao Li, Yixin Ji, Zhenlin Yang, Tong Liu, Qingrong Xia, Xinyu Duan, Zhefeng Wang, Baoxing Huai, Min Zhang
Abstract: Large Language Models (LLMs) for Generative AI have achieved remarkable progress, evolving into sophisticated and versatile tools widely adopted across various domains and applications. However, the substantial memory overhead caused by their vast number of parameters, combined with the high computational demands of the attention mechanism, poses significant challenges in achieving low latency and high throughput for LLM inference services. Recent advancements, driven by groundbreaking research, have significantly accelerated progress in this field. This paper provides a comprehensive survey of these methods, covering fundamental instance-level approaches, in-depth cluster-level strategies, emerging scenario directions, and other miscellaneous but important areas. At the instance level, we review model placement, request scheduling, decoding length prediction, storage management, and the disaggregation paradigm. At the cluster level, we explore GPU cluster deployment, multi-instance load balancing, and cloud service solutions. For emerging scenarios, we organize the discussion around specific tasks, modules, and auxiliary methods. To ensure a holistic overview, we also highlight several niche yet critical areas. Finally, we outline potential research directions to further advance the field of LLM inference serving.
Authors: Rupert Polley, Nikolai Polley, Dominik Heid, Marc Heinrich, Sven Ochs, J. Marius Z\"ollner
Abstract: Traffic light perception is an essential component of the camera-based perception system for autonomous vehicles, enabling accurate detection and interpretation of traffic lights to ensure safe navigation through complex urban environments. In this work, we propose a modularized perception framework that integrates state-of-the-art detection models with a novel real-time association and decision framework, enabling seamless deployment into an autonomous driving stack. To address the limitations of existing public datasets, we introduce the ATLAS dataset, which provides comprehensive annotations of traffic light states and pictograms across diverse environmental conditions and camera setups. This dataset is publicly available at https://url.fzi.de/ATLAS. We train and evaluate several state-of-the-art traffic light detection architectures on ATLAS, demonstrating significant performance improvements in both accuracy and robustness. Finally, we evaluate the framework in real-world scenarios by deploying it in an autonomous vehicle to make decisions at traffic light-controlled intersections, highlighting its reliability and effectiveness for real-time operation.
Authors: Yingbin Bai, Sylvie Thiebaux, Felipe Trevizan
Abstract: Learning-based planners leveraging Graph Neural Networks can learn search guidance applicable to large search spaces, yet their potential to address symmetries remains largely unexplored. In this paper, we introduce a graph representation of planning problems allying learning efficiency with the ability to detect symmetries, along with two pruning methods, action pruning and state pruning, designed to manage symmetries during search. The integration of these techniques into Fast Downward achieves a first-time success over LAMA on the latest IPC learning track dataset. Code is released at: https://github.com/bybeye/Distincter.
Authors: Aditi Nachnani, Kai K. Li-Caldwell, Saptarshi Biswas, Prince Sharma, Gaoyuan Ouyang, Prashant Singh
Abstract: We present a machine-learning guided approach to predict saturation magnetization (MS) and coercivity (HC) in Fe-rich soft magnetic alloys, particularly Fe-Si-B systems. ML models trained on experimental data reveals that increasing Si and B content reduces MS from 1.81T (DFT~2.04 T) to ~1.54 T (DFT~1.56T) in Fe-Si-B, which is attributed to decreased magnetic density and structural modifications. Experimental validation of ML predicted magnetic saturation on Fe-1Si-1B (2.09T), Fe-5Si-5B (2.01T) and Fe-10Si-10B (1.54T) alloy compositions further support our findings. These trends are consistent with density functional theory (DFT) predictions, which link increased electronic disorder and band broadening to lower MS values. Experimental validation on selected alloys confirms the predictive accuracy of the ML model, with good agreement across compositions. Beyond predictive accuracy, detailed uncertainty quantification and model interpretability including through feature importance and partial dependence analysis reveals that MS is governed by a nonlinear interplay between Fe content, early transition metal ratios, and annealing temperature, while HC is more sensitive to processing conditions such as ribbon thickness and thermal treatment windows. The ML framework was further applied to Fe-Si-B/Cr/Cu/Zr/Nb alloys in a pseudo-quaternary compositional space, which shows comparable magnetic properties to NANOMET (Fe84.8Si0.5B9.4Cu0.8 P3.5C1), FINEMET (Fe73.5Si13.5B9 Cu1Nb3), NANOPERM (Fe88Zr7B4Cu1), and HITPERM (Fe44Co44Zr7B4Cu1. Our fundings demonstrate the potential of ML framework for accelerated search of high-performance, Co- and Ni-free, soft magnetic materials.
Authors: Gang Mao, Tousif Rahman, Sidharth Maheshwari, Bob Pattison, Zhuang Shao, Rishad Shafik, Alex Yakovlev
Abstract: The increased demand for data privacy and security in machine learning (ML) applications has put impetus on effective edge training on Internet-of-Things (IoT) nodes. Edge training aims to leverage speed, energy efficiency and adaptability within the resource constraints of the nodes. Deploying and training Deep Neural Networks (DNNs)-based models at the edge, although accurate, posit significant challenges from the back-propagation algorithm's complexity, bit precision trade-offs, and heterogeneity of DNN layers. This paper presents a Dynamic Tsetlin Machine (DTM) training accelerator as an alternative to DNN implementations. DTM utilizes logic-based on-chip inference with finite-state automata-driven learning within the same Field Programmable Gate Array (FPGA) package. Underpinned on the Vanilla and Coalesced Tsetlin Machine algorithms, the dynamic aspect of the accelerator design allows for a run-time reconfiguration targeting different datasets, model architectures, and model sizes without resynthesis. This makes the DTM suitable for targeting multivariate sensor-based edge tasks. Compared to DNNs, DTM trains with fewer multiply-accumulates, devoid of derivative computation. It is a data-centric ML algorithm that learns by aligning Tsetlin automata with input data to form logical propositions enabling efficient Look-up-Table (LUT) mapping and frugal Block RAM usage in FPGA training implementations. The proposed accelerator offers 2.54x more Giga operations per second per Watt (GOP/s per W) and uses 6x less power than the next-best comparable design.
Authors: Erblin Isaku, Hassan Sartaj, Shaukat Ali
Abstract: An autonomous vessel (AV) is a complex cyber-physical system (CPS) with software enabling many key functionalities, e.g., navigation software enables an AV to autonomously or semi-autonomously follow a path to its destination. Digital twins of such AVs enable advanced functionalities such as running what-if scenarios, performing predictive maintenance, and enabling fault diagnosis. Due to technological improvements, real-time analyses using continuous data from vessels' real-time operations have become increasingly possible. However, the literature has little explored developing advanced analyses in real-time data in AVs with digital twins built with machine learning techniques. To this end, we present a novel digital twin-based approach (ODDIT) to detect future out-of-distribution (OOD) states of an AV before reaching them, enabling proactive intervention. Such states may indicate anomalies requiring attention (e.g., manual correction by the ship master) and assist testers in scenario-centered testing. The digital twin consists of two machine-learning models predicting future vessel states and whether the predicted state will be OOD. We evaluated ODDIT with five vessels across waypoint and zigzag maneuvering under simulated conditions, including sensor and actuator noise and environmental disturbances i.e., ocean current. ODDIT achieved high accuracy in detecting OOD states, with AUROC and TNR@TPR95 scores reaching 99\% across multiple vessels.
Authors: Daniel Kienzle, Robin Sch\"on, Rainer Lienhart, Shin'Ichi Satoh
Abstract: Analyzing a player's technique in table tennis requires knowledge of the ball's 3D trajectory and spin. While, the spin is not directly observable in standard broadcasting videos, we show that it can be inferred from the ball's trajectory in the video. We present a novel method to infer the initial spin and 3D trajectory from the corresponding 2D trajectory in a video. Without ground truth labels for broadcast videos, we train a neural network solely on synthetic data. Due to the choice of our input data representation, physically correct synthetic training data, and using targeted augmentations, the network naturally generalizes to real data. Notably, these simple techniques are sufficient to achieve generalization. No real data at all is required for training. To the best of our knowledge, we are the first to present a method for spin and trajectory prediction in simple monocular broadcast videos, achieving an accuracy of 92.0% in spin classification and a 2D reprojection error of 0.19% of the image diagonal.
Authors: Ke Hong, Lufang Chen, Zhong Wang, Xiuhong Li, Qiuli Mao, Jianping Ma, Chao Xiong, Guanyu Wu, Buhe Han, Guohao Dai, Yun Liang, Yu Wang
Abstract: Existing large language model (LLM) serving systems fall into two categories: 1) a unified system where prefill phase and decode phase are co-located on the same GPU, sharing the unified computational resource and storage, and 2) a disaggregated system where the two phases are disaggregated to different GPUs. The design of the disaggregated system addresses the latency interference and sophisticated scheduling issues in the unified system but leads to storage challenges including 1) replicated weights for both phases that prevent flexible deployment, 2) KV cache transfer overhead between the two phases, 3) storage imbalance that causes substantial wasted space of the GPU capacity, and 4) suboptimal resource adjustment arising from the difficulties in migrating KV cache. Such storage inefficiency delivers poor serving performance under high request rates. In this paper, we identify that the advantage of the disaggregated system lies in the disaggregated computation, i.e., partitioning the computational resource to enable the asynchronous computation of two phases. Thus, we propose a novel LLM serving system, semi-PD, characterized by disaggregated computation and unified storage. In semi-PD, we introduce a computation resource controller to achieve disaggregated computation at the streaming multi-processor (SM) level, and a unified memory manager to manage the asynchronous memory access from both phases. semi-PD has a low-overhead resource adjustment mechanism between the two phases, and a service-level objective (SLO) aware dynamic partitioning algorithm to optimize the SLO attainment. Compared to state-of-the-art systems, semi-PD maintains lower latency at higher request rates, reducing the average end-to-end latency per request by 1.27-2.58x on DeepSeek series models, and serves 1.55-1.72x more requests adhering to latency constraints on Llama series models.
Authors: Athinagoras Skiadopoulos, Mark Zhao, Swapnil Gandhi, Thomas Norrie, Shrijeet Mukherjee, Christos Kozyrakis
Abstract: Mixture-of-Experts (MoE) models have become a widely adopted solution to continue scaling model sizes without a corresponding linear increase in compute. During MoE model training, each input token is dynamically routed to a subset of experts -- sparsely-activated feed-forward networks -- within each transformer layer. The distribution of tokens assigned to each expert varies widely and rapidly over the course of training. To handle the wide load imbalance across experts, current systems are forced to either drop tokens assigned to popular experts, degrading convergence, or frequently rebalance resources allocated to each expert based on popularity, incurring high state migration overheads. To break this performance-accuracy tradeoff, we introduce SwiftMoE, an adaptive MoE training system. The key insight of SwiftMoE is to decouple the placement of expert parameters from their large optimizer state. SwiftMoE statically partitions the optimizer of each expert across all training nodes. Meanwhile, SwiftMoE dynamically adjusts the placement of expert parameters by repurposing existing weight updates, avoiding migration overheads. In doing so, SwiftMoE right-sizes the GPU resources allocated to each expert, on a per-iteration basis, with minimal overheads. Compared to state-of-the-art MoE training systems, DeepSpeed and FlexMoE, SwiftMoE is able to achieve a 30.5% and 25.9% faster time-to-convergence, respectively.
Authors: Riccardo Lo Bianco, Willem van Jaarsveld, Jeroen Middelhuis, Luca Begnardi, Remco Dijkman
Abstract: The Dynamic Task Assignment Problem (DTAP) concerns matching resources to tasks in real time while minimizing some objectives, like resource costs or task cycle time. In this work, we consider a DTAP variant where every task is a case composed of a stochastic sequence of activities. The DTAP, in this case, involves the decision of which employee to assign to which activity to process requests as quickly as possible. In recent years, Deep Reinforcement Learning (DRL) has emerged as a promising tool for tackling this DTAP variant, but most research is limited to solving small-scale, synthetic problems, neglecting the challenges posed by real-world use cases. To bridge this gap, this work proposes a DRL-based Decision Support System (DSS) for real-world scale DTAPS. To this end, we introduce a DRL agent with two novel elements: a graph structure for observations and actions that can effectively represent any DTAP and a reward function that is provably equivalent to the objective of minimizing the average cycle time of tasks. The combination of these two novelties allows the agent to learn effective and generalizable assignment policies for real-world scale DTAPs. The proposed DSS is evaluated on five DTAP instances whose parameters are extracted from real-world logs through process mining. The experimental evaluation shows how the proposed DRL agent matches or outperforms the best baseline in all DTAP instances and generalizes on different time horizons and across instances.
Authors: Shubhada Agrawal, Aaditya Ramdas
Abstract: We prove two lower bounds for stopping times of sequential tests between general composite nulls and alternatives. The first lower bound is for the setting where the type-1 error level $\alpha$ approaches zero, and equals $\log(1/\alpha)$ divided by a certain infimum KL divergence, termed $\operatorname{KL_{inf}}$. The second lower bound applies to the setting where $\alpha$ is fixed and $\operatorname{KL_{inf}}$ approaches 0 (meaning that the null and alternative sets are not separated) and equals $c \operatorname{KL_{inf}}^{-1} \log \log \operatorname{KL_{inf}}^{-1}$ for a universal constant $c > 0$. We also provide a sufficient condition for matching the upper bounds and show that this condition is met in several special cases. Given past work, these upper and lower bounds are unsurprising in their form; our main contribution is the generality in which they hold, for example, not requiring reference measures or compactness of the classes.
Authors: Adway Das, Agnimitra Sengupta, S. Ilgin Guler
Abstract: Traffic flow prediction is a critical component of intelligent transportation systems, yet accurately forecasting traffic remains challenging due to the interaction between long-term trends and short-term fluctuations. Standard deep learning models often struggle with these challenges because their architectures inherently smooth over fine-grained fluctuations while focusing on general trends. This limitation arises from low-pass filtering effects, gate biases favoring stability, and memory update mechanisms that prioritize long-term information retention. To address these shortcomings, this study introduces a hybrid deep learning framework that integrates both long-term trend and short-term fluctuation information using two input features processed in parallel, designed to capture complementary aspects of traffic flow dynamics. Further, our approach leverages attention mechanisms, specifically Bahdanau attention, to selectively focus on critical time steps within traffic data, enhancing the model's ability to predict congestion and other transient phenomena. Experimental results demonstrate that features learned from both branches are complementary, significantly improving the goodness-of-fit statistics across multiple prediction horizons compared to a baseline model. Notably, the attention mechanism enhances short-term forecast accuracy by directly targeting immediate fluctuations, though challenges remain in fully integrating long-term trends. This framework can contribute to more effective congestion mitigation and urban mobility planning by advancing the robustness and precision of traffic prediction models.
Authors: Yomn Alkabakibi, Congwei Xie, Artem R. Oganov
Abstract: Nonlinear optical (NLO) materials for generating lasers via second harmonic generation (SHG) are highly sought in today's technology. However, discovering novel materials with considerable SHG is challenging due to the time-consuming and costly nature of both experimental methods and first-principles calculations. In this study, we present a deep learning approach using the Atomistic Line Graph Neural Network (ALIGNN) to predict NLO properties. Sourcing data from the Novel Opto-Electronic Materials Discovery (NOEMD) database and using the Kurtz-Perry (KP) coefficient as the key target, we developed a robust model capable of accurately estimating nonlinear optical responses. Our results demonstrate that the model achieves 82.5% accuracy at a tolerated absolute error up to 1 pm/V and relative error not exceeding 0.5. This work highlights the potential of deep learning in accelerating the discovery and design of advanced optical materials with desired properties.
Authors: Ioannis Kontogiorgakis, Iason Tsardanidis, Dimitrios Bormpoudakis, Ilias Tsoumas, Dimitra A. Loka, Christos Noulas, Alexandros Tsitouras, Charalampos Kontoes
Abstract: Effective weed management is crucial for improving agricultural productivity, as weeds compete with crops for vital resources like nutrients and water. Accurate maps of weed management methods are essential for policymakers to assess farmer practices, evaluate impacts on vegetation health, biodiversity, and climate, as well as ensure compliance with policies and subsidies. However, monitoring weed management methods is challenging as commonly rely on on-ground field surveys, which are often costly, time-consuming and subject to delays. In order to tackle this problem, we leverage Earth Observation (EO) data and Machine Learning (ML). Specifically, we developed an ML approach for mapping four distinct weed management methods (Mowing, Tillage, Chemical-spraying, and No practice) in orchards using satellite image time series (SITS) data from two different sources: Sentinel-2 (S2) and PlanetScope (PS). The findings demonstrate the potential of ML-driven remote sensing to enhance the efficiency and accuracy of weed management mapping in orchards.
Authors: Andreas Kalogeras, Dimitrios Bormpoudakis, Iason Tsardanidis, Dimitra A. Loka, Charalampos Kontoes
Abstract: The widespread use of Exogenous Organic Matter in agriculture necessitates monitoring to assess its effects on soil and crop health. This study evaluates optical Sentinel-2 satellite imagery for detecting digestate application, a practice that enhances soil fertility but poses environmental risks like microplastic contamination and nitrogen losses. In the first instance, Sentinel-2 satellite image time series (SITS) analysis of specific indices (EOMI, NDVI, EVI) was used to characterize EOM's spectral behavior after application on the soils of four different crop types in Thessaly, Greece. Furthermore, Machine Learning (ML) models (namely Random Forest, k-NN, Gradient Boosting and a Feed-Forward Neural Network), were used to investigate digestate presence detection, achieving F1-scores up to 0.85. The findings highlight the potential of combining remote sensing and ML for scalable and cost-effective monitoring of EOM applications, supporting precision agriculture and sustainability.
Authors: Rishika Sen, Sujoy Roychowdhury, Sumit Soman, H. G. Ranjani, Srikhetra Mohanty
Abstract: Knowledge Distillation (KD) is one of the approaches to reduce the size of Large Language Models (LLMs). A LLM with smaller number of model parameters (student) is trained to mimic the performance of a LLM of a larger size (teacher model) on a specific task. For domain-specific tasks, it is not clear if teacher or student model, or both, must be considered for domain adaptation. In this work, we study this problem from perspective of telecom domain Question-Answering (QA) task. We systematically experiment with Supervised Fine-tuning (SFT) of teacher only, SFT of student only and SFT of both prior to KD. We design experiments to study the impact of vocabulary (same and different) and KD algorithms (vanilla KD and Dual Space KD, DSKD) on the distilled model. Multi-faceted evaluation of the distillation using 14 different metrics (N-gram, embedding and LLM-based metrics) is considered. Experimental results show that SFT of teacher improves performance of distilled model when both models have same vocabulary, irrespective of algorithm and metrics. Overall, SFT of both teacher and student results in better performance across all metrics, although the statistical significance of the same depends on the vocabulary of the teacher models.
Authors: Jing Wang, Yan Jin, Hamid Taghavifar, Fei Ding, Chongfeng Wei
Abstract: Since the emergence of autonomous driving technology, it has advanced rapidly over the past decade. It is becoming increasingly likely that autonomous vehicles (AVs) would soon coexist with human-driven vehicles (HVs) on the roads. Currently, safety and reliable decision-making remain significant challenges, particularly when AVs are navigating lane changes and interacting with surrounding HVs. Therefore, precise estimation of the intentions of surrounding HVs can assist AVs in making more reliable and safe lane change decision-making. This involves not only understanding their current behaviors but also predicting their future motions without any direct communication. However, distinguishing between the passing and yielding intentions of surrounding HVs still remains ambiguous. To address the challenge, we propose a social intention estimation algorithm rooted in Directed Acyclic Graph (DAG), coupled with a decision-making framework employing Deep Reinforcement Learning (DRL) algorithms. To evaluate the method's performance, the proposed framework can be tested and applied in a lane-changing scenario within a simulated environment. Furthermore, the experiment results demonstrate how our approach enhances the ability of AVs to navigate lane changes safely and efficiently on roads.
Authors: Aditya Vatsavai, Ganesh Narasimha, Yongtao Liu, Jan-Chi Yang, Hiroshu Funakubo, Maxim Ziatdinov, Rama Vasudevan
Abstract: Rapidly determining structure-property correlations in materials is an important challenge in better understanding fundamental mechanisms and greatly assists in materials design. In microscopy, imaging data provides a direct measurement of the local structure, while spectroscopic measurements provide relevant functional property information. Deep kernel active learning approaches have been utilized to rapidly map local structure to functional properties in microscopy experiments, but are computationally expensive for multi-dimensional and correlated output spaces. Here, we present an alternative lightweight curiosity algorithm which actively samples regions with unexplored structure-property relations, utilizing a deep-learning based surrogate model for error prediction. We show that the algorithm outperforms random sampling for predicting properties from structures, and provides a convenient tool for efficient mapping of structure-property relationships in materials science.
Authors: Pritika Rohera, Chaitrali Ginimav, Gayatri Sawant, Raviraj Joshi
Abstract: Multilingual Large Language Models (LLMs) have demonstrated significant effectiveness across various languages, particularly in high-resource languages such as English. However, their performance in terms of factual accuracy across other low-resource languages, especially Indic languages, remains an area of investigation. In this study, we assess the factual accuracy of LLMs - GPT-4o, Gemma-2-9B, Gemma-2-2B, and Llama-3.1-8B - by comparing their performance in English and Indic languages using the IndicQuest dataset, which contains question-answer pairs in English and 19 Indic languages. By asking the same questions in English and their respective Indic translations, we analyze whether the models are more reliable for regional context questions in Indic languages or when operating in English. Our findings reveal that LLMs often perform better in English, even for questions rooted in Indic contexts. Notably, we observe a higher tendency for hallucination in responses generated in low-resource Indic languages, highlighting challenges in the multilingual understanding capabilities of current LLMs.
Authors: Roman Garipov, Fedor Velikonivtsev, Ruslan Svirschevski, Vage Egiazarian, Max Ryabinin
Abstract: We introduce AutoJudge, a framework that accelerates large language model (LLM) inference with task-specific lossy speculative decoding. Instead of matching the original model output distribution token-by-token, we identify which of the generated tokens affect the downstream quality of the generated response, relaxing the guarantee so that the "unimportant" tokens can be generated faster. Our approach relies on a semi-greedy search algorithm to test which of the mismatches between target and draft model should be corrected to preserve quality, and which ones may be skipped. We then train a lightweight classifier based on existing LLM embeddings to predict, at inference time, which mismatching tokens can be safely accepted without compromising the final answer quality. We test our approach with Llama 3.2 1B (draft) and Llama 3.1 8B (target) models on zero-shot GSM8K reasoning, where it achieves up to 1.5x more accepted tokens per verification cycle with under 1% degradation in answer accuracy compared to standard speculative decoding and over 2x with small loss in accuracy. When applied to the LiveCodeBench benchmark, our approach automatically detects other, programming-specific important tokens and shows similar speedups, demonstrating its ability to generalize across tasks.
Authors: John T\"ornblom, Emil Karlsson, Simin Nadjm-Tehrani
Abstract: The ability to explain why a machine learning model arrives at a particular prediction is crucial when used as decision support by human operators of critical systems. The provided explanations must be provably correct, and preferably without redundant information, called minimal explanations. In this paper, we aim at finding explanations for predictions made by tree ensembles that are not only minimal, but also minimum with respect to a cost function. To this end, we first present a highly efficient oracle that can determine the correctness of explanations, surpassing the runtime performance of current state-of-the-art alternatives by several orders of magnitude when computing minimal explanations. Secondly, we adapt an algorithm called MARCO from related works (calling it m-MARCO) for the purpose of computing a single minimum explanation per prediction, and demonstrate an overall speedup factor of two compared to the MARCO algorithm which enumerates all minimal explanations. Finally, we study the obtained explanations from a range of use cases, leading to further insights of their characteristics. In particular, we observe that in several cases, there are more than 100,000 minimal explanations to choose from for a single prediction. In these cases, we see that only a small portion of the minimal explanations are also minimum, and that the minimum explanations are significantly less verbose, hence motivating the aim of this work.
Authors: Yiran Ye, Thai Le, Dongwon Lee
Abstract: Online texts with toxic content are a clear threat to the users on social media in particular and society in general. Although many platforms have adopted various measures (e.g., machine learning-based hate-speech detection systems) to diminish their effect, toxic content writers have also attempted to evade such measures by using cleverly modified toxic words, so-called human-written text perturbations. Therefore, to help build automatic detection tools to recognize those perturbations, prior methods have developed sophisticated techniques to generate diverse adversarial samples. However, we note that these ``algorithms"-generated perturbations do not necessarily capture all the traits of ``human"-written perturbations. Therefore, in this paper, we introduce a novel, high-quality dataset of human-written perturbations, named as NoisyHate, that was created from real-life perturbations that are both written and verified by human-in-the-loop. We show that perturbations in NoisyHate have different characteristics than prior algorithm-generated toxic datasets show, and thus can be in particular useful to help develop better toxic speech detection solutions. We thoroughly validate NoisyHate against state-of-the-art language models, such as BERT and RoBERTa, and black box APIs, such as Perspective API, on two tasks, such as perturbation normalization and understanding.
Authors: Hiroki Naganuma, Ryuichiro Hataya, Kotaro Yoshida, Ioannis Mitliagkas
Abstract: In the field of computer vision, fine-tuning pre-trained models has become a prevalent strategy for out-of-distribution (OOD) generalization tasks. Different from most prior work that has focused on advancing learning algorithms, we systematically examined how pre-trained model size, pre-training dataset size, and training strategies impact generalization and confidence calibration on downstream tasks. We evaluated 100 models across diverse pre-trained model sizes, five pre-training datasets, and five data augmentations through extensive experiments on four distribution shift datasets totaling over 120,000 GPU hours. Our results demonstrate the significant impact of pre-trained model selection, with optimal choices substantially improving OOD accuracy over algorithm improvement alone. Additionally, we find that larger models and bigger pre-training datasets not only enhance OOD performance but also improve calibration, helping to mitigate overconfidence, contrary to some prior studies that found modern deep networks to calibrate worse than classical shallow models. Our work underscores the overlooked importance of pre-trained model selection for out-of-distribution generalization and calibration.
Authors: Rumeng Li, Xun Wang, Dan Berlowitz, Brian Silver, Wen Hu, Heather Keating, Raelene Goodwin, Weisong Liu, Honghuang Lin, Hong Yu
Abstract: Early prediction of Alzheimer's disease (AD) is crucial for timely intervention and treatment. This study aims to use machine learning approaches to analyze longitudinal electronic health records (EHRs) of patients with AD and identify signs and symptoms that can predict AD onset earlier. We used a case-control design with longitudinal EHRs from the U.S. Department of Veterans Affairs Veterans Health Administration (VHA) from 2004 to 2021. Cases were VHA patients with AD diagnosed after 1/1/2016 based on ICD-10-CM codes, matched 1:9 with controls by age, sex and clinical utilization with replacement. We used a panel of AD-related keywords and their occurrences over time in a patient's longitudinal EHRs as predictors for AD prediction with four machine learning models. We performed subgroup analyses by age, sex, and race/ethnicity, and validated the model in a hold-out and "unseen" VHA stations group. Model discrimination, calibration, and other relevant metrics were reported for predictions up to ten years before ICD-based diagnosis. The study population included 16,701 cases and 39,097 matched controls. The average number of AD-related keywords (e.g., "concentration", "speaking") per year increased rapidly for cases as diagnosis approached, from around 10 to over 40, while remaining flat at 10 for controls. The best model achieved high discriminative accuracy (ROCAUC 0.997) for predictions using data from at least ten years before ICD-based diagnoses. The model was well-calibrated (Hosmer-Lemeshow goodness-of-fit p-value = 0.99) and consistent across subgroups of age, sex and race/ethnicity, except for patients younger than 65 (ROCAUC 0.746). Machine learning models using AD-related keywords identified from EHR notes can predict future AD diagnoses, suggesting its potential use for identifying AD risk using EHR notes, offering an affordable way for early screening on large population.
Authors: Surojit Ganguli, Zeyu Zhou, Christopher G. Brinton, David I. Inouye
Abstract: When each edge device of a network only perceives a local part of the environment, collaborative inference across multiple devices is often needed to predict global properties of the environment. In safety-critical applications, collaborative inference must be robust to significant network failures caused by environmental disruptions or extreme weather. Existing collaborative learning approaches, such as privacy-focused Vertical Federated Learning (VFL), typically assume a centralized setup or that one device never fails. However, these assumptions make prior approaches susceptible to significant network failures. To address this problem, we first formalize the problem of robust collaborative inference over a dynamic network of devices that could experience significant network faults. Then, we develop a minimalistic yet impactful method called Multiple Aggregation with Gossip Rounds and Simulated Faults (MAGS) that synthesizes simulated faults via dropout, replication, and gossiping to significantly improve robustness over baselines. We also theoretically analyze our proposed approach to explain why each component enhances robustness. Extensive empirical results validate that MAGS is robust across a range of fault rates-including extreme fault rates.
Authors: Jiabin Lin, Shana Moothedath
Abstract: We present conservative distributed multi-task learning in stochastic linear contextual bandits with heterogeneous agents. This extends conservative linear bandits to a distributed setting where M agents tackle different but related tasks while adhering to stage-wise performance constraints. The exact context is unknown, and only a context distribution is available to the agents as in many practical applications that involve a prediction mechanism to infer context, such as stock market prediction and weather forecast. We propose a distributed upper confidence bound (UCB) algorithm, DiSC-UCB. Our algorithm constructs a pruned action set during each round to ensure the constraints are met. Additionally, it includes synchronized sharing of estimates among agents via a central server using well-structured synchronization steps. We prove the regret and communication bounds on the algorithm. We extend the problem to a setting where the agents are unaware of the baseline reward. For this setting, we provide a modified algorithm, DiSC-UCB2, and we show that the modified algorithm achieves the same regret and communication bounds. We empirically validated the performance of our algorithm on synthetic data and real-world Movielens-100K data.
Authors: Jamelle Watson-Daniels, Flavio du Pin Calmon, Alexander D'Amour, Carol Long, David C. Parkes, Berk Ustun
Abstract: Issues can arise when research focused on fairness, transparency, or safety is conducted separately from research driven by practical deployment concerns and vice versa. This separation creates a growing need for translational work that bridges the gap between independently studied concepts that may be fundamentally related. This paper explores connections between two seemingly unrelated concepts of predictive inconsistency that share intriguing parallels. The first, known as predictive multiplicity, occurs when models that perform similarly (e.g., nearly equivalent training loss) produce conflicting predictions for individual samples. This concept is often emphasized in algorithmic fairness research as a means of promoting transparency in ML model development. The second concept, predictive churn, examines the differences in individual predictions before and after model updates, a key challenge in deploying ML models in consumer-facing applications. We present theoretical and empirical results that uncover links between these previously disconnected concepts.
Authors: Sungmin Cha, Kyunghyun Cho
Abstract: Continual learning (CL) aims to train a model on a sequence of tasks (i.e., a CL scenario) while balancing the trade-off between plasticity (learning new tasks) and stability (retaining prior knowledge). The dominantly adopted conventional evaluation protocol for CL algorithms selects the best hyperparameters (e.g., learning rate, mini-batch size, regularization strengths, etc.) within a given scenario and then evaluates the algorithms using these hyperparameters in the same scenario. However, this protocol has significant shortcomings: it overestimates the CL capacity of algorithms and relies on unrealistic hyperparameter tuning, which is not feasible for real-world applications. From the fundamental principles of evaluation in machine learning, we argue that the evaluation of CL algorithms should focus on assessing the generalizability of their CL capacity to unseen scenarios. Based on this, we propose the Generalizable Two-phase Evaluation Protocol (GTEP) consisting of hyperparameter tuning and evaluation phases. Both phases share the same scenario configuration (e.g., number of tasks) but are generated from different datasets. Hyperparameters of CL algorithms are tuned in the first phase and applied in the second phase to evaluate the algorithms. We apply this protocol to class-incremental learning, both with and without pretrained models. Across more than 8,000 experiments, our results show that most state-of-the-art algorithms fail to replicate their reported performance, highlighting that their CL capacity has been significantly overestimated in the conventional evaluation protocol. Our implementation can be found in https://github.com/csm9493/GTEP.
Authors: Jiayuan Dong, Christian Jacobsen, Mehdi Khalloufi, Maryam Akram, Wanjiao Liu, Karthik Duraisamy, Xun Huan
Abstract: Bayesian optimal experimental design (OED) seeks experiments that maximize the expected information gain (EIG) in model parameters. Directly estimating the EIG using nested Monte Carlo is computationally expensive and requires an explicit likelihood. Variational OED (vOED), in contrast, estimates a lower bound of the EIG without likelihood evaluations by approximating the posterior distributions with variational forms, and then tightens the bound by optimizing its variational parameters. We introduce the use of normalizing flows (NFs) for representing variational distributions in vOED; we call this approach vOED-NFs. Specifically, we adopt NFs with a conditional invertible neural network architecture built from compositions of coupling layers, and enhanced with a summary network for data dimension reduction. We present Monte Carlo estimators to the lower bound along with gradient expressions to enable a gradient-based simultaneous optimization of the variational parameters and the design variables. The vOED-NFs algorithm is then validated in two benchmark problems, and demonstrated on a partial differential equation-governed application of cathodic electrophoretic deposition and an implicit likelihood case with stochastic modeling of aphid population. The findings suggest that a composition of 4--5 coupling layers is able to achieve lower EIG estimation bias, under a fixed budget of forward model runs, compared to previous approaches. The resulting NFs produce approximate posteriors that agree well with the true posteriors, able to capture non-Gaussian and multi-modal features effectively.
Authors: Geert De Paepe, Lesley De Cruz
Abstract: In meteorology, finding similar weather patterns or analogs in historical datasets can be useful for data assimilation, forecasting, and postprocessing. In climate science, analogs in historical and climate projection data are used for attribution and impact studies. However, most of the time, those large weather and climate datasets are nearline. This means that they must be downloaded, which takes a lot of bandwidth and disk space, before the computationally expensive search can be executed. We propose a dimension reduction technique based on autoencoder (AE) neural networks to compress the datasets and perform the search in an interpretable, compressed latent space. A distance-regularized Siamese twin autoencoder (DIRESA) architecture is designed to preserve distance in latent space while capturing the nonlinearities in the datasets. Using conceptual climate models of different complexities, we show that the latent components thus obtained provide physical insight into the dominant modes of variability in the system. Compressing datasets with DIRESA reduces the online storage and keeps the latent components uncorrelated, while the distance (ordering) preservation and reconstruction fidelity robustly outperform Principal Component Analysis (PCA) and other dimension reduction techniques such as UMAP or variational autoencoders.
Authors: Ruichen Wang, Dinesh Manocha
Abstract: We present a novel machine-learning (ML) approach (EM-GANSim) for real-time electromagnetic (EM) propagation that is used for wireless communication simulation in 3D indoor environments. Our approach uses a modified conditional Generative Adversarial Network (GAN) that incorporates encoded geometry and transmitter location while adhering to the electromagnetic propagation theory. The overall physically-inspired learning is able to predict the power distribution in 3D scenes, which is represented using heatmaps. We evaluated our method on 15 complex 3D indoor environments, with 4 additional scenarios later included in the results, showcasing the generalizability of the model across diverse conditions. Our overall accuracy is comparable to ray tracing-based EM simulation, as evidenced by lower mean squared error values. Furthermore, our GAN-based method drastically reduces the computation time, achieving a 5X speedup on complex benchmarks. In practice, it can compute the signal strength in a few milliseconds on any location in 3D indoor environments. We also present a large dataset of 3D models and EM ray tracing-simulated heatmaps. To the best of our knowledge, EM-GANSim is the first real-time algorithm for EM simulation in complex 3D indoor environments. We plan to release the code and the dataset.
Authors: Emi Soroka, Rohan Sinha, Sanjay Lall
Abstract: Temporal logic rules are often used in control and robotics to provide structured, human-interpretable descriptions of trajectory data. These rules have numerous applications including safety validation using formal methods, constraining motion planning among autonomous agents, and classifying data. However, existing methods for learning temporal logic predicates from data do not provide assurances about the correctness of the resulting predicate. We present a novel method to learn temporal logic predicates from data with finite-sample correctness guarantees. Our approach leverages expression optimization and conformal prediction to learn predicates that correctly describe future trajectories under mild statistical assumptions. We provide experimental results showing the performance of our approach on a simulated trajectory dataset and perform ablation studies to understand how each component of our algorithm contributes to its performance.
Authors: Yedi Zhang, Andrew Saxe, Peter E. Latham
Abstract: We investigate the implications of removing bias in ReLU networks regarding their expressivity and learning dynamics. We first show that two-layer bias-free ReLU networks have limited expressivity: the only odd function two-layer bias-free ReLU networks can express is a linear one. We then show that, under symmetry conditions on the data, these networks have the same learning dynamics as linear networks. This enables us to give analytical time-course solutions to certain two-layer bias-free (leaky) ReLU networks outside the lazy learning regime. While deep bias-free ReLU networks are more expressive than their two-layer counterparts, they still share a number of similarities with deep linear networks. These similarities enable us to leverage insights from linear networks to understand certain ReLU networks. Overall, our results show that some properties previously established for bias-free ReLU networks arise due to equivalence to linear networks.
Authors: Xinxing Shi, Thomas Baldwin-McDonald, Mauricio A. \'Alvarez
Abstract: Deep Gaussian Processes (DGPs) leverage a compositional structure to model non-stationary processes. DGPs typically rely on local inducing point approximations across intermediate GP layers. Recent advances in DGP inference have shown that incorporating global Fourier features from the Reproducing Kernel Hilbert Space (RKHS) can enhance the DGPs' capability to capture complex non-stationary patterns. This paper extends the use of these features to compositional GPs involving linear transformations. In particular, we introduce Ordinary Differential Equation(ODE)--based RKHS Fourier features that allow for adaptive amplitude and phase modulation through convolution operations. This convolutional formulation relates our work to recently proposed deep latent force models, a multi-layer structure designed for modelling nonlinear dynamical systems. By embedding these adjustable RKHS Fourier features within a doubly stochastic variational inference framework, our model exhibits improved predictive performance across various regression tasks.
Authors: Lucas Joos, Daniel A. Keim, Maximilian T. Fischer
Abstract: Systematic literature reviews (SLRs) are essential but labor-intensive due to high publication volumes and inefficient keyword-based filtering. To streamline this process, we evaluate Large Language Models (LLMs) for enhancing efficiency and accuracy in corpus filtration while minimizing manual effort. Our open-source tool LLMSurver presents a visual interface to utilize LLMs for literature filtration, evaluate the results, and refine queries in an interactive way. We assess the real-world performance of our approach in filtering over 8.3k articles during a recent survey construction, comparing results with human efforts. The findings show that recent LLM models can reduce filtering time from weeks to minutes. A consensus scheme ensures recall rates >98.8%, surpassing typical human error thresholds and improving selection accuracy. This work advances literature review methodologies and highlights the potential of responsible human-AI collaboration in academic research.
Authors: Joana Reuss, Jan Macdonald, Simon Becker, Lorenz Richter, Marco K\"orner
Abstract: We introduce EuroCropsML, an analysis-ready remote sensing machine learning dataset for time series crop type classification of agricultural parcels in Europe. It is the first dataset designed to benchmark transnational few-shot crop type classification algorithms that supports advancements in algorithmic development and research comparability. It comprises 706 683 multi-class labeled data points across 176 classes, featuring annual time series of per-parcel median pixel values from Sentinel-2 L1C data for 2021, along with crop type labels and spatial coordinates. Based on the open-source EuroCrops collection, EuroCropsML is publicly available on Zenodo.
Authors: Agathe Fernandes Machado, Arthur Charpentier, Ewen Gallic
Abstract: In this paper, we link two existing approaches to derive counterfactuals: adaptations based on a causal graph, and optimal transport. We extend "Knothe's rearrangement" and "triangular transport" to probabilistic graphical models, and use this counterfactual approach, referred to as sequential transport, to discuss fairness at the individual level. After establishing the theoretical foundations of the proposed method, we demonstrate its application through numerical experiments on both synthetic and real datasets.
Authors: Pinak Mandal, Georg A. Gottwald, Nicholas Cranch
Abstract: The computationally cheap machine learning architecture of random feature maps can be viewed as a single-layer feedforward network in which the weights of the hidden layer are random but fixed and only the outer weights are learned via linear regression. The internal weights are typically chosen from a prescribed distribution. The choice of the internal weights significantly impacts the accuracy of random feature maps. We address here the task of how to best select the internal weights. In particular, we consider the forecasting problem whereby random feature maps are used to learn a one-step propagator map for a dynamical system. We provide a computationally cheap hit-and-run algorithm to select good internal weights which lead to good forecasting skill. We show that the number of good features is the main factor controlling the forecasting skill of random feature maps and acts as an effective feature dimension. Lastly, we compare random feature maps with single-layer feedforward neural networks in which the internal weights are now learned using gradient descent. We find that random feature maps have superior forecasting capabilities whilst having several orders of magnitude lower computational cost.
Authors: Mohammad Mohammadi, Sreejita Ghosh
Abstract: Classification of sets of inputs (e.g., images and texts) is an active area of research within both computer vision (CV) and natural language processing (NLP). A common way to represent a set of vectors is to model them as linear subspaces. In this contribution, we present a prototype-based approach for learning on the manifold formed from such linear subspaces, the Grassmann manifold. Our proposed method learns a set of subspace prototypes capturing the representative characteristics of classes and a set of relevance factors automating the selection of the dimensionality of the subspaces. This leads to a transparent classifier model which presents the computed impact of each input vector on its decision. Through experiments on benchmark image and text datasets, we have demonstrated the efficiency of our proposed classifier, compared to the transformer-based models in terms of not only performance and explainability but also computational resource requirements.
Authors: Yuxia Wu, Lizi Liao, Yuan Fang
Abstract: Modeling dynamic graphs, such as those found in social networks, recommendation systems, and e-commerce platforms, is crucial for capturing evolving relationships and delivering relevant insights over time. Traditional approaches primarily rely on graph neural networks with temporal components or sequence generation models, which often focus narrowly on the historical context of target nodes. This limitation restricts the ability to adapt to new and emerging patterns in dynamic graphs. To address this challenge, we propose a novel framework, Retrieval-Augmented Generation for Dynamic Graph modeling (RAG4DyG), which enhances dynamic graph predictions by incorporating contextually and temporally relevant examples from broader graph structures. Our approach includes a time- and context-aware contrastive learning module to identify high-quality demonstrations and a graph fusion strategy to effectively integrate these examples with historical contexts. The proposed framework is designed to be effective in both transductive and inductive scenarios, ensuring adaptability to previously unseen nodes and evolving graph structures. Extensive experiments across multiple real-world datasets demonstrate the effectiveness of RAG4DyG in improving predictive accuracy and adaptability for dynamic graph modeling. The code and datasets are publicly available at https://github.com/YuxiaWu/RAG4DyG.
Authors: Amogh Joshi, Adarsh Kumar Kosta, Kaushik Roy
Abstract: The ability of neural networks to perform robotic perception and control tasks such as depth and optical flow estimation, simultaneous localization and mapping (SLAM), and automatic control has led to their widespread adoption in recent years. Deep Reinforcement Learning has been used extensively in these settings, as it does not have the unsustainable training costs associated with supervised learning. However, DeepRL suffers from poor sample efficiency, i.e., it requires a large number of environmental interactions to converge to an acceptable solution. Modern RL algorithms such as Deep Q Learning and Soft Actor-Critic attempt to remedy this shortcoming but can not provide the explainability required in applications such as autonomous robotics. Humans intuitively understand the long-time-horizon sequential tasks common in robotics. Properly using such intuition can make RL policies more explainable while enhancing their sample efficiency. In this work, we propose SHIRE, a novel framework for encoding human intuition using Probabilistic Graphical Models (PGMs) and using it in the Deep RL training pipeline to enhance sample efficiency. Our framework achieves 25-78% sample efficiency gains across the environments we evaluate at negligible overhead cost. Additionally, by teaching RL agents the encoded elementary behavior, SHIRE enhances policy explainability. A real-world demonstration further highlights the efficacy of policies trained using our framework.
Authors: Barak Gahtan, Robert J. Shahla, Reuven Cohen, Alex M. Bronstein
Abstract: QUIC, a new and increasingly used transport protocol, enhances TCP by offering improved security, performance, and stream multiplexing. These features, however, also impose challenges for network middle-boxes that need to monitor and analyze web traffic. This paper proposes a novel method to estimate the number of HTTP/3 responses in a given QUIC connection by an observer. This estimation reveals server behavior, client-server interactions, and data transmission efficiency, which is crucial for various applications such as designing a load balancing solution and detecting HTTP/3 flood attacks. The proposed scheme transforms QUIC connection traces into image sequences and uses machine learning (ML) models, guided by a tailored loss function, to predict response counts. Evaluations on more than seven million images-derived from 100,000 traces collected across 44,000 websites over four months-achieve up to 97% accuracy in both known and unknown server settings and 92% accuracy on previously unseen complete QUIC traces.
Authors: Noam Razin, Sadhika Malladi, Adithya Bhaskar, Danqi Chen, Sanjeev Arora, Boris Hanin
Abstract: Direct Preference Optimization (DPO) and its variants are increasingly used for aligning language models with human preferences. Although these methods are designed to teach a model to generate preferred responses more frequently relative to dispreferred responses, prior work has observed that the likelihood of preferred responses often decreases during training. The current work sheds light on the causes and implications of this counter-intuitive phenomenon, which we term likelihood displacement. We demonstrate that likelihood displacement can be catastrophic, shifting probability mass from preferred responses to responses with an opposite meaning. As a simple example, training a model to prefer $\texttt{No}$ over $\texttt{Never}$ can sharply increase the probability of $\texttt{Yes}$. Moreover, when aligning the model to refuse unsafe prompts, we show that such displacement can unintentionally lead to unalignment, by shifting probability mass from preferred refusal responses to harmful responses (e.g., reducing the refusal rate of Llama-3-8B-Instruct from 74.4% to 33.4%). We theoretically characterize that likelihood displacement is driven by preferences that induce similar embeddings, as measured by a centered hidden embedding similarity (CHES) score. Empirically, the CHES score enables identifying which training samples contribute most to likelihood displacement in a given dataset. Filtering out these samples effectively mitigated unintentional unalignment in our experiments. More broadly, our results highlight the importance of curating data with sufficiently distinct preferences, for which we believe the CHES score may prove valuable.
Authors: Lothar Sebastian Krapp, Laura Wirth
Abstract: The Fundamental Theorem of Statistical Learning states that a hypothesis space is PAC learnable if and only if its VC dimension is finite. For the agnostic model of PAC learning, the literature so far presents proofs of this theorem that often tacitly impose several measurability assumptions on the involved sets and functions. We scrutinize these proofs from a measure-theoretic perspective in order to explicitly extract the assumptions needed for a rigorous argument. This leads to a sound statement as well as a detailed and self-contained proof of the Fundamental Theorem of Statistical Learning in the agnostic setting, showcasing the minimal measurability requirements needed. As the Fundamental Theorem of Statistical Learning underpins a wide range of further theoretical developments, our results are of foundational importance: A careful analysis of measurability aspects is essential, especially when the theorem is used in settings where measure-theoretic subtleties play a role. We particularly discuss applications in Model Theory, considering NIP and o-minimal structures. Our main theorem presents sufficient conditions for the PAC learnability of hypothesis spaces defined over o-minimal expansions of the reals. This class of hypothesis spaces covers all artificial neural networks for binary classification that use commonly employed activation functions like ReLU and the sigmoid function.
Authors: Zhaoyang Wang, Weilei He, Zhiyuan Liang, Xuchao Zhang, Chetan Bansal, Ying Wei, Weitong Zhang, Huaxiu Yao
Abstract: Recent self-rewarding large language models (LLM) have successfully applied LLM-as-a-Judge to iteratively improve the alignment performance without the need of human annotations for preference data. These methods commonly utilize the same LLM to act as both the policy model (which generates responses) and the reward model (which scores and ranks those responses). The ranked responses are then used as preference pairs to train the LLM via direct alignment technologies (e.g. DPO). However, it is noteworthy that throughout this process, there is no guarantee of accuracy in the rewarding and ranking, which is critical for ensuring accurate rewards and high-quality preference data. Empirical results from relatively small LLMs (e.g., 7B parameters) also indicate that improvements from self-rewarding may diminish after several iterations in certain situations, which we hypothesize is due to accumulated bias in the reward system. This bias can lead to unreliable preference data for training the LLM. To address this issue, we first formulate and analyze the generalized iterative preference fine-tuning framework for self-rewarding language model. We then introduce the regularization to this generalized framework to mitigate the overconfident preference labeling in the self-rewarding process. Based on this theoretical insight, we propose a Consistency Regularized sElf-rewarding lAnguage Model (CREAM) that leverages the consistency of rewards across different iterations to regularize the self-rewarding training, helping the model to learn from more reliable preference data. With this explicit regularization, our empirical results demonstrate the superiority of CREAM in improving both reward consistency and alignment performance. The code is publicly available at https://github.com/Raibows/CREAM.
Authors: Yann Bouteiller, Karthik Soma, Giovanni Beltrame
Abstract: The universe involves many independent co-learning agents as an ever-evolving part of our observed environment. Yet, in practice, Multi-Agent Reinforcement Learning (MARL) applications are typically constrained to small, homogeneous populations and remain computationally intensive. We propose a methodology that enables simulating populations of Reinforcement Learning agents at evolutionary scale. More specifically, we derive a fast, parallelizable implementation of Policy Gradient (PG) and Opponent-Learning Awareness (LOLA), tailored for evolutionary simulations where agents undergo random pairwise interactions in stateless normal-form games. We demonstrate our approach by simulating the evolution of very large populations made of heterogeneous co-learning agents, under both naive and advanced learning strategies. In our experiments, 200,000 PG or LOLA agents evolve in the classic games of Hawk-Dove, Stag-Hunt, and Rock-Paper-Scissors. Each game provides distinct insights into how populations evolve under both naive and advanced MARL rules, including compelling ways in which Opponent-Learning Awareness affects social evolution.
Authors: Michael Noukhovitch, Shengyi Huang, Sophie Xhonneux, Arian Hosseini, Rishabh Agarwal, Aaron Courville
Abstract: The dominant paradigm for RLHF is online and on-policy RL: synchronously generating from the large language model (LLM) policy, labelling with a reward model, and learning using feedback on the LLM's own outputs. While performant, this paradigm is computationally inefficient. Inspired by classical deep RL literature, we propose separating generation and learning in RLHF. This enables asynchronous generation of new samples while simultaneously training on old samples, leading to faster training and more compute-optimal scaling. However, asynchronous training relies on an underexplored regime, online but off-policy RLHF: learning on samples from previous iterations of our model which give a worse training signal. We tackle the fundamental challenge in this regime: how much off-policyness can we tolerate for asynchronous training to speed up learning but maintain performance? Among several RLHF algorithms we test, online DPO is found to be most robust to off-policy data, and robustness increases with the scale of the policy model. We study further compute optimizations for asynchronous RLHF but find that they come at a performance cost, giving rise to a trade-off. We verify the scalability of asynchronous RLHF by training a general-purpose chatbot from LLaMA 3.1 8B on an instruction-following task ~40% faster than a synchronous run while matching final performance. Finally, we extend our results to math and reasoning to demonstrate asynchronous RL can finetune Rho 1B on GSM8k ~70% faster while matching synchronous accuracy.
Authors: Jong-Ik Park, Srinivasa Pranav, Jos\'e M. F. Moura, Carlee Joe-Wong
Abstract: Foundation models are now a major focus of leading technology organizations due to their ability to generalize across diverse tasks. Existing approaches for adapting foundation models to new applications often rely on Federated Learning (FL) and disclose the foundation model weights to clients when using it to initialize the global model. While these methods ensure client data privacy, they compromise model and information security. In this paper, we introduce Federated Learning Aggregation Biased by a Foundation Model (FedBaF), a novel method for dynamically integrating pre-trained foundation model weights during the FL aggregation phase. Unlike conventional methods, FedBaF preserves the confidentiality of the foundation model while still leveraging its power to train more accurate models, especially in non-IID and adversarial scenarios. Our comprehensive experiments use Pre-ResNet and foundation models like Vision Transformer to demonstrate that FedBaF not only matches, but often surpasses the test accuracy of traditional weight initialization methods by up to 11.4% in IID and up to 15.8% in non-IID settings. Additionally, FedBaF applied to a Transformer-based language model significantly reduced perplexity by up to 39.2%.
Authors: Hongjia Wu, Hui Zeng, Zehui Xiong, Jiawen Kang, Zhiping Cai, Tse-Tin Chan, Dusit Niyato, Zhu Han
Abstract: Timely updating of Internet of Things (IoT) data is crucial for immersive vehicular metaverse services. However, challenges such as latency caused by massive data transmissions, privacy risks associated with user data, and computational burdens on metaverse service providers (MSPs) hinder continuous collection of high-quality data. To address these issues, we propose an immersion-aware model trading framework that facilitates data provision for services while ensuring privacy through federated learning (FL). Specifically, we first develop a novel multi-dimensional metric, the immersion of model (IoM), which assesses model value comprehensively by considering freshness and accuracy of learning models, as well as the amount and potential value of raw data used for training. Then, we design an incentive mechanism to incentivize metaverse users (MUs) to contribute high-value models under resource constraints. The trading interactions between MSPs and MUs are modeled as an equilibrium problem with equilibrium constraints (EPEC) to analyze and balance their costs and gains, where MSPs as leaders determine rewards, while MUs as followers optimize resource allocation. Furthermore, considering dynamic network conditions and privacy concerns, we formulate the reward decisions of MSPs as a multi-agent Markov decision process. To solve this, we develop a fully distributed dynamic reward algorithm based on deep reinforcement learning, without accessing any private information about MUs and other MSPs. Experimental results demonstrate that the proposed framework outperforms state-of-the-art benchmarks, achieving improvements in IoM of 38.3% and 37.2%, and reductions in training time to reach the target accuracy of 43.5% and 49.8%, on average, for the MNIST and GTSRB datasets, respectively.
Authors: Marcel Binz, Elif Akata, Matthias Bethge, Franziska Br\"andle, Fred Callaway, Julian Coda-Forno, Peter Dayan, Can Demircan, Maria K. Eckstein, No\'emi \'Eltet\H{o}, Thomas L. Griffiths, Susanne Haridi, Akshay K. Jagadish, Li Ji-An, Alexander Kipnis, Sreejan Kumar, Tobias Ludwig, Marvin Mathony, Marcelo Mattar, Alireza Modirshanechi, Surabhi S. Nath, Joshua C. Peterson, Milena Rmus, Evan M. Russek, Tankred Saanum, Johannes A. Schubert, Luca M. Schulze Buschoff, Nishad Singhi, Xin Sui, Mirko Thalmann, Fabian Theis, Vuong Truong, Vishaal Udandarao, Konstantinos Voudouris, Robert Wilson, Kristin Witte, Shuchen Wu, Dirk Wulff, Huadong Xiong, Eric Schulz
Abstract: Establishing a unified theory of cognition has been a major goal of psychology. While there have been previous attempts to instantiate such theories by building computational models, we currently do not have one model that captures the human mind in its entirety. A first step in this direction is to create a model that can predict human behavior in a wide range of settings. Here we introduce Centaur, a computational model that can predict and simulate human behavior in any experiment expressible in natural language. We derived Centaur by finetuning a state-of-the-art language model on a novel, large-scale data set called Psych-101. Psych-101 reaches an unprecedented scale, covering trial-by-trial data from over 60,000 participants performing over 10,000,000 choices in 160 experiments. Centaur not only captures the behavior of held-out participants better than existing cognitive models, but also generalizes to new cover stories, structural task modifications, and entirely new domains. Furthermore, we find that the model's internal representations become more aligned with human neural activity after finetuning. Taken together, our results demonstrate that it is possible to discover computational models that capture human behavior across a wide range of domains. We believe that such models provide tremendous potential for guiding the development of cognitive theories and present a case study to demonstrate this.
Authors: Hanshi Sun, Li-Wen Chang, Wenlei Bao, Size Zheng, Ningxin Zheng, Xin Liu, Harry Dong, Yuejie Chi, Beidi Chen
Abstract: With the widespread deployment of long-context large language models (LLMs), there has been a growing demand for efficient support of high-throughput inference. However, as the key-value (KV) cache expands with the sequence length, the increasing memory footprint and the need to access it for each token generation both result in low throughput when serving long-context LLMs. While various dynamic sparse attention methods have been proposed to speed up inference while maintaining generation quality, they either fail to sufficiently reduce GPU memory consumption or introduce significant decoding latency by offloading the KV cache to the CPU. We present ShadowKV, a high-throughput long-context LLM inference system that stores the low-rank key cache and offloads the value cache to reduce the memory footprint for larger batch sizes and longer sequences. To minimize decoding latency, ShadowKV employs an accurate KV selection strategy that reconstructs minimal sparse KV pairs on-the-fly. By evaluating ShadowKV on a broad range of benchmarks, including RULER, LongBench, and Needle In A Haystack, and models like Llama-3.1-8B, Llama-3-8B-1M, GLM-4-9B-1M, Yi-9B-200K, Phi-3-Mini-128K, and Qwen2-7B-128K, we demonstrate that it can support up to 6$\times$ larger batch sizes and boost throughput by up to 3.04$\times$ on an A100 GPU without sacrificing accuracy, even surpassing the performance achievable with infinite batch size under the assumption of infinite GPU memory. The code is available at https://github.com/bytedance/ShadowKV.
Authors: Bryan Chan, Xinyi Chen, Andr\'as Gy\"orgy, Dale Schuurmans
Abstract: It has recently been demonstrated empirically that in-context learning emerges in transformers when certain distributional properties are present in the training data, but this ability can also diminish upon further training. We provide a new theoretical understanding of these phenomena by identifying simplified distributional properties that give rise to the emergence and eventual disappearance of in-context learning. We do so by first analyzing a simplified model that uses a gating mechanism to choose between an in-weight and an in-context predictor. Through a combination of a generalization error and regret analysis we identify conditions where in-context and in-weight learning emerge. These theoretical findings are then corroborated experimentally by comparing the behaviour of a full transformer on the simplified distributions to that of the stylized model, demonstrating aligned results. We then extend the study to a full large language model, showing how fine-tuning on various collections of natural language prompts can elicit similar in-context and in-weight learning behaviour.
Authors: Md Abrar Jahin, Md. Akmol Masud, Md Wahiduzzaman Suva, M. F. Mridha, Nilanjan Dey
Abstract: The rapid data surge from the high-luminosity Large Hadron Collider introduces critical computational challenges requiring novel approaches for efficient data processing in particle physics. Quantum machine learning, with its capability to leverage the extensive Hilbert space of quantum hardware, offers a promising solution. However, current quantum graph neural networks (GNNs) lack robustness to noise and are often constrained by fixed symmetry groups, limiting adaptability in complex particle interaction modeling. This paper demonstrates that replacing the Lorentz Group Equivariant Block modules in LorentzNet with a dressed quantum circuit significantly enhances performance despite using nearly 5.5 times fewer parameters. Additionally, quantum circuits effectively replace MLPs by inherently preserving symmetries, with Lorentz symmetry integration ensuring robust handling of relativistic invariance. Our Lorentz-Equivariant Quantum Graph Neural Network (Lorentz-EQGNN) achieved $74.00\%$ test accuracy and an AUC of $87.38\%$ on the Quark-Gluon jet tagging dataset, outperforming the classical and quantum GNNs with a reduced architecture using only 4 qubits. On the Electron-Photon dataset, Lorentz-EQGNN reached $67.00\%$ test accuracy and an AUC of $68.20\%$, demonstrating competitive results with just 800 training samples. Evaluation of our model on generic MNIST and FashionMNIST datasets confirmed Lorentz-EQGNN's efficiency, achieving $88.10\%$ and $74.80\%$ test accuracy, respectively. Ablation studies validated the impact of quantum components on performance, with notable improvements in background rejection rates over classical counterparts. These results highlight Lorentz-EQGNN's potential for immediate applications in noise-resilient jet tagging, event classification, and broader data-scarce HEP tasks.
Authors: Xiaohui Tu, Yossiri Adulyasak, Nima Akbarzadeh, Erick Delage
Abstract: We consider fair resource allocation in sequential decision-making environments modeled as weakly coupled Markov decision processes, where resource constraints couple the action spaces of $N$ sub-Markov decision processes (sub-MDPs) that would otherwise operate independently. We adopt a fairness definition using the generalized Gini function instead of the traditional utilitarian (total-sum) objective. After introducing a general but computationally prohibitive solution scheme based on linear programming, we focus on the homogeneous case where all sub-MDPs are identical. For this case, we show for the first time that the problem reduces to optimizing the utilitarian objective over the class of "permutation invariant" policies. This result is particularly useful as we can exploit Whittle index policies in the restless bandits setting while, for the more general setting, we introduce a count-proportion-based deep reinforcement learning approach. Finally, we validate our theoretical findings with comprehensive experiments, confirming the effectiveness of our proposed method in achieving fairness.
Authors: Yuzong Chen, Ahmed F. AbouElhamayed, Xilai Dai, Yang Wang, Marta Andronic, George A. Constantinides, Mohamed S. Abdelfattah
Abstract: Large language models (LLMs) have demonstrated remarkable performance across various machine learning tasks. Yet the substantial memory footprint of LLMs significantly hinders their deployment. In this paper, we improve the accessibility of LLMs through BitMoD, an algorithm-hardware co-design solution that enables efficient LLM acceleration at low weight precision. On the algorithm side, BitMoD introduces fine-grained data type adaptation that uses a different numerical data type to quantize a group of (e.g., 128) weights. Through the careful design of these new data types, BitMoD is able to quantize LLM weights to very low precision (e.g., 4 bits and 3 bits) while maintaining high accuracy. On the hardware side, BitMoD employs a bit-serial processing element to easily support multiple numerical precisions and data types; our hardware design includes two key innovations: First, it employs a unified representation to process different weight data types, thus reducing the hardware cost. Second, it adopts a bit-serial dequantization unit to rescale the per-group partial sum with minimal hardware overhead. Our evaluation on six representative LLMs demonstrates that BitMoD significantly outperforms state-of-the-art LLM quantization and acceleration methods. For discriminative tasks, BitMoD can quantize LLM weights to 4-bit with $<\!0.5\%$ accuracy loss on average. For generative tasks, BitMoD is able to quantize LLM weights to 3-bit while achieving better perplexity than prior LLM quantization scheme. Combining the superior model performance with an efficient accelerator design, BitMoD achieves an average of $1.69\times$ and $1.48\times$ speedups compared to prior LLM accelerators ANT and OliVe, respectively.
Authors: Jakob Schloer, Matthew Newman, Jannik Thuemmel, Antonietta Capotondi, Bedartha Goswami
Abstract: While deep-learning models have demonstrated skillful El Ni\~no Southern Oscillation (ENSO) forecasts up to one year in advance, they are predominantly trained on climate model simulations that provide thousands of years of training data at the expense of introducing climate model biases. Simpler Linear Inverse Models (LIMs) trained on the much shorter observational record also make skillful ENSO predictions but do not capture predictable nonlinear processes. This motivates a hybrid approach, combining the LIMs modest data needs with a deep-learning non-Markovian correction of the LIM. For O(100 yr) datasets, our resulting Hybrid model is more skillful than the LIM while also exceeding the skill of a full deep-learning model. Additionally, while the most predictable ENSO events are still identified in advance by the LIM, they are better predicted by the Hybrid model, especially in the western tropical Pacific for leads beyond about 9 months, by capturing the subsequent asymmetric (warm versus cold phases) evolution of ENSO.
Authors: Jari Peeperkorn, Simon De Vos
Abstract: Predictive process monitoring focuses on forecasting future states of ongoing process executions, such as predicting the outcome of a particular case. In recent years, the application of machine learning models in this domain has garnered significant scientific attention. When using historical execution data, which may contain biases or exhibit unfair behavior, these biases may be encoded into the trained models. Consequently, when such models are deployed to make decisions or guide interventions for new cases, they risk perpetuating this unwanted behavior. This work addresses group fairness in predictive process monitoring by investigating independence, i.e. ensuring predictions are unaffected by sensitive group membership. We explore independence through metrics for demographic parity such as $\Delta$DP, as well as recently introduced, threshold-independent distribution-based alternatives. Additionally, we propose a composite loss function existing of binary cross-entropy and a distribution-based loss (Wasserstein) to train models that balance predictive performance and fairness, and allow for customizable trade-offs. The effectiveness of both the fairness metrics and the composite loss functions is validated through a controlled experimental setup.
Authors: Taiyan Zhang, Renchi Yang, Yurui Lai, Mingyu Yan, Xiaochun Ye, Dongrui Fan
Abstract: Graph neural networks (GNNs) have become the preferred models for node classification in graph data due to their robust capabilities in integrating graph structures and attributes. However, these models heavily depend on a substantial amount of high-quality labeled data for training, which is often costly to obtain. With the rise of large language models (LLMs), a promising approach is to utilize their exceptional zero-shot capabilities and extensive knowledge for node labeling. Despite encouraging results, this approach either requires numerous queries to LLMs or suffers from reduced performance due to noisy labels generated by LLMs. To address these challenges, we introduce Locle, an active self-training framework that does Label-free node Classification with LLMs cost-Effectively. Locle iteratively identifies small sets of "critical" samples using GNNs and extracts informative pseudo-labels for them with both LLMs and GNNs, serving as additional supervision signals to enhance model training. Specifically, Locle comprises three key components: (i) an effective active node selection strategy for initial annotations; (ii) a careful sample selection scheme to identify "critical" nodes based on label disharmonicity and entropy; and (iii) a label refinement module that combines LLMs and GNNs with a rewired topology. Extensive experiments on five benchmark text-attributed graph datasets demonstrate that Locle significantly outperforms state-of-the-art methods under the same query budget to LLMs in terms of label-free node classification. Notably, on the DBLP dataset with 14.3k nodes, Locle achieves an 8.08% improvement in accuracy over the state-of-the-art at a cost of less than one cent. Our code is available at https://github.com/HKBU-LAGAS/Locle.
Authors: Weijiang Xiong, Robert Fonod, Alexandre Alahi, Nikolas Geroliminis
Abstract: Traffic forecasting is a fundamental task in transportation research, however the scope of current research has mainly focused on a single data modality of loop detectors. Recently, the advances in Artificial Intelligence and drone technologies have made possible novel solutions for efficient, accurate and flexible aerial observations of urban traffic. As a promising traffic monitoring approach, drone-captured data can create an accurate multi-sensor mobility observatory for large-scale urban networks, when combined with existing infrastructure. Therefore, this paper investigates the problem of multi-source traffic speed prediction, simultaneously using drone and loop detector data. A simple yet effective graph-based model HiMSNet is proposed to integrate multiple data modalities and learn spatio-temporal correlations. Detailed analysis shows that predicting accurate segment-level speed is more challenging than the regional speed, especially under high-demand scenarios with heavier congestions and varying traffic dynamics. Utilizing both drone and loop detector data, the prediction accuracy can be improved compared to single-modality cases, when the sensors have lower coverages and are subject to noise. Our simulation study based on vehicle trajectories in a real urban road network has highlighted the added value of integrating drones in traffic forecasting and monitoring.
Authors: G Dhinesh Chandran, Srinivas Reddy Kota, Srikrishna Bhashyam
Abstract: We study the problem of online clustering within the multi-armed bandit framework under the fixed confidence setting. In this multi-armed bandit problem, we have $M$ arms, each providing i.i.d. samples that follow a multivariate Gaussian distribution with an {\em unknown} mean and a known unit covariance. The arms are grouped into $K$ clusters based on the distance between their means using the Single Linkage (SLINK) clustering algorithm on the means of the arms. Since the true means are unknown, the objective is to obtain the above clustering of the arms with the minimum number of samples drawn from the arms, subject to an upper bound on the error probability. We introduce a novel algorithm, Average Tracking Bandit Online Clustering (ATBOC), and prove that this algorithm is order optimal, meaning that the upper bound on its expected sample complexity for given error probability $\delta$ is within a factor of 2 of an instance-dependent lower bound as $\delta \rightarrow 0$. Furthermore, we propose a computationally more efficient algorithm, Lower and Upper Confidence Bound-based Bandit Online Clustering (LUCBBOC), inspired by the LUCB algorithm for best arm identification. Simulation results demonstrate that the performance of LUCBBOC is comparable to that of ATBOC. We numerically assess the effectiveness of the proposed algorithms through numerical experiments on both synthetic datasets and the real-world MovieLens dataset. To the best of our knowledge, this is the first work on bandit online clustering that allows arms with different means in a cluster and $K$ greater than 2.
Authors: Anindita Chattopadhyay, Siddharth Bisariya, Vijay Kumar Sutrakar
Abstract: In the evolving landscape of integrated circuit design, detecting Hardware Trojans (HTs) within a multi entity based design cycle presents significant challenges. This research proposes an innovative machine learning-based methodology for identifying malicious logic gates in gate-level netlists. By focusing on path retrace algorithms. The methodology is validated across three distinct cases, each employing different machine learning models to classify HTs. Case I utilizes a decision tree algorithm for node-to-node comparisons, significantly improving detection accuracy through the integration of Principal Component Analysis (PCA). Case II introduces a graph-to-graph classification using a Graph Neural Network (GNN) model, enabling the differentiation between normal and Trojan-infected circuit designs. Case III applies GNN-based node classification to identify individual compromised nodes and its location. Additionally, nearest neighbor (NN) method has been combined with GNN graph-to-graph in Case II and GNN node-to-node in Case III. Despite the potential of GNN model graph-to-graph classification, NN approach demonstrated superior performance, with the first nearest neighbor (1st NN) achieving 73.2% accuracy and the second nearest neighbor (2nd NN) method reaching 97.7%. In comparison, the GNN model achieved an accuracy of 62.8%. Similarly, GNN model node-to-node classification, NN approach demonstrated superior performance, with the 1st NN achieving 93% accuracy and the 2nd NN method reaching 97.7%. In comparison, the GNN model achieved an accuracy of 79.8%. However, higher and higher NN will lead to large code coverage for the identification of HTs.
Authors: Eirini Schoinas, Adyah Rastogi, Anissa Carter, Jacob Granley, Michael Beyeler
Abstract: Human-in-the-loop optimization (HILO) is a promising approach for personalizing visual prostheses by iteratively refining stimulus parameters based on user feedback. Previous work demonstrated HILO's efficacy in simulation, but its performance with human participants remains untested. Here we evaluate HILO using sighted participants viewing simulated prosthetic vision to assess its ability to optimize stimulation strategies under realistic conditions. Participants selected between phosphenes generated by competing encoders to iteratively refine a deep stimulus encoder (DSE). We tested HILO in three conditions: standard optimization, threshold misspecifications, and out-of-distribution parameter sampling. Participants consistently preferred HILO-generated stimuli over both a naive encoder and the DSE alone, with log odds favoring HILO across all conditions. We also observed key differences between human and simulated decision-making, highlighting the importance of validating optimization strategies with human participants. These findings support HILO as a viable approach for adapting visual prostheses to individuals. Clinical relevance: Validating HILO with sighted participants viewing simulated prosthetic vision is an important step toward personalized calibration of future visual prostheses.
Authors: Chen Fan, Mark Schmidt, Christos Thrampoulidis
Abstract: In the optimization of overparameterized models, different gradient-based methods can achieve zero training error yet converge to distinctly different solutions inducing different generalization properties. Despite a decade of research on implicit optimization bias, important questions remain open even in the foundational case of linear classification with separable data. We address this gap by characterizing the implicit bias of both Adam and Sign gradient descent (SignGD) in multi-class cross-entropy minimization: we prove that their iterates converge to solutions maximizing the margin with respect to the classifier matrix's max-norm, and we establish the corresponding convergence rates. We then generalize our analysis to p-norm normalized steepest descent (NSD) algorithms. This includes Spectral Descent, which we show converges to the max-margin solution with respect to the spectral norm. A key insight is that the analysis of general entry-wise and Schatten p-norms can be reduced to the analysis of NSD with max-norm (i.e., SignGD) by exploiting a natural ordering property between all p-norms relative to the max-norm and its dual sum-norm. Our results demonstrate that the multi-class linear setting, which is inherently richer than the binary counterpart, provides the most transparent playground for studying implicit biases of matrix-parameter optimization algorithms.
Authors: Georgios Syros, Anshuman Suri, Farinaz Koushanfar, Cristina Nita-Rotaru, Alina Oprea
Abstract: Federated Learning is vulnerable to adversarial manipulation, where malicious clients can inject poisoned updates to influence the global model's behavior. While existing defense mechanisms have made notable progress, they fail to protect against adversaries that aim to induce targeted backdoors under different learning and attack configurations. To address this limitation, we introduce DROP (Distillation-based Reduction Of Poisoning), a novel defense mechanism that combines clustering and activity-tracking techniques with extraction of benign behavior from clients via knowledge distillation to tackle stealthy adversaries that manipulate low data poisoning rates and diverse malicious client ratios within the federation. Through extensive experimentation, our approach demonstrates superior robustness compared to existing defenses across a wide range of learning configurations. Finally, we evaluate existing defenses and our method under the challenging setting of non-IID client data distribution and highlight the challenges of designing a resilient FL defense in this setting.
Authors: Evgeniia Tokarchuk, Hua Chang Bakker, Vlad Niculae
Abstract: Learning well-separated features in high-dimensional spaces, such as text or image embeddings, is crucial for many machine learning applications. Achieving such separation can be effectively accomplished through the dispersion of embeddings, where unrelated vectors are pushed apart as much as possible. By constraining features to be on a hypersphere, we can connect dispersion to well-studied problems in mathematics and physics, where optimal solutions are known for limited low-dimensional cases. However, in representation learning we typically deal with a large number of features in high-dimensional space, and moreover, dispersion is usually traded off with some other task-oriented training objective, making existing theoretical and numerical solutions inapplicable. Therefore, it is common to rely on gradient-based methods to encourage dispersion, usually by minimizing some function of the pairwise distances. In this work, we first give an overview of existing methods from disconnected literature, making new connections and highlighting similarities. Next, we introduce some new angles. We propose to reinterpret pairwise dispersion using a maximum mean discrepancy (MMD) motivation. We then propose an online variant of the celebrated Lloyd's algorithm, of K-Means fame, as an effective alternative regularizer for dispersion on generic domains. Finally, we derive a novel dispersion method that directly exploits properties of the hypersphere. Our experiments show the importance of dispersion in image classification and natural language processing tasks, and how algorithms exhibit different trade-offs in different regimes.
Authors: John D. Jakeman, Lorena A. Barba, Joaquim R. R. A. Martins, Thomas O'Leary-Roseberry
Abstract: Scientific machine learning (SciML) models are transforming many scientific disciplines. However, the development of good modeling practices to increase the trustworthiness of SciML has lagged behind its application, limiting its potential impact. The goal of this paper is to start a discussion on establishing consensus-based good practices for predictive SciML. We identify key challenges in applying existing computational science and engineering guidelines, such as verification and validation protocols, and provide recommendations to address these challenges. Our discussion focuses on predictive SciML, which uses machine learning models to learn, improve, and accelerate numerical simulations of physical systems. While centered on predictive applications, our 16 recommendations aim to help researchers conduct and document their modeling processes rigorously across all SciML domains.
Authors: Juan L. Gamella, Simon Bing, Jakob Runge
Abstract: We evaluate methods for causal representation learning (CRL) on a simple, real-world system where these methods are expected to work. The system consists of a controlled optical experiment specifically built for this purpose, which satisfies the core assumptions of CRL and where the underlying causal factors (the inputs to the experiment) are known, providing a ground truth. We select methods representative of different approaches to CRL and find that they all fail to recover the underlying causal factors. To understand the failure modes of the evaluated algorithms, we perform an ablation on the data by substituting the real data-generating process with a simpler synthetic equivalent. The results reveal a reproducibility problem, as most methods already fail on this synthetic ablation despite its simple data-generating process. Additionally, we observe that common assumptions on the mixing function are crucial for the performance of some of the methods but do not hold in the real data. Our efforts highlight the contrast between the theoretical promise of the state of the art and the challenges in its application. We hope the benchmark serves as a simple, real-world sanity check to further develop and validate methodology, bridging the gap towards CRL methods that work in practice. We make all code and datasets publicly available at github.com/simonbing/CRLSanityCheck
Authors: Md Abrar Jahin, Shahriar Soudeep, Fahmid Al Farid, M. F. Mridha, Raihan Kabir, Md Rashedul Islam, Hezerul Abdul Karim
Abstract: Cybersecurity threats are growing, making network intrusion detection essential. Traditional machine learning models remain effective in resource-limited environments due to their efficiency, requiring fewer parameters and less computational time. However, handling short and highly imbalanced datasets remains challenging. In this study, we propose the fusion of a Contrastive Attentive Graph Network and Graph Attention Network (CAGN-GAT Fusion) and benchmark it against 15 other models, including both Graph Neural Networks (GNNs) and traditional ML models. Our evaluation is conducted on four benchmark datasets (KDD-CUP-1999, NSL-KDD, UNSW-NB15, and CICIDS2017) using a short and proportionally imbalanced dataset with a constant size of 5000 samples to ensure fairness in comparison. Results show that CAGN-GAT Fusion demonstrates stable and competitive accuracy, recall, and F1-score, even though it does not achieve the highest performance in every dataset. Our analysis also highlights the impact of adaptive graph construction techniques, including small changes in connections (edge perturbation) and selective hiding of features (feature masking), improving detection performance. The findings confirm that GNNs, particularly CAGN-GAT Fusion, are robust and computationally efficient, making them well-suited for resource-constrained environments. Future work will explore GraphSAGE layers and multiview graph construction techniques to further enhance adaptability and detection accuracy.
Authors: Chuanyin Wang, Yifei Zhang, Neng Gao, Qiang Luo
Abstract: Personalized federated learning is extensively utilized in scenarios characterized by data heterogeneity, facilitating more efficient and automated local training on data-owning terminals. This includes the automated selection of high-performance model parameters for upload, thereby enhancing the overall training process. However, it entails significant risks of privacy leakage. Existing studies have attempted to mitigate these risks by utilizing differential privacy. Nevertheless, these studies present two major limitations: (1) The integration of differential privacy into personalized federated learning lacks sufficient personalization, leading to the introduction of excessive noise into the model. (2) It fails to adequately control the spatial scope of model update information, resulting in a suboptimal balance between data privacy and model effectiveness in differential privacy federated learning. In this paper, we propose a differentially private personalized federated learning approach that employs dynamically sparsified client updates through reparameterization and adaptive norm(DP-pFedDSU). Reparameterization training effectively selects personalized client update information, thereby reducing the quantity of updates. This approach minimizes the introduction of noise to the greatest extent possible. Additionally, dynamic adaptive norm refers to controlling the norm space of model updates during the training process, mitigating the negative impact of clipping on the update information. These strategies substantially enhance the effective integration of differential privacy and personalized federated learning. Experimental results on EMNIST, CIFAR-10, and CIFAR-100 demonstrate that our proposed scheme achieves superior performance and is well-suited for more complex personalized federated learning scenarios.
Authors: Lauren Harrell, Christine Kaeser-Chen, Burcu Karagol Ayan, Keith Anderson, Michelangelo Conserva, Elise Kleeman, Maxim Neumann, Matt Overlan, Melissa Chapman, Drew Purves
Abstract: Species distribution models (SDMs) are necessary for measuring and predicting occurrences and habitat suitability of species and their relationship with environmental factors. We introduce a novel presence-only SDM with graph neural networks (GNN). In our model, species and locations are treated as two distinct node sets, and the learning task is predicting detection records as the edges that connect locations to species. Using GNN for SDM allows us to model fine-grained interactions between species and the environment. We evaluate the potential of this methodology on the six-region dataset compiled by National Center for Ecological Analysis and Synthesis (NCEAS) for benchmarking SDMs. For each of the regions, the heterogeneous GNN model is comparable to or outperforms previously-benchmarked single-species SDMs as well as a feed-forward neural network baseline model.
Authors: Liying Han, Gaofeng Dong, Xiaomin Ouyang, Lance Kaplan, Federico Cerutti, Mani Srivastava
Abstract: Complex events (CEs) play a crucial role in CPS-IoT applications, enabling high-level decision-making in domains such as smart monitoring and autonomous systems. However, most existing models focus on short-span perception tasks, lacking the long-term reasoning required for CE detection. CEs consist of sequences of short-time atomic events (AEs) governed by spatiotemporal dependencies. Detecting them is difficult due to long, noisy sensor data and the challenge of filtering out irrelevant AEs while capturing meaningful patterns. This work explores CE detection as a case study for CPS-IoT foundation models capable of long-term reasoning. We evaluate three approaches: (1) leveraging large language models (LLMs), (2) employing various neural architectures that learn CE rules from data, and (3) adopting a neurosymbolic approach that integrates neural models with symbolic engines embedding human knowledge. Our results show that the state-space model, Mamba, which belongs to the second category, outperforms all methods in accuracy and generalization to longer, unseen sensor traces. These findings suggest that state-space models could be a strong backbone for CPS-IoT foundation models for long-span reasoning tasks.
Authors: Kai Uwe Barthel, Florian Barthel, Peter Eisert
Abstract: Sorting and permutation learning are key concepts in optimization and machine learning, especially when organizing high-dimensional data into meaningful spatial layouts. The Gumbel-Sinkhorn method, while effective, requires N*N parameters to determine a full permutation matrix, making it computationally expensive for large datasets. Low-rank matrix factorization approximations reduce memory requirements to 2NM (with M << N), but they still struggle with very large problems. SoftSort, by providing a continuous relaxation of the argsort operator, allows differentiable 1D sorting, but it faces challenges with multidimensional data and complex permutations. In this paper, we present a novel method for learning permutations using only N parameters, which dramatically reduces storage costs. Our method extends SoftSort by iteratively shuffling the N indices of the elements and applying a few SoftSort optimization steps per iteration. This modification significantly improves sorting quality, especially for multidimensional data and complex optimization criteria, and outperforms pure SoftSort. Our method offers improved memory efficiency and scalability compared to existing approaches, while maintaining high-quality permutation learning. Its dramatically reduced memory requirements make it particularly well-suited for large-scale optimization tasks, such as "Self-Organizing Gaussians", where efficient and scalable permutation learning is critical.
Authors: Ari Blondal, Hamed Hatami, Pooya Hatami, Chavdar Lalov, Sivan Tretiak
Abstract: Recent remarkable advances in learning theory have established that, for total concept classes, list replicability, global stability, differentially private (DP) learnability, and shared-randomness replicability all coincide with the finiteness of Littlestone dimension. Does this equivalence extend to partial concept classes? We answer this question by proving that the list replicability number of $d$-dimensional $\gamma$-margin half-spaces satisfies \[ \frac{d}{2}+1 \le \mathrm{LR}(H^d_\gamma) \le d, \] which grows with dimension. Consequently, for partial classes, list replicability and global stability do not necessarily follow from bounded Littlestone dimension, pure DP-learnability, or shared-randomness replicability. Applying our main theorem, we resolve several open problems: $\bullet$ Every disambiguation of infinite-dimensional large-margin half-spaces to a total concept class has unbounded Littlestone dimension, answering an open question of Alon et al. (FOCS '21). $\bullet$ The maximum list-replicability number of any finite set of points and homogeneous half-spaces in $d$-dimensional Euclidean space is $d$, resolving a problem of Chase et al. (FOCS '23). $\bullet$ Every disambiguation of the Gap Hamming Distance problem in the large gap regime has unbounded public-coin randomized communication complexity. This answers an open question of Fang et al. (STOC '25). $\bullet$ There exists a partial concept class with Littlestone dimension $1$ such that all its disambiguations have infinite Littlestone dimension. This answers a question of Cheung et al. (ICALP '23). Our lower bound follows from a topological argument based on the local Borsuk-Ulam theorem of Chase, Chornomaz, Moran, and Yehudayoff (STOC '24). For the upper bound, we construct a list-replicable learning rule using the generalization properties of SVMs.
Authors: Qingsong Wang
Abstract: Symmetric matrix decomposition is an active research area in machine learning. This paper focuses on exploiting the low-rank structure of non-negative and sparse symmetric matrices via the rectified linear unit (ReLU) activation function. We propose the ReLU-based nonlinear symmetric matrix decomposition (ReLU-NSMD) model, introduce an accelerated alternating partial Bregman (AAPB) method for its solution, and present the algorithm's convergence results. Our algorithm leverages the Bregman proximal gradient framework to overcome the challenge of estimating the global $L$-smooth constant in the classic proximal gradient algorithm. Numerical experiments on synthetic and real datasets validate the effectiveness of our model and algorithm.
Authors: Leander Kurscheidt, Paolo Morettin, Roberto Sebastiani, Andrea Passerini, Antonio Vergari
Abstract: In safety-critical applications, guaranteeing the satisfaction of constraints over continuous environments is crucial, e.g., an autonomous agent should never crash into obstacles or go off-road. Neural models struggle in the presence of these constraints, especially when they involve intricate algebraic relationships. To address this, we introduce a differentiable probabilistic layer that guarantees the satisfaction of non-convex algebraic constraints over continuous variables. This probabilistic algebraic layer (PAL) can be seamlessly plugged into any neural architecture and trained via maximum likelihood without requiring approximations. PAL defines a distribution over conjunctions and disjunctions of linear inequalities, parameterized by polynomials. This formulation enables efficient and exact renormalization via symbolic integration, which can be amortized across different data points and easily parallelized on a GPU. We showcase PAL and our integration scheme on a number of benchmarks for algebraic constraint integration and on real-world trajectory data.
Authors: Zhihan Zhang, Xunkai Li, Guang Zeng, Hongchao Qin, Ronghua Li, Guoren Wang
Abstract: Recently, the emergence of LLMs has prompted researchers to integrate language descriptions into graphs, aiming to enhance model encoding capabilities from a data-centric perspective. This graph representation is called text-attributed graphs (TAGs). A review of prior advancements highlights that graph structure learning (GSL) is a pivotal technique for improving data utility, making it highly relevant to efficient TAG learning. However, most GSL methods are tailored for traditional graphs without textual information, underscoring the necessity of developing a new GSL paradigm. Despite clear motivations, it remains challenging: (1) How can we define a reasonable optimization objective for GSL in the era of LLMs, considering the massive parameters in LLM? (2) How can we design an efficient model architecture that enables seamless integration of LLM for this optimization objective? For Question 1, we reformulate existing GSL optimization objectives as a tree optimization framework, shifting the focus from obtaining a well-trained edge predictor to a language-aware tree sampler. For Question 2, we propose decoupled and training-free model design principles for LLM integration, shifting the focus from computation-intensive fine-tuning to more efficient inference. Based on this, we propose Large Language and Tree Assistant (LLaTA), which leverages tree-based LLM in-context learning to enhance the understanding of topology and text, enabling reliable inference and generating improved graph structure. Extensive experiments on 10 datasets demonstrate that LLaTA enjoys flexibility-incorporated with any backbone; scalability-outperforms other LLM-enhanced graph learning methods; effectiveness-achieves SOTA predictive performance.
Authors: Fadi Alharbi, Nishant Budhiraja, Aleksandar Vakanski, Boyu Zhang, Murtada K. Elbashir, Hrshith Gudur, Mohanad Mohammed
Abstract: The integration of heterogeneous multi-omics datasets at a systems level remains a central challenge for developing analytical and computational models in precision cancer diagnostics. This paper introduces Multi-Omics Graph Kolmogorov-Arnold Network (MOGKAN), a deep learning framework that utilizes messenger-RNA, micro-RNA sequences, and DNA methylation samples together with Protein-Protein Interaction (PPI) networks for cancer classification across 31 different cancer types. The proposed approach combines differential gene expression with DESeq2, Linear Models for Microarray (LIMMA), and Least Absolute Shrinkage and Selection Operator (LASSO) regression to reduce multi-omics data dimensionality while preserving relevant biological features. The model architecture is based on the Kolmogorov-Arnold theorem principle and uses trainable univariate functions to enhance interpretability and feature analysis. MOGKAN achieves classification accuracy of 96.28 percent and exhibits low experimental variability in comparison to related deep learning-based models. The biomarkers identified by MOGKAN were validated as cancer-related markers through Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis. By integrating multi-omics data with graph-based deep learning, our proposed approach demonstrates robust predictive performance and interpretability with potential to enhance the translation of complex multi-omics data into clinically actionable cancer diagnostics.
Authors: Shengrui XU, Tianchi Lu, Zikun Wang, Jixiu Zhai
Abstract: Protein-protein interaction (PPI) prediction plays a pivotal role in deciphering cellular functions and disease mechanisms. To address the limitations of traditional experimental methods and existing computational approaches in cross-modal feature fusion and false-negative suppression, we propose SCMPPI-a novel supervised contrastive multimodal framework. By effectively integrating sequence-based features (AAC, DPC, ESMC-CKSAAP) with network topology (Node2Vec embeddings) and incorporating an enhanced contrastive learning strategy with negative sample filtering, SCMPPI achieves superior prediction performance. Extensive experiments on eight benchmark datasets demonstrate its state-of-the-art accuracy(98.13%) and AUC(99.69%), along with excellent cross-species generalization (AUC>99%). Successful applications in CD9 networks, Wnt pathway analysis, and cancer-specific networks further highlight its potential for disease target discovery, establishing SCMPPI as a powerful tool for multimodal biological data analysis.
Authors: Sharan Sahu
Abstract: We study regret minimization under privacy constraints in episodic inhomogeneous linear Markov Decision Processes (MDPs), motivated by the growing use of reinforcement learning (RL) in personalized decision-making systems that rely on sensitive user data. In this setting, both transition probabilities and reward functions are assumed to be linear in a feature mapping $\phi(s, a)$, and we aim to ensure privacy through joint differential privacy (JDP), a relaxation of differential privacy suited to online learning. Prior work has established suboptimal regret bounds by privatizing the LSVI-UCB algorithm, which achieves $\widetilde{O}(\sqrt{d^3 H^4 K})$ regret in the non-private setting. Building on recent advances that improve this to near minimax optimal regret $\widetilde{O}(d\sqrt{H^{3}K})$ via LSVI-UCB++ with Bernstein-style bonuses, we design a new differentially private algorithm by privatizing LSVI-UCB++ and adapting techniques for variance-aware analysis from offline RL. Our algorithm achieves a regret bound of $\widetilde{O}(d \sqrt{H^3 K} + H^{15/4} d^{7/6} K^{1/2} / \epsilon)$, improving over previous private methods. Empirical results show that our algorithm retains near-optimal utility compared to non-private baselines, indicating that privacy can be achieved with minimal performance degradation in this setting.
Authors: Changming Xu, Debangshu Banerjee, Deepak Vasisht, Gagandeep Singh
Abstract: Variational Autoencoders (VAEs) have become increasingly popular and deployed in safety-critical applications. In such applications, we want to give certified probabilistic guarantees on performance under adversarial attacks. We propose a novel method, CIVET, for certified training of VAEs. CIVET depends on the key insight that we can bound worst-case VAE error by bounding the error on carefully chosen support sets at the latent layer. We show this point mathematically and present a novel training algorithm utilizing this insight. We show in an extensive evaluation across different datasets (in both the wireless and vision application areas), architectures, and perturbation magnitudes that our method outperforms SOTA methods achieving good standard performance with strong robustness guarantees.
Authors: Hyunji Jung, Hanseul Cho, Chulhee Yun
Abstract: We study continual learning on multiple linear classification tasks by sequentially running gradient descent (GD) for a fixed budget of iterations per task. When all tasks are jointly linearly separable and are presented in a cyclic/random order, we show the directional convergence of the trained linear classifier to the joint (offline) max-margin solution. This is surprising because GD training on a single task is implicitly biased towards the individual max-margin solution for the task, and the direction of the joint max-margin solution can be largely different from these individual solutions. Additionally, when tasks are given in a cyclic order, we present a non-asymptotic analysis on cycle-averaged forgetting, revealing that (1) alignment between tasks is indeed closely tied to catastrophic forgetting and backward knowledge transfer and (2) the amount of forgetting vanishes to zero as the cycle repeats. Lastly, we analyze the case where the tasks are no longer jointly separable and show that the model trained in a cyclic order converges to the unique minimum of the joint loss function.
Authors: Yaoyao Ding, Bohan Hou, Xiao Zhang, Allan Lin, Tianqi Chen, Cody Yu Hao, Yida Wang, Gennady Pekhimenko
Abstract: Serving Large Language Models (LLMs) is critical for AI-powered applications but demands substantial computational resources, particularly in memory bandwidth and computational throughput. Low-precision computation has emerged as a key technique to improve efficiency while reducing resource consumption. Existing approaches for generating low-precision kernels are limited to weight bit widths that are powers of two and suffer from suboptimal performance due to high-level GPU programming abstractions. These abstractions restrict critical optimizations, such as fine-grained register management and optimized memory access patterns, which are essential for efficient low-precision computations. In this paper, we introduce a virtual machine (VM) designed for General-Purpose GPU (GPGPU) computing, enabling support for low-precision data types with arbitrary bit widths while maintaining GPU programmability. The proposed VM features a thread-block-level programming model, a hierarchical memory space, a novel algebraic layout system, and extensive support for diverse low-precision data types. VM programs are compiled into highly efficient GPU programs with automatic vectorization and instruction selection. Extensive experiments demonstrate that our VM efficiently supports a full spectrum of low-precision data types, and outperforms state-of-the-art low-precision kernels on their supported types. Compared to existing compilers like Triton and Ladder, as well as hand-optimized kernels such as QuantLLM and Marlin, our VM achieves performance improvements of 1.75x, 2.61x, 1.29x and 1.03x, respectively.
Authors: Wenxin Zhang, Jingxing Zhong, Guangzhen Yao, Renda Han, Xiaojian Lin, Zeyu Zhang, Cuicui Luo
Abstract: Fraudulent activities have significantly increased across various domains, such as e-commerce, online review platforms, and social networks, making fraud detection a critical task. Spatial Graph Neural Networks (GNNs) have been successfully applied to fraud detection tasks due to their strong inductive learning capabilities. However, existing spatial GNN-based methods often enhance the graph structure by excluding heterophilic neighbors during message passing to align with the homophilic bias of GNNs. Unfortunately, this approach can disrupt the original graph topology and increase uncertainty in predictions. To address these limitations, this paper proposes a novel framework, Dual-channel Heterophilic Message Passing (DHMP), for fraud detection. DHMP leverages a heterophily separation module to divide the graph into homophilic and heterophilic subgraphs, mitigating the low-pass inductive bias of traditional GNNs. It then applies shared weights to capture signals at different frequencies independently and incorporates a customized sampling strategy for training. This allows nodes to adaptively balance the contributions of various signals based on their labels. Extensive experiments on three real-world datasets demonstrate that DHMP outperforms existing methods, highlighting the importance of separating signals with different frequencies for improved fraud detection. The code is available at https://github.com/shaieesss/DHMP.
Authors: Steve Hanneke, Shay Moran, Alexander Shlimovich, Amir Yehudayoff
Abstract: Learning theory has traditionally followed a model-centric approach, focusing on designing optimal algorithms for a fixed natural learning task (e.g., linear classification or regression). In this paper, we adopt a complementary data-centric perspective, whereby we fix a natural learning rule and focus on optimizing the training data. Specifically, we study the following question: given a learning rule $\mathcal{A}$ and a data selection budget $n$, how well can $\mathcal{A}$ perform when trained on at most $n$ data points selected from a population of $N$ points? We investigate when it is possible to select $n \ll N$ points and achieve performance comparable to training on the entire population. We address this question across a variety of empirical risk minimizers. Our results include optimal data-selection bounds for mean estimation, linear classification, and linear regression. Additionally, we establish two general results: a taxonomy of error rates in binary classification and in stochastic convex optimization. Finally, we propose several open questions and directions for future research.
Authors: Muhammad Qasim Elahi, Somtochukwu Oguchienti, Maheed H. Ahmed, Mahsa Ghasemi
Abstract: Designing an effective reward function has long been a challenge in reinforcement learning, particularly for complex tasks in unstructured environments. To address this, various learning paradigms have emerged that leverage different forms of human input to specify or refine the reward function. Reinforcement learning from human feedback is a prominent approach that utilizes human comparative feedback, expressed as a preference for one behavior over another, to tackle this problem. In contrast to comparative feedback, we explore multi-level human feedback, which is provided in the form of a score at the end of each episode. This type of feedback offers more coarse but informative signals about the underlying reward function than binary feedback. Additionally, it can handle non-Markovian rewards, as it is based on the evaluation of an entire episode. We propose an algorithm to efficiently learn both the reward function and the optimal policy from this form of feedback. Moreover, we show that the proposed algorithm achieves sublinear regret and demonstrate its empirical effectiveness through extensive simulations.
Authors: Kazuhisa Fujita
Abstract: Artificial neural networks are powerful tools capable of addressing various tasks. Although the backpropagation algorithm has become a standard training method for these neural networks, its lack of biological plausibility has inspired the development of alternative learning approaches. One such alternative is Kaneko's Error Diffusion Learning Algorithm (EDLA), a biologically motivated approach wherein a single global error signal diffuses throughout a network composed of paired excitatory-inhibitory sublayers, thereby eliminating the necessity for layer-wise backpropagation. This study presents a contemporary formulation of the EDLA framework and evaluates its effectiveness through parity check, regression, and image classification tasks. Our experimental results indicate that EDLA networks can consistently achieve high accuracy across these benchmarks, with performance efficiency and convergence speed notably influenced by the choice of learning rate, neuron count, and network depth. Further investigation of the internal representations formed by EDLA networks reveals their capacity for meaningful feature extraction, similar to traditional neural networks. These results suggest that EDLA is a biologically motivated alternative for training feedforward networks and will motivate future work on extending this method to biologically inspired neural networks.
Authors: Simone Papicchio, Simone Rossi, Luca Cagliero, Paolo Papotti
Abstract: Large Language Models (LLMs) have shown impressive capabilities in transforming natural language questions about relational databases into SQL queries. Despite recent improvements, small LLMs struggle to handle questions involving multiple tables and complex SQL patterns under a Zero-Shot Learning (ZSL) setting. Supervised Fine-Tuning (SFT) partially compensates for the knowledge deficits in pretrained models but falls short while dealing with queries involving multi-hop reasoning. To bridge this gap, different LLM training strategies to reinforce reasoning capabilities have been proposed, ranging from leveraging a thinking process within ZSL, including reasoning traces in SFT, or adopt Reinforcement Learning (RL) strategies. However, the influence of reasoning on Text2SQL performance is still largely unexplored. This paper investigates to what extent LLM reasoning capabilities influence their Text2SQL performance on four benchmark datasets. To this end, it considers the following LLM settings: (1) ZSL, including general-purpose reasoning or not; (2) SFT, with and without task-specific reasoning traces; (3) RL, exploring the use of different rewarding functions, both the established EXecution accuracy (EX) and a mix with fine-grained ones that also account the precision, recall, and cardinality of partially correct answers; (4) SFT+RL, i.e, a two-stage approach that combines SFT and RL. The results show that general-purpose reasoning under ZSL proves to be ineffective in tackling complex Text2SQL cases. Small LLMs benefit from SFT with reasoning much more than larger ones. RL is generally beneficial across all tested models and datasets. The use of the fine-grained metrics turns out to be the most effective RL strategy. Thanks to RL and the novel text2SQL rewards, the 7B Qwen-Coder-2.5 model performs on par with 400+ Billion ones (including gpt-4o) on the Bird dataset.
Authors: Rahul Vishwakarma, Shrey Dharmendra Modi, Vishwanath Seshagiri
Abstract: The generation of high-quality synthetic data presents significant challenges in machine learning research, particularly regarding statistical fidelity and uncertainty quantification. Existing generative models produce compelling synthetic samples but lack rigorous statistical guarantees about their relation to the underlying data distribution, limiting their applicability in critical domains requiring robust error bounds. We address this fundamental limitation by presenting a novel framework that incorporates conformal prediction methodologies into Generative Adversarial Networks (GANs). By integrating multiple conformal prediction paradigms including Inductive Conformal Prediction (ICP), Mondrian Conformal Prediction, Cross-Conformal Prediction, and Venn-Abers Predictors, we establish distribution-free uncertainty quantification in generated samples. This approach, termed Conformalized GAN (cGAN), demonstrates enhanced calibration properties while maintaining the generative power of traditional GANs, producing synthetic data with provable statistical guarantees. We provide rigorous mathematical proofs establishing finite-sample validity guarantees and asymptotic efficiency properties, enabling the reliable application of synthetic data in high-stakes domains including healthcare, finance, and autonomous systems.
Authors: Radia Berreziga, Mohammed Brahimi, Khairedine Kraim, Hamid Azzoune
Abstract: Virtual screening plays a critical role in modern drug discovery by enabling the identification of promising candidate molecules for experimental validation. Traditional machine learning methods such, as Support Vector Machines (SVM) and XGBoost, rely on predefined molecular representations, often leading to information loss and potential bias. In contrast, deep learning approaches-particularly Graph Convolutional Networks (GCNs)-offer a more expressive and unbiased alternative by operating directly on molecular graphs. Meanwhile, Large Language Models (LLMs) have recently demonstrated state-of-the-art performance in drug design, thanks to their capacity to capture complex chemical patterns from large-scale data via attention mechanisms. In this paper, we propose a hybrid architecture that integrates GCNs with LLM-derived embeddings to combine localized structural learning with global chemical knowledge. The LLM embeddings can be precomputed and stored in a molecular feature library, removing the need to rerun the LLM during training or inference and thus maintaining computational efficiency. We found that concatenating the LLM embeddings after each GCN layer-rather than only at the final layer-significantly improves performance, enabling deeper integration of global context throughout the network. The resulting model achieves superior results, with an F1-score of (88.8\%), outperforming standalone GCN (87.9%), XGBoost (85.5%), and SVM (85.4%) baselines.
Authors: Lei Wang, Yu Cheng, Yining Shi, Zhengju Tang, Zhiwen Mo, Wenhao Xie, Lingxiao Ma, Yuqing Xia, Jilong Xue, Fan Yang, Zhi Yang
Abstract: Modern AI workloads rely heavily on optimized computing kernels for both training and inference. These AI kernels follow well-defined data-flow patterns, such as moving tiles between DRAM and SRAM and performing a sequence of computations on those tiles. However, writing high-performance kernels remains complex despite the clarity of these patterns. Achieving peak performance requires careful, hardware-centric optimizations to fully leverage modern accelerators. While domain-specific compilers attempt to reduce the burden of writing high-performance kernels, they often struggle with usability and expressiveness gaps. In this paper, we present TileLang, a generalized tiled programming model for more efficient AI Kernel programming. TileLang decouples scheduling space (thread binding, layout, tensorize and pipeline) from dataflow, and encapsulated them as a set of customization annotations and primitives. This approach allows users to focus on the kernel's data-flow itself, while leaving most other optimizations to compilers. We conduct comprehensive experiments on commonly-used devices, across numerous experiments, our evaluation shows that TileLang can achieve state-of-the-art performance in key kernels, demonstrating that its unified block-and-thread paradigm and transparent scheduling capabilities deliver both the power and flexibility demanded by modern AI system development.
Authors: Fan Yang, Hongyang R. Zhang, Sen Wu, Christopher R\'e, Weijie J. Su
Abstract: The problem of learning one task with samples from another task is central to transfer learning (TL). In this paper, we examine a fundamental question: When does combining the data samples from a source task and a target task perform better than single-task learning with the target task alone? This question is motivated by an intriguing phenomenon known as negative transfer often observed in the TL literature. Precise quantification of TL effects -- even within simple statistical models -- has remained elusive in the statistical learning literature. A critical challenge is that to compare TL to single-task learning, we would need to compare the risks between two different estimators in a very precise way. In particular, the comparative advantage of one estimator over another would depend on the specific distribution shifts between the two tasks. This paper applies recent developments in the random matrix theory literature to tackle this challenge in a high-dimensional linear regression setting with two tasks. We provide precise high-dimensional asymptotics for the bias and variance of hard parameter sharing (HPS) estimators in the proportional limit, when the sample sizes of both tasks increase proportionally with dimension at fixed ratios. The precise asymptotics are expressed as a function of the sample sizes of both tasks, the covariate shift between their feature population covariate matrices, and the model shift. We provide illustrative examples of our results in a random-effects model to determine positive and negative transfers. For example, we can identify a phase transition in the high-dimensional linear regression setting from positive transfer to negative transfer under a model shift between the source and target tasks. The finding regarding phase transition can be extended to a multiple-task learning setting where the feature covariates are shared across all tasks.
Authors: P. Schulze, S. Wiegrebe, P. W. Thurner, C. Heumann, M. A{\ss}enmacher
Abstract: The objective of advanced topic modeling is not only to explore latent topical structures, but also to estimate relationships between the discovered topics and theoretically relevant metadata. Methods used to estimate such relationships must take into account that the topical structure is not directly observed, but instead being estimated itself in an unsupervised fashion, usually by common topic models. A frequently used procedure to achieve this is the method of composition, a Monte Carlo sampling technique performing multiple repeated linear regressions of sampled topic proportions on metadata covariates. In this paper, we propose two modifications of this approach: First, we substantially refine the existing implementation of the method of composition from the R package stm by replacing linear regression with the more appropriate Beta regression. Second, we provide a fundamental enhancement of the entire estimation framework by substituting the current blending of frequentist and Bayesian methods with a fully Bayesian approach. This allows for a more appropriate quantification of uncertainty. We illustrate our improved methodology by investigating relationships between Twitter posts by German parliamentarians and different metadata covariates related to their electoral districts, using the Structural Topic Model to estimate topic proportions.
Authors: Kawin Ethayarajh, Yejin Choi, Swabha Swayamdipta
Abstract: Estimating the difficulty of a dataset typically involves comparing state-of-the-art models to humans; the bigger the performance gap, the harder the dataset is said to be. However, this comparison provides little understanding of how difficult each instance in a given distribution is, or what attributes make the dataset difficult for a given model. To address these questions, we frame dataset difficulty -- w.r.t. a model $\mathcal{V}$ -- as the lack of $\mathcal{V}$-$\textit{usable information}$ (Xu et al., 2019), where a lower value indicates a more difficult dataset for $\mathcal{V}$. We further introduce $\textit{pointwise $\mathcal{V}$-information}$ (PVI) for measuring the difficulty of individual instances w.r.t. a given distribution. While standard evaluation metrics typically only compare different models for the same dataset, $\mathcal{V}$-$\textit{usable information}$ and PVI also permit the converse: for a given model $\mathcal{V}$, we can compare different datasets, as well as different instances/slices of the same dataset. Furthermore, our framework allows for the interpretability of different input attributes via transformations of the input, which we use to discover annotation artefacts in widely-used NLP benchmarks.
Authors: Armin Moin, Ukrit Wattanavaekin, Alexandra Lungu, Stephan R\"ossler, Stephan G\"unnemann
Abstract: We propose a novel approach to enable Automated Machine Learning (AutoML) for Non-Intrusive Appliance Load Monitoring (NIALM), also known as Energy Disaggregation, through Bayesian Optimization. NIALM offers a cost-effective alternative to smart meters for measuring the energy consumption of electric devices and appliances. NIALM methods analyze the entire power consumption signal of a household and predict the type of appliances as well as their individual power consumption (i.e., their contributions to the aggregated signal). We enable NIALM domain experts and practitioners who typically have no deep data analytics or Machine Learning (ML) skills to benefit from state-of-the-art ML approaches to NIALM. Further, we conduct a survey and benchmarking of the state of the art and show that in many cases, simple and basic ML models and algorithms, such as Decision Trees, outperform the state of the art. Finally, we present our open-source tool, AutoML4NIALM, which will facilitate the exploitation of existing methods for NIALM in the industry.
Authors: Amogh Joshi, Adarsh Kosta, Wachirawit Ponghiran, Manish Nagaraj, Kaushik Roy
Abstract: The ability of resource-constrained biological systems such as fruitflies to perform complex and high-speed maneuvers in cluttered environments has been one of the prime sources of inspiration for developing vision-based autonomous systems. To emulate this capability, the perception pipeline of such systems must integrate information cues from tasks including optical flow and depth estimation, object detection and tracking, and segmentation, among others. However, the conventional approach of employing slow, synchronous inputs from standard frame-based cameras constrains these perception capabilities, particularly during high-speed maneuvers. Recently, event-based sensors have emerged as low latency and low energy alternatives to standard frame-based cameras for capturing high-speed motion, effectively speeding up perception and hence navigation. For coherence, all the perception tasks must be trained on the same input data. However, present-day datasets are curated mainly for a single or a handful of tasks and are limited in the rate of the provided ground truths. To address these limitations, we present Flying Event Dataset fOr Reactive behAviour (FEDORA) - a fully synthetic dataset for perception tasks, with raw data from frame-based cameras, event-based cameras, and Inertial Measurement Units (IMU), along with ground truths for depth, pose, and optical flow at a rate much higher than existing datasets.
Authors: Qingyu Chen, Yan Hu, Xueqing Peng, Qianqian Xie, Qiao Jin, Aidan Gilson, Maxwell B. Singer, Xuguang Ai, Po-Ting Lai, Zhizheng Wang, Vipina Kuttichi Keloth, Kalpana Raja, Jiming Huang, Huan He, Fongci Lin, Jingcheng Du, Rui Zhang, W. Jim Zheng, Ron A. Adelman, Zhiyong Lu, Hua Xu
Abstract: The rapid growth of biomedical literature poses challenges for manual knowledge curation and synthesis. Biomedical Natural Language Processing (BioNLP) automates the process. While Large Language Models (LLMs) have shown promise in general domains, their effectiveness in BioNLP tasks remains unclear due to limited benchmarks and practical guidelines. We perform a systematic evaluation of four LLMs, GPT and LLaMA representatives on 12 BioNLP benchmarks across six applications. We compare their zero-shot, few-shot, and fine-tuning performance with traditional fine-tuning of BERT or BART models. We examine inconsistencies, missing information, hallucinations, and perform cost analysis. Here we show that traditional fine-tuning outperforms zero or few shot LLMs in most tasks. However, closed-source LLMs like GPT-4 excel in reasoning-related tasks such as medical question answering. Open source LLMs still require fine-tuning to close performance gaps. We find issues like missing information and hallucinations in LLM outputs. These results offer practical insights for applying LLMs in BioNLP.
Authors: Hui Lan, Vasilis Syrgkanis
Abstract: Accurate estimation of conditional average treatment effects (CATE) is at the core of personalized decision making. While there is a plethora of models for CATE estimation, model selection is a nontrivial task, due to the fundamental problem of causal inference. Recent empirical work provides evidence in favor of proxy loss metrics with double robust properties and in favor of model ensembling. However, theoretical understanding is lacking. Direct application of prior theoretical work leads to suboptimal oracle model selection rates due to the non-convexity of the model selection problem. We provide regret rates for the major existing CATE ensembling approaches and propose a new CATE model ensembling approach based on Q-aggregation using the doubly robust loss. Our main result shows that causal Q-aggregation achieves statistically optimal oracle model selection regret rates of $\frac{\log(M)}{n}$ (with $M$ models and $n$ samples), with the addition of higher-order estimation error terms related to products of errors in the nuisance functions. Crucially, our regret rate does not require that any of the candidate CATE models be close to the truth. We validate our new method on many semi-synthetic datasets and also provide extensions of our work to CATE model selection with instrumental variables and unobserved confounding.
Authors: Nikolaos Louloudakis, Perry Gibson, Jos\'e Cano, Ajitha Rajan
Abstract: Converting deep learning models between frameworks is a common step to maximize model compatibility across devices and leverage optimization features that may be exclusively provided in one deep learning framework. However, this conversion process may be riddled with bugs, making the converted models either undeployable or problematic, considerably degrading their prediction correctness. In this paper, we propose an automated approach for fault localization and repair, FetaFix, during model conversion between deep learning frameworks. FetaFix is capable of detecting and fixing faults introduced in model input, parameters, hyperparameters, and the model graph during conversion. FetaFix uses a set of fault types (mined from surveying common conversion issues reported in code repositories and forums) to localize potential conversion faults in the converted target model and then repair them appropriately, e.g., replacing the parameters of the target model with those from the source model. This is done iteratively for every image in the dataset, comparing output label differences between the source model and the converted target model until all differences are resolved. We evaluate the effectiveness of FetaFix in fixing model conversion bugs of three widely used image recognition models converted across four different deep learning frameworks. Overall, FetaFix was able to fix $462$ out of $755$ detected conversion faults, either completely repairing or significantly improving the performance of $14$ out of the $15$ erroneous conversion cases.
Authors: Vladyslav Gapyak, Corinna Rentschler, Thomas M\"arz, Andreas Weinmann
Abstract: Objective: Magnetic particle imaging (MPI) is an emerging medical imaging modality which has gained increasing interest in recent years. Among the benefits of MPI are its high temporal resolution, and that the technique does not expose the specimen to any kind of ionizing radiation. It is based on the non-linear response of magnetic nanoparticles to an applied magnetic field. From the electric signal measured in receive coils, the particle concentration has to be reconstructed. Due to the ill-posedness of the reconstruction problem, various regularization methods have been proposed for reconstruction ranging from early stopping methods, via classical Tikhonov regularization and iterative methods to modern machine learning approaches. In this work, we contribute to the latter class: we propose a plug-and-play approach based on a generic zero-shot denoiser with an $\ell^1$-prior. Approach: We validate the reconstruction parameters of the method on a hybrid dataset and compare it with the baseline Tikhonov, DIP and the previous PP-MPI, which is a plug-and-play method with denoiser trained on MPI-friendly data. Main results: We offer a quantitative and qualitative evaluation of the zero-shot plug-and-play approach on the 3D Open MPI dataset. Moreover, we show the quality of the approach with different levels of preprocessing of the data. Significance: The proposed method employs a zero-shot denoiser which has not been trained for the MPI task and therefore saves the cost for training. Moreover, it offers a method that can be potentially applied in future MPI contexts.
Authors: Sadaf Salehkalaibar, Jun Chen, Ashish Khisti, Wei Yu
Abstract: This paper studies the rate-distortion-perception (RDP) tradeoff for a memoryless source model in the asymptotic limit of large block-lengths. The perception measure is based on a divergence between the distributions of the source and reconstruction sequences \emph{conditioned} on the encoder output, first proposed by Mentzer et al. We consider the case when there is no shared randomness between the encoder and the decoder and derive a single-letter characterization of the RDP function for the case of discrete memoryless sources. This is in contrast to the marginal-distribution metric case (introduced by Blau and Michaeli), whose RDP characterization remains open when there is no shared randomness. The achievability scheme is based on lossy source coding with a posterior reference map. For the case of continuous valued sources under the squared error distortion measure and the squared quadratic Wasserstein perception measure, we also derive a single-letter characterization and show that the decoder can be restricted to a noise-adding mechanism. Interestingly, the RDP function characterized for the case of zero perception loss coincides with that of the marginal metric, and further zero perception loss can be achieved with a 3-dB penalty in minimum distortion. Finally we specialize to the case of Gaussian sources, and derive the RDP function for Gaussian vector case and propose a reverse water-filling type solution. We also partially characterize the RDP function for a mixture of Gaussian vector sources.
Authors: Jacob Russin, Ellie Pavlick, Michael J. Frank
Abstract: Human learning embodies a striking duality: sometimes, we appear capable of following logical, compositional rules and benefit from structured curricula (e.g., in formal education), while other times, we rely on an incremental approach or trial-and-error, learning better from curricula that are randomly interleaved. Influential psychological theories explain this seemingly disparate behavioral evidence by positing two qualitatively different learning systems -- one for rapid, rule-based inferences and another for slow, incremental adaptation. It remains unclear how to reconcile such theories with neural networks, which learn via incremental weight updates and are thus a natural model for the latter type of learning, but are not obviously compatible with the former. However, recent evidence suggests that metalearning neural networks and large language models are capable of "in-context learning" (ICL) -- the ability to flexibly grasp the structure of a new task from a few examples. Here, we show that the dynamic interplay between ICL and default in-weight learning (IWL) naturally captures a broad range of learning phenomena observed in humans, reproducing curriculum effects on category-learning and compositional tasks, and recapitulating a tradeoff between flexibility and retention. Our work shows how emergent ICL can equip neural networks with fundamentally different learning properties that can coexist with their native IWL, thus offering a novel perspective on dual-process theories and human cognitive flexibility.
Authors: Yutong He, Alexander Robey, Naoki Murata, Yiding Jiang, Joshua Nathaniel Williams, George J. Pappas, Hamed Hassani, Yuki Mitsufuji, Ruslan Salakhutdinov, J. Zico Kolter
Abstract: Prompt engineering is an effective but labor-intensive way to control text-to-image (T2I) generative models. Its time-intensive nature and complexity have spurred the development of algorithms for automated prompt generation. However, these methods often struggle with transferability across T2I models, require white-box access to the underlying model, or produce non-intuitive prompts. In this work, we introduce PRISM, an algorithm that automatically produces human-interpretable and transferable prompts that can effectively generate desired concepts given only black-box access to T2I models. Inspired by large language model (LLM) jailbreaking, PRISM leverages the in-context learning ability of LLMs to iteratively refine the candidate prompt distribution built upon the reference images. Our experiments demonstrate the versatility and effectiveness of PRISM in generating accurate prompts for objects, styles, and images across multiple T2I models, including Stable Diffusion, DALL-E, and Midjourney.
Authors: Atsuyuki Miyai, Jingkang Yang, Jingyang Zhang, Yifei Ming, Qing Yu, Go Irie, Yixuan Li, Hai Li, Ziwei Liu, Kiyoharu Aizawa
Abstract: This paper introduces a novel task to evaluate the robust understanding capability of Large Multimodal Models (LMMs), termed $\textbf{Unsolvable Problem Detection (UPD)}$. Multiple-choice question answering (MCQA) is widely used to assess the understanding capability of LMMs, but it does not guarantee that LMMs truly comprehend the answer. UPD assesses the LMM's ability to withhold answers when encountering unsolvable problems of MCQA, verifying whether the model truly understands the answer. UPD encompasses three problems: Absent Answer Detection (AAD), Incompatible Answer Set Detection (IASD), and Incompatible Visual Question Detection (IVQD), covering unsolvable cases like answer-lacking or incompatible choices and image-question mismatches. For the evaluation, we introduce the MM-UPD Bench, a benchmark for assessing performance across various ability dimensions. Our experiments reveal that even most LMMs, which demonstrate adequate performance on existing benchmarks, struggle significantly with MM-UPD, underscoring a novel aspect of trustworthiness that current benchmarks have overlooked. A detailed analysis shows that LMMs have different bottlenecks and chain-of-thought and self-reflection improved performance for LMMs with the bottleneck in their LLM capability. We hope our insights will enhance the broader understanding and development of more reliable LMMs.
Authors: Tianyu Zhu, Myong Chol Jung, Jesse Clark
Abstract: Contrastive learning has gained widespread adoption for retrieval tasks due to its minimal requirement for manual annotations. However, popular training frameworks typically learn from binary (positive/negative) relevance, making them ineffective at incorporating desired rankings. As a result, the poor ranking performance of these models forces systems to employ a re-ranker, which increases complexity, maintenance effort and inference time. To address this, we introduce Generalized Contrastive Learning (GCL), a training framework designed to learn from continuous ranking scores beyond binary relevance. GCL encodes both relevance and ranking information into a unified embedding space by applying ranking scores to the loss function. This enables a single-stage retrieval system. In addition, during our research, we identified a lack of public multi-modal datasets that benchmark both retrieval and ranking capabilities. To facilitate this and future research for ranked retrieval, we curated a large-scale MarqoGS-10M dataset using GPT-4 and Google Shopping, providing ranking scores for each of the 10 million query-document pairs. Our results show that GCL achieves a 29.3% increase in NDCG@10 for in-domain evaluations and 6.0% to 10.0% increases for cold-start evaluations compared to the finetuned CLIP baseline with MarqoGS-10M. Additionally, we evaluated GCL offline on a proprietary user interaction data. GCL shows an 11.2% gain for in-domain evaluations. The dataset and the method are available at: https://github.com/marqo-ai/GCL.
Authors: Xiao-Yin Liu, Guotao Li, Xiao-Hu Zhou, Xu Liang, Zeng-Guang Hou
Abstract: Accurate recognition of human motion intention (HMI) is beneficial for exoskeleton robots to improve the wearing comfort level and achieve natural human-robot interaction. A classifier trained on labeled source subjects (domains) performs poorly on unlabeled target subject since the difference in individual motor characteristics. The unsupervised domain adaptation (UDA) method has become an effective way to this problem. However, the labeled data are collected from multiple source subjects that might be different not only from the target subject but also from each other. The current UDA methods for HMI recognition ignore the difference between each source subject, which reduces the classification accuracy. Therefore, this paper considers the differences between source subjects and develops a novel theory and algorithm for UDA to recognize HMI, where the margin disparity discrepancy (MDD) is extended to multi-source UDA theory and a novel weight-aware-based multi-source UDA algorithm (WMDD) is proposed. The source domain weight, which can be adjusted adaptively by the MDD between each source subject and target subject, is incorporated into UDA to measure the differences between source subjects. The developed multi-source UDA theory is theoretical and the generalization error on target subject is guaranteed. The theory can be transformed into an optimization problem for UDA, successfully bridging the gap between theory and algorithm. Moreover, a lightweight network is employed to guarantee the real-time of classification and the adversarial learning between feature generator and ensemble classifiers is utilized to further improve the generalization ability. The extensive experiments verify theoretical analysis and show that WMDD outperforms previous UDA methods on HMI recognition tasks.
Authors: Christoforos Vasilatos, Dunia J. Mahboobeh, Hithem Lamri, Manaar Alam, Michail Maniatakos
Abstract: Industrial Control Systems (ICS) are extensively used in critical infrastructures ensuring efficient, reliable, and continuous operations. However, their increasing connectivity and addition of advanced features make them vulnerable to cyber threats, potentially leading to severe disruptions in essential services. In this context, honeypots play a vital role by acting as decoy targets within ICS networks, or on the Internet, helping to detect, log, analyze, and develop mitigations for ICS-specific cyber threats. Deploying ICS honeypots, however, is challenging due to the necessity of accurately replicating industrial protocols and device characteristics, a crucial requirement for effectively mimicking the unique operational behavior of different industrial systems. Moreover, this challenge is compounded by the significant manual effort required in also mimicking the control logic the PLC would execute, in order to capture attacker traffic aiming to disrupt critical infrastructure operations. In this paper, we propose LLMPot, a novel approach for designing honeypots in ICS networks harnessing the potency of Large Language Models (LLMs). LLMPot aims to automate and optimize the creation of realistic honeypots with vendor-agnostic configurations, and for any control logic, aiming to eliminate the manual effort and specialized knowledge traditionally required in this domain. We conducted extensive experiments focusing on a wide array of parameters, demonstrating that our LLM-based approach can effectively create honeypot devices implementing different industrial protocols and diverse control logic.
Authors: Javier Lopez-Piqueres, Jing Chen
Abstract: In this study, we introduce a novel family of tensor networks, termed constrained matrix product states (MPS), designed to incorporate exactly arbitrary discrete linear constraints, including inequalities, into sparse block structures. These tensor networks are particularly tailored for modeling distributions with support strictly over the feasible space, offering benefits such as reducing the search space in optimization problems, alleviating overfitting, improving training efficiency, and decreasing model size. Central to our approach is the concept of a quantum region, an extension of quantum numbers traditionally used in U(1) symmetric tensor networks, adapted to capture any linear constraint, including the unconstrained scenario. We further develop a novel canonical form for these new MPS, which allow for the merging and factorization of tensor blocks according to quantum region fusion rules and permit optimal truncation schemes. Utilizing this canonical form, we apply an unsupervised training strategy to optimize arbitrary objective functions subject to discrete linear constraints. Our method's efficacy is demonstrated by solving the quadratic knapsack problem, achieving superior performance compared to a leading nonlinear integer programming solver. Additionally, we analyze the complexity and scalability of our approach, demonstrating its potential in addressing complex constrained combinatorial optimization problems.
Authors: Amit Kadan, Kevin Ryczko, Erika Lloyd, Adrian Roitberg, Takeshi Yamazaki
Abstract: Generative AI has the potential to revolutionize drug discovery. Yet, despite recent advances in deep learning, existing models cannot generate molecules that satisfy all desired physicochemical properties. Herein, we describe IDOLpro, a generative chemistry AI combining diffusion with multi-objective optimization for structure-based drug design. Differentiable scoring functions guide the latent variables of the diffusion model to explore uncharted chemical space and generate novel ligands in silico, optimizing a plurality of target physicochemical properties. We demonstrate our platform's effectiveness by generating ligands with optimized binding affinity and synthetic accessibility on two benchmark sets. IDOLpro produces ligands with binding affinities over 10%-20% better than the next best state-of-the-art method on each test set, producing more drug-like molecules with generally better synthetic accessibility scores than other methods. We do a head-to-head comparison of IDOLpro against a classic virtual screen of a large database of drug-like molecules. We show that IDOLpro can generate molecules for a range of important disease-related targets with better binding affinity and synthetic accessibility than any molecule found in the virtual screen while being over 100x faster and less expensive to run. On a test set of experimental complexes, IDOLpro is the first to produce molecules with better binding affinities than experimentally observed ligands. IDOLpro can accommodate other scoring functions (e.g. ADME-Tox) to accelerate hit-finding, hit-to-lead, and lead optimization for drug discovery.
Authors: Chirag Wadhwa, Mina Doosti
Abstract: In this work, we study the learnability of quantum circuits in the near term. We demonstrate the natural robustness of quantum statistical queries for learning quantum processes, motivating their use as a theoretical tool for near-term learning problems. We adapt a learning algorithm for constant-depth quantum circuits to the quantum statistical query setting, and show that such circuits can be learned in our setting with only a linear overhead in the query complexity. We prove average-case quantum statistical query lower bounds for learning, within diamond distance, random quantum circuits with depth at least logarithmic and at most linear in the system size. Finally, we prove that pseudorandom unitaries (PRUs) cannot be constructed using circuits of constant depth by constructing an efficient distinguisher using existing learning algorithms. To show the correctness of our distinguisher, we prove a new variation of the quantum no free lunch theorem.
Authors: Jake A. Soloff, Rina Foygel Barber, Rebecca Willett
Abstract: We propose a new framework for algorithmic stability in the context of multiclass classification. In practice, classification algorithms often operate by first assigning a continuous score (for instance, an estimated probability) to each possible label, then taking the maximizer -- i.e., selecting the class that has the highest score. A drawback of this type of approach is that it is inherently unstable, meaning that it is very sensitive to slight perturbations of the training data, since taking the maximizer is discontinuous. Motivated by this challenge, we propose a pipeline for constructing stable classifiers from data, using bagging (i.e., resampling and averaging) to produce stable continuous scores, and then using a stable relaxation of argmax, which we call the "inflated argmax," to convert these scores to a set of candidate labels. The resulting stability guarantee places no distributional assumptions on the data, does not depend on the number of classes or dimensionality of the covariates, and holds for any base classifier. Using a common benchmark data set, we demonstrate that the inflated argmax provides necessary protection against unstable classifiers, without loss of accuracy.
Authors: Ilia Azizi, Marc-Olivier Boldi, Val\'erie Chavez-Demoulin
Abstract: This work introduces the Supervised Expectation-Maximization Framework (SEMF), a versatile and model-agnostic approach for generating prediction intervals with any ML model. SEMF extends the Expectation-Maximization algorithm, traditionally used in unsupervised learning, to a supervised context, leveraging latent variable modeling for uncertainty estimation. Through extensive empirical evaluation of diverse simulated distributions and 11 real-world tabular datasets, SEMF consistently produces narrower prediction intervals while maintaining the desired coverage probability, outperforming traditional quantile regression methods. Furthermore, without using the quantile (pinball) loss, SEMF allows point predictors, including gradient-boosted trees and neural networks, to be calibrated with conformal quantile regression. The results indicate that SEMF enhances uncertainty quantification under diverse data distributions and is particularly effective for models that otherwise struggle with inherent uncertainty representation.
Authors: Fang Chen, Gourav Datta, Mujahid Al Rafi, Hyeran Jeon, Meng Tang
Abstract: The expansion of neural network sizes and the enhanced resolution of modern image sensors result in heightened memory and power demands to process modern computer vision models. In order to deploy these models in extremely resource-constrained edge devices, it is crucial to reduce their peak memory, which is the maximum memory consumed during the execution of a model. A naive approach to reducing peak memory is aggressive down-sampling of feature maps via pooling with large stride, which often results in unacceptable degradation in network performance. To mitigate this problem, we propose residual encoded distillation (ReDistill) for peak memory reduction in a teacher-student framework, in which a student network with less memory is derived from the teacher network using aggressive pooling. We apply our distillation method to multiple problems in computer vision, including image classification and diffusion-based image generation. For image classification, our method yields 4x-5x theoretical peak memory reduction with less degradation in accuracy for most CNN-based architectures. For diffusion-based image generation, our proposed distillation method yields a denoising network with 4x lower theoretical peak memory while maintaining decent diversity and fidelity for image generation. Experiments demonstrate our method's superior performance compared to other feature-based and response-based distillation methods when applied to the same student network. The code is available at https://github.com/mengtang-lab/ReDistill.
Authors: Md Saiful Islam, Tariq Adnan, Jan Freyberg, Sangwu Lee, Abdelrahman Abdelkader, Meghan Pawlik, Cathe Schwartz, Karen Jaffe, Ruth B. Schneider, E Ray Dorsey, Ehsan Hoque
Abstract: Limited accessibility to neurological care leads to underdiagnosed Parkinson's Disease (PD), preventing early intervention. Existing AI-based PD detection methods primarily focus on unimodal analysis of motor or speech tasks, overlooking the multifaceted nature of the disease. To address this, we introduce a large-scale, multi-task video dataset consisting of 1102 sessions (each containing videos of finger tapping, facial expression, and speech tasks captured via webcam) from 845 participants (272 with PD). We propose a novel Uncertainty-calibrated Fusion Network (UFNet) that leverages this multimodal data to enhance diagnostic accuracy. UFNet employs independent task-specific networks, trained with Monte Carlo Dropout for uncertainty quantification, followed by self-attended fusion of features, with attention weights dynamically adjusted based on task-specific uncertainties. To ensure patient-centered evaluation, the participants were randomly split into three sets: 60% for training, 20% for model selection, and 20% for final performance evaluation. UFNet significantly outperformed single-task models in terms of accuracy, area under the ROC curve (AUROC), and sensitivity while maintaining non-inferior specificity. Withholding uncertain predictions further boosted the performance, achieving 88.0+-0.3%$ accuracy, 93.0+-0.2% AUROC, 79.3+-0.9% sensitivity, and 92.6+-0.3% specificity, at the expense of not being able to predict for 2.3+-0.3% data (+- denotes 95% confidence interval). Further analysis suggests that the trained model does not exhibit any detectable bias across sex and ethnic subgroups and is most effective for individuals aged between 50 and 80. Requiring only a webcam and microphone, our approach facilitates accessible home-based PD screening, especially in regions with limited healthcare resources.
Authors: Ilya Timofeyev, Alexey Schwarzmann, Dmitri Kuzmin
Abstract: We propose a combination of machine learning and flux limiting for property-preserving subgrid scale modeling in the context of flux-limited finite volume methods for the one-dimensional shallow-water equations. The numerical fluxes of a conservative target scheme are fitted to the coarse-mesh averages of a monotone fine-grid discretization using a neural network to parametrize the subgrid scale components. To ensure positivity preservation and the validity of local maximum principles, we use a flux limiter that constrains the intermediate states of an equivalent fluctuation form to stay in a convex admissible set. The results of our numerical studies confirm that the proposed combination of machine learning with monolithic convex limiting produces meaningful closures even in scenarios for which the network was not trained.
Authors: Jinming Nian, Zhiyuan Peng, Qifan Wang, Yi Fang
Abstract: In knowledge-intensive tasks such as open-domain question answering (OpenQA), large language models (LLMs) often struggle to generate factual answers, relying solely on their internal (parametric) knowledge. To address this limitation, Retrieval-Augmented Generation (RAG) systems enhance LLMs by retrieving relevant information from external sources, thereby positioning the retriever as a pivotal component. Although dense retrieval demonstrates state-of-the-art performance, its training poses challenges due to the scarcity of ground-truth evidence, largely attributed to the high costs of human annotation. In this paper, we propose W-RAG, a method that draws weak training signals from the downstream task (such as OpenQA) of an LLM, and fine-tunes the retriever to prioritize passages that most benefit the task. Specifically, we rerank the top-$k$ passages retrieved via BM25 by assessing the probability that the LLM will generate the correct answer for a question given each passage. The highest-ranking passages are then used as positive fine-tuning examples for dense retrieval. We conduct comprehensive experiments across four publicly available OpenQA datasets to demonstrate that our approach enhances both retrieval and OpenQA performance compared to baseline models, achieving results comparable to models fine-tuned with human-labeled data.
Authors: Xin Liu, Weijia Zhang, Min-Ling Zhang
Abstract: Although attention-based multi-instance learning (MIL) algorithms have achieved impressive performance on slide-level whole slide image (WSI) classification tasks, they are prone to mistakenly focusing on irrelevant patterns such as staining conditions and tissue morphology, leading to incorrect patch-level predictions and unreliable interpretability. In this paper, we analyze why attention-based methods tend to rely on spurious correlations in their predictions. Furthermore, we revisit max-pooling-based approaches and examine the reasons behind the underperformance of existing methods. We argue that well-trained max-pooling-based MIL models can make predictions based on causal factors and avoid relying on spurious correlations. Building on these insights, we propose a simple yet effective max-pooling-based MIL method (FocusMIL) that outperforms existing mainstream attention-based methods on two datasets. In this position paper, we advocate renewed attention to max-pooling-based methods to achieve more robust and interpretable predictions.
Authors: Steve Hanneke, Samory Kpotufe
Abstract: Transfer Learning aims to optimally aggregate samples from a target distribution, with related samples from a so-called source distribution to improve target risk. Multiple procedures have been proposed over the last two decades to address this problem, each driven by one of a multitude of possible divergence measures between source and target distributions. A first question asked in this work is whether there exist unified algorithmic approaches that automatically adapt to many of these divergence measures simultaneously. We show that this is indeed the case for a large family of divergences proposed across classification and regression tasks, as they all happen to upper-bound the same measure of continuity between source and target risks, which we refer to as a weak modulus of transfer. This more unified view allows us, first, to identify algorithmic approaches that are simultaneously adaptive to these various divergence measures via a reduction to particular confidence sets. Second, it allows for a more refined understanding of the statistical limits of transfer under such divergences, and in particular, reveals regimes with faster rates than might be expected under coarser lenses. We then turn to situations that are not well captured by the weak modulus and corresponding divergences: these are situations where the aggregate of source and target data can improve target performance significantly beyond what's possible with either source or target data alone. We show that common such situations -- as may arise, e.g., under certain causal models with spurious correlations -- are well described by a so-called strong modulus of transfer which supersedes the weak modulus. We finally show that the strong modulus also admits adaptive procedures, which achieve near optimal rates in terms of the unknown strong modulus, and therefore apply in more general settings.
Authors: Guanglei Zhou, Bhargav Korrapati, Gaurav Rajavendra Reddy, Chen-Chia Chang, Jingyu Pan, Jiang Hu, Yiran Chen, Dipto G. Thakurta
Abstract: Generating diverse VLSI layout patterns is essential for various downstream tasks in design for manufacturing, as design rules continually evolve during the development of new technology nodes. However, existing training-based methods for layout pattern generation rely on large datasets. In practical scenarios, especially when developing a new technology node, obtaining such extensive layout data is challenging. Consequently, training models with large datasets becomes impractical, limiting the scalability and adaptability of prior approaches. To this end, we propose PatternPaint, a diffusion-based framework capable of generating legal patterns with limited design-rule-compliant training samples. PatternPaint simplifies complex layout pattern generation into a series of inpainting processes with a template-based denoising scheme. Furthermore, we perform few-shot finetuning on a pretrained image foundation model with only 20 design-rule-compliant samples. Experimental results show that using a sub-3nm technology node (Intel 18A), our model is the only one that can generate legal patterns in complex 2D metal interconnect design rule settings among all previous works and achieves a high diversity score. Additionally, our few-shot finetuning can boost the legality rate with 1.87X improvement compared to the original pretrained model. As a result, we demonstrate a production-ready approach for layout pattern generation in developing new technology nodes.
Authors: Ashirbad Mishra, Soumik Dey, Marshall Wu, Jinyu Zhao, He Yu, Kaichen Ni, Binbin Li, Kamesh Madduri
Abstract: Online sellers and advertisers are recommended keyphrases for their listed products, which they bid on to enhance their sales. One popular paradigm that generates such recommendations is Extreme Multi-Label Classification (XMC), which involves tagging/mapping keyphrases to items. We outline the limitations of using traditional item-query based tagging or mapping techniques for keyphrase recommendations on E-Commerce platforms. We introduce GraphEx, an innovative graph-based approach that recommends keyphrases to sellers using extraction of token permutations from item titles. Additionally, we demonstrate that relying on traditional metrics such as precision/recall can be misleading in practical applications, thereby necessitating a combination of metrics to evaluate performance in real-world scenarios. These metrics are designed to assess the relevance of keyphrases to items and the potential for buyer outreach. GraphEx outperforms production models at eBay, achieving the objectives mentioned above. It supports near real-time inferencing in resource-constrained production environments and scales effectively for billions of items.
Authors: Jan Schnabel, Marco Roth
Abstract: Since the entry of kernel theory in the field of quantum machine learning, quantum kernel methods (QKMs) have gained increasing attention with regard to both probing promising applications and delivering intriguing research insights. Benchmarking these methods is crucial to gain robust insights and to understand their practical utility. In this work, we present a comprehensive large-scale study examining QKMs based on fidelity quantum kernels (FQKs) and projected quantum kernels (PQKs) across a manifold of design choices. Our investigation encompasses both classification and regression tasks for five dataset families and 64 datasets, systematically comparing the use of FQKs and PQKs quantum support vector machines and kernel ridge regression. This resulted in over 20,000 models that were trained and optimized using a state-of-the-art hyperparameter search to ensure robust and comprehensive insights. We delve into the importance of hyperparameters on model performance scores and support our findings through rigorous correlation analyses. Additionally, we provide an in-depth analysis addressing the design freedom of PQKs and explore the underlying principles responsible for learning. Our goal is not to identify the best-performing model for a specific task but to uncover the mechanisms that lead to effective QKMs and reveal universal patterns.
Authors: Yetao Wu, Yihong Wang, Teng Chen, Ningyuan Xi, Qingqing Gu, Hongyang Lei, Luo Ji
Abstract: Hallucination is a major challenge for large language models (LLMs), preventing their further application in some fields. The skeptical thinking of humankind could be useful for LLMs to self-cognition, self-reflection and alleviate their hallucinations. Inspired by this consideration, we propose a novel approach called LaMsS, which combines the semantic understanding capability of LLMs with self-skepticism. By introducing a series of skepticism tokens and augmenting them into the vocabulary, we conduct both pertaining and finetuning, which allow the LLM to decode each normal token followed by a skeptical token, representing different skepticism levels. By calculating the response skepticism given a query, one can define a new self-aware LLM which is only willing to answer with relative lower skepticism level than the threshold. By examining the accuracy, AUC and AP of willingly answering questions, we demonstrate that LaMsS achieves better performance than baselines on both multi-choice questions and open-domain question-answering benchmarks, and can generalize to multi-task and out-of-domain settings. Our study sheds some lights on the self-skepticism modeling on further artificial intelligence. Project code and model checkpoints can be found in https://anonymous.4open.science/r/SM-1E76.
Authors: Justin Wang, Haimin Hu, Duy Phuong Nguyen, Jaime Fern\'andez Fisac
Abstract: While robust optimal control theory provides a rigorous framework to compute robot control policies that are provably safe, it struggles to scale to high-dimensional problems, leading to increased use of deep learning for tractable synthesis of robot safety. Unfortunately, existing neural safety synthesis methods often lack convergence guarantees and solution interpretability. In this paper, we present Minimax Actors Guided by Implicit Critic Stackelberg (MAGICS), a novel adversarial reinforcement learning (RL) algorithm that guarantees local convergence to a minimax equilibrium solution. We then build on this approach to provide local convergence guarantees for a general deep RL-based robot safety synthesis algorithm. Through both simulation studies on OpenAI Gym environments and hardware experiments with a 36-dimensional quadruped robot, we show that MAGICS can yield robust control policies outperforming the state-of-the-art neural safety synthesis methods.
Authors: Fernando M. Quintana, Maryada, Pedro L. Galindo, Elisa Donati, Giacomo Indiveri, Fernando Perez-Pe\~na
Abstract: Developing dedicated mixed-signal neuromorphic computing systems optimized for real-time sensory-processing in extreme edge-computing applications requires time-consuming design, fabrication, and deployment of full-custom neuromorphic processors. To ensure that initial prototyping efforts, exploring the properties of different network architectures and parameter settings, lead to realistic results, it is important to use simulation frameworks that match as best as possible the properties of the final hardware. This is particularly challenging for neuromorphic hardware platforms made using mixed-signal analog/digital circuits, due to the variability and noise sensitivity of their components. In this paper, we address this challenge by developing a software spiking neural network simulator explicitly designed to account for the properties of mixed-signal neuromorphic circuits, including device mismatch variability. The simulator, called ARCANA (A Realistic Simulation Framework for Analog/Digital Neuromorphic Architectures), is designed to reproduce the dynamics of mixed-signal synapse and neuron electronic circuits with autogradient differentiation for parameter optimization and GPU acceleration. We demonstrate the effectiveness of this approach by matching software simulation results with measurements made from an existing neuromorphic processor. We show how the results obtained provide a reliable estimate of the behavior of the spiking neural network trained in software, once deployed in hardware. This framework enables the development and innovation of new learning rules and processing architectures in neuromorphic embedded systems.
Authors: Jiashuo Li, Shaokun Wang, Bo Qian, Yuhang He, Xing Wei, Qiang Wang, Yihong Gong
Abstract: Non-exemplar class Incremental Learning (NECIL) enables models to continuously acquire new classes without retraining from scratch and storing old task exemplars, addressing privacy and storage issues. However, the absence of data from earlier tasks exacerbates the challenge of catastrophic forgetting in NECIL. In this paper, we propose a novel framework called Dynamic Integration of task-specific Adapters (DIA), which comprises two key components: Task-Specific Adapter Integration (TSAI) and Patch-Level Model Alignment. TSAI boosts compositionality through a patch-level adapter integration strategy, which provides a more flexible compositional solution while maintaining low computation costs. Patch-Level Model Alignment maintains feature consistency and accurate decision boundaries via two specialized mechanisms: Patch-Level Distillation Loss (PDL) and Patch-Level Feature Reconstruction method (PFR). Specifically, the PDL preserves feature-level consistency between successive models by implementing a distillation loss based on the contributions of patch tokens to new class learning. The PFR facilitates accurate classifier alignment by reconstructing old class features from previous tasks that adapt to new task knowledge. Extensive experiments validate the effectiveness of our DIA, revealing significant improvements on benchmark datasets in the NECIL setting, maintaining an optimal balance between computational complexity and accuracy.
Authors: Yunkee Chae, Woosung Choi, Yuhta Takida, Junghyun Koo, Yukara Ikemiya, Zhi Zhong, Kin Wai Cheuk, Marco A. Mart\'inez-Ram\'irez, Kyogu Lee, Wei-Hsiang Liao, Yuki Mitsufuji
Abstract: Recent state-of-the-art neural audio compression models have progressively adopted residual vector quantization (RVQ). Despite this success, these models employ a fixed number of codebooks per frame, which can be suboptimal in terms of rate-distortion tradeoff, particularly in scenarios with simple input audio, such as silence. To address this limitation, we propose variable bitrate RVQ (VRVQ) for audio codecs, which allows for more efficient coding by adapting the number of codebooks used per frame. Furthermore, we propose a gradient estimation method for the non-differentiable masking operation that transforms from the importance map to the binary importance mask, improving model training via a straight-through estimator. We demonstrate that the proposed training framework achieves superior results compared to the baseline method and shows further improvement when applied to the current state-of-the-art codec.
Authors: Clement Neo, Luke Ong, Philip Torr, Mor Geva, David Krueger, Fazl Barez
Abstract: Vision-Language Models (VLMs) are powerful tools for processing and understanding text and images. We study the processing of visual tokens in the language model component of LLaVA, a prominent VLM. Our approach focuses on analyzing the localization of object information, the evolution of visual token representations across layers, and the mechanism of integrating visual information for predictions. Through ablation studies, we demonstrated that object identification accuracy drops by over 70\% when object-specific tokens are removed. We observed that visual token representations become increasingly interpretable in the vocabulary space across layers, suggesting an alignment with textual tokens corresponding to image content. Finally, we found that the model extracts object information from these refined representations at the last token position for prediction, mirroring the process in text-only language models for factual association tasks. These findings provide crucial insights into how VLMs process and integrate visual information, bridging the gap between our understanding of language and vision models, and paving the way for more interpretable and controllable multimodal systems.
Authors: Evans Xu Han, Linghao Jin, Xiaofeng Liu, Paul Pu Liang
Abstract: Despite the impressive text-to-image (T2I) synthesis capabilities of diffusion models, they often struggle to understand compositional relationships between objects and attributes, especially in complex settings. Existing solutions have tackled these challenges by optimizing the cross-attention mechanism or learning from the caption pairs with minimal semantic changes. However, can we generate high-quality complex contrastive images that diffusion models can directly discriminate based on visual representations? In this work, we leverage large-language models (LLMs) to compose realistic, complex scenarios and harness Visual-Question Answering (VQA) systems alongside diffusion models to automatically curate a contrastive dataset, ConPair, consisting of 15k pairs of high-quality contrastive images. These pairs feature minimal visual discrepancies and cover a wide range of attribute categories, especially complex and natural scenarios. To learn effectively from these error cases, i.e., hard negative images, we propose EvoGen, a new multi-stage curriculum for contrastive learning of diffusion models. Through extensive experiments across a wide range of compositional scenarios, we showcase the effectiveness of our proposed framework on compositional T2I benchmarks.
Authors: Jun Bai, Yiliao Song, Di Wu, Atul Sajjanhar, Yong Xiang, Wei Zhou, Xiaohui Tao, Yan Li, Yue Li
Abstract: One-Shot Federated Learning (OSFL) restricts communication between the server and clients to a single round, significantly reducing communication costs and minimizing privacy leakage risks compared to traditional Federated Learning (FL), which requires multiple rounds of communication. However, existing OSFL frameworks remain vulnerable to distributional heterogeneity, as they primarily focus on model heterogeneity while neglecting data heterogeneity. To bridge this gap, we propose FedHydra, a unified, data-free, OSFL framework designed to effectively address both model and data heterogeneity. Unlike existing OSFL approaches, FedHydra introduces a novel two-stage learning mechanism. Specifically, it incorporates model stratification and heterogeneity-aware stratified aggregation to mitigate the challenges posed by both model and data heterogeneity. By this design, the data and model heterogeneity issues are simultaneously monitored from different aspects during learning. Consequently, FedHydra can effectively mitigate both issues by minimizing their inherent conflicts. We compared FedHydra with five SOTA baselines on four benchmark datasets. Experimental results show that our method outperforms the previous OSFL methods in both homogeneous and heterogeneous settings. Our code is available at https://anonymous.4open.science/r/Fed-SA-A4D7.
Authors: Massimo Bilancia, Samuele Magro
Abstract: This paper presents a variant of the Multinomial mixture model tailored to the unsupervised classification of short text data. While the Multinomial probability vector is traditionally assigned a Dirichlet prior distribution, this work explores an alternative formulation based on the Beta-Liouville distribution, which offers a more flexible correlation structure than the Dirichlet. We examine the theoretical properties of the Beta-Liouville distribution, with particular focus on its conjugacy with the Multinomial likelihood. This property enables the derivation of update equations for a CAVI (Coordinate Ascent Variational Inference) algorithm, facilitating approximate posterior inference of the model parameters. In addition, we introduce a stochastic variant of the CAVI algorithm to enhance scalability. The paper concludes with empirical examples demonstrating effective strategies for selecting the Beta-Liouville hyperparameters.
Authors: Yangtao Deng, Xiang Shi, Zhuo Jiang, Xingjian Zhang, Lei Zhang, Zhang Zhang, Bo Li, Zuquan Song, Hang Zhu, Gaohong Liu, Fuliang Li, Shuguang Wang, Haibin Lin, Jianxi Ye, Minlan Yu
Abstract: Large-scale distributed model training requires simultaneous training on up to thousands of machines. Faulty machine detection is critical when an unexpected fault occurs in a machine. From our experience, a training task can encounter two faults per day on average, possibly leading to a halt for hours. To address the drawbacks of the time-consuming and labor-intensive manual scrutiny, we propose Minder, an automatic faulty machine detector for distributed training tasks. The key idea of Minder is to automatically and efficiently detect faulty distinctive monitoring metric patterns, which could last for a period before the entire training task comes to a halt. Minder has been deployed in our production environment for over one year, monitoring daily distributed training tasks where each involves up to thousands of machines. In our real-world fault detection scenarios, Minder can accurately and efficiently react to faults within 3.6 seconds on average, with a precision of 0.904 and F1-score of 0.893.
Authors: Hiroshi Yokoyama, Ryusei Shingaki, Kaneharu Nishino, Shohei Shimizu, Thong Pham
Abstract: Recent rapid advancements of machine learning have greatly enhanced the accuracy of prediction models, but most models remain "black boxes", making prediction error diagnosis challenging, especially with outliers. This lack of transparency hinders trust and reliability in industrial applications. Heuristic attribution methods, while helpful, often fail to capture true causal relationships, leading to inaccurate error attributions. Various root-cause analysis methods have been developed using Shapley values, yet they typically require predefined causal graphs, limiting their applicability for prediction errors in machine learning models. To address these limitations, we introduce the Causal-Discovery-based Root-Cause Analysis (CD-RCA) method that estimates causal relationships between the prediction error and the explanatory variables, without needing a pre-defined causal graph. By simulating synthetic error data, CD-RCA can identify variable contributions to outliers in prediction errors by Shapley values. Extensive experiments show CD-RCA outperforms current heuristic attribution methods.
Authors: Nardiena A. Pratama, Shaoyang Fan, Gianluca Demartini
Abstract: Human-annotated content is often used to train machine learning (ML) models. However, recently, language and multi-modal foundational models have been used to replace and scale-up human annotator's efforts. This study explores the similarity between human-generated and ML-generated annotations of images across diverse socio-economic contexts (RQ1) and their impact on ML model performance and bias (RQ2). We aim to understand differences in perception and identify potential biases in content interpretation. Our dataset comprises images of people from various geographical regions and income levels, covering various daily activities and home environments. ML captions and human labels show highest similarity at a low-level, i.e., types of words that appear and sentence structures, but all annotations are consistent in how they perceive images across regions. ML Captions resulted in best overall region classification performance, while ML Objects and ML Captions performed best overall for income regression. ML annotations worked best for action categories, while human input was more effective for non-action categories. These findings highlight the notion that both human and machine annotations are important, and that human-generated annotations are yet to be replaceable.
Authors: Ivan \v{C}ili\'c, Anna Lackinger, Pantelis Frangoudis, Ivana Podnar \v{Z}arko, Alireza Furutanpey, Ilir Murturi, Schahram Dustdar
Abstract: Deploying a Hierarchical Federated Learning (HFL) pipeline across the computing continuum (CC) requires careful organization of participants into a hierarchical structure with intermediate aggregation nodes between FL clients and the global FL server. This is challenging to achieve due to (i) cost constraints, (ii) varying data distributions, and (iii) the volatile operating environment of the CC. In response to these challenges, we present a framework for the adaptive orchestration of HFL pipelines, designed to be reactive to client churn and infrastructure-level events, while balancing communication cost and ML model accuracy. Our mechanisms identify and react to events that cause HFL reconfiguration actions at runtime, building on multi-level monitoring information (model accuracy, resource availability, resource cost). Moreover, our framework introduces a generic methodology for estimating reconfiguration costs to continuously re-evaluate the quality of adaptation actions, while being extensible to optimize for various HFL performance criteria. By extending the Kubernetes ecosystem, our framework demonstrates the ability to react promptly and effectively to changes in the operating environment, making the best of the available communication cost budget and effectively balancing costs and ML performance at runtime.
Authors: Kasra Arabi, Benjamin Feuer, R. Teal Witter, Chinmay Hegde, Niv Cohen
Abstract: As the quality of image generators continues to improve, deepfakes become a topic of considerable societal debate. Image watermarking allows responsible model owners to detect and label their AI-generated content, which can mitigate the harm. Yet, current state-of-the-art methods in image watermarking remain vulnerable to forgery and removal attacks. This vulnerability occurs in part because watermarks distort the distribution of generated images, unintentionally revealing information about the watermarking techniques. In this work, we first demonstrate a distortion-free watermarking method for images, based on a diffusion model's initial noise. However, detecting the watermark requires comparing the initial noise reconstructed for an image to all previously used initial noises. To mitigate these issues, we propose a two-stage watermarking framework for efficient detection. During generation, we augment the initial noise with generated Fourier patterns to embed information about the group of initial noises we used. For detection, we (i) retrieve the relevant group of noises, and (ii) search within the given group for an initial noise that might match our image. This watermarking approach achieves state-of-the-art robustness to forgery and removal against a large battery of attacks.
Authors: Xiayin Lou, Peng Luo, Liqiu Meng
Abstract: Spatial prediction is a fundamental task in geography. In recent years, with advances in geospatial artificial intelligence (GeoAI), numerous models have been developed to improve the accuracy of geographic variable predictions. Beyond achieving higher accuracy, it is equally important to obtain predictions with uncertainty measures to enhance model credibility and support responsible spatial prediction. Although geostatistic methods like Kriging offer some level of uncertainty assessment, such as Kriging variance, these measurements are not always accurate and lack general applicability to other spatial models. To address this issue, we propose a model-agnostic uncertainty assessment method called GeoConformal Prediction, which incorporates geographical weighting into conformal prediction. We applied it to two classic spatial prediction cases, spatial regression and spatial interpolation, to evaluate its reliability. First, in the spatial regression case, we used XGBoost to predict housing prices, followed by GeoConformal to calculate uncertainty. Our results show that GeoConformal achieved a coverage rate of 93.67%, while Bootstrap methods only reached a maximum coverage of 81.00% after 2000 runs. Next, we applied GeoConformal to spatial interpolation models. We found that the uncertainty obtained from GeoConformal aligned closely with the variance in Kriging. Finally, using GeoConformal, we analyzed the sources of uncertainty in spatial prediction. We found that explicitly including local features in AI models can significantly reduce prediction uncertainty, especially in areas with strong local dependence. Our findings suggest that GeoConformal holds potential not only for geographic knowledge discovery but also for guiding the design of future GeoAI models, paving the way for more reliable and interpretable spatial prediction frameworks.
Authors: ChonLam Lao, Minlan Yu, Aditya Akella, Jiamin Cao, Yu Guan, Pengcheng Zhang, Zhilong Zheng, Yichi Xu, Ennan Zhai, Dennis Cai, Jiaqi Gao
Abstract: Large-scale ML training jobs are frequently interrupted by hardware and software anomalies, failures, and management events. Existing solutions like checkpointing or runtime reconfiguration suffer from long downtimes, degraded performance, or undesired changes to training strategies. We present TrainMover, a resilient runtime that leverages standby machines to handle interruptions with minimal downtime and zero memory overhead. To achieve these goals, TrainMover introduces two key techniques: two-phase, delta-based communication group setups and communication-free sandboxed shadow iterations. Our evaluation shows that TrainMover consistently achieves second-level downtime across all evaluated models during migration, maintaining 99\% training efficiency during periodic 10-minute rebalancing. We also demonstrate the effectiveness of TrainMover in handling various interruptions.
Authors: Benjamin Steenhoek, Kalpathy Sivaraman, Renata Saldivar Gonzalez, Yevhen Mohylevskyy, Roshanak Zilouchian Moghaddam, Wei Le
Abstract: This paper presents the first empirical study of a vulnerability detection and fix tool with professional software developers on real projects that they own. We implemented DeepVulGuard, an IDE-integrated tool based on state-of-the-art detection and fix models, and show that it has promising performance on benchmarks of historic vulnerability data. DeepVulGuard scans code for vulnerabilities (including identifying the vulnerability type and vulnerable region of code), suggests fixes, provides natural-language explanations for alerts and fixes, leveraging chat interfaces. We recruited 17 professional software developers at Microsoft, observed their usage of the tool on their code, and conducted interviews to assess the tool's usefulness, speed, trust, relevance, and workflow integration. We also gathered detailed qualitative feedback on users' perceptions and their desired features. Study participants scanned a total of 24 projects, 6.9k files, and over 1.7 million lines of source code, and generated 170 alerts and 50 fix suggestions. We find that although state-of-the-art AI-powered detection and fix tools show promise, they are not yet practical for real-world use due to a high rate of false positives and non-applicable fixes. User feedback reveals several actionable pain points, ranging from incomplete context to lack of customization for the user's codebase. Additionally, we explore how AI features, including confidence scores, explanations, and chat interaction, can apply to vulnerability detection and fixing. Based on these insights, we offer practical recommendations for evaluating and deploying AI detection and fix models. Our code and data are available at https://doi.org/10.6084/m9.figshare.26367139.
Authors: Dominik Chevalier, Marie-Pier C\^ot\'e
Abstract: Gradient boosting for decision tree algorithms are increasingly used in actuarial applications as they show superior predictive performance over traditional generalised linear models. Many enhancements to the first gradient boosting machine algorithm exist. We present in a unified notation, and contrast, all the existing point and probabilistic gradient boosting for decision tree algorithms: GBM, XGBoost, DART, LightGBM, CatBoost, EGBM, PGBM, XGBoostLSS, cyclic GBM, and NGBoost. In this comprehensive numerical study, we compare their performance on five publicly available datasets for claim frequency and severity, of various sizes and comprising different numbers of (high cardinality) categorical variables. We explain how varying exposure-to-risk can be handled with boosting in frequency models. We compare the algorithms on the basis of computational efficiency, predictive performance, and model adequacy. LightGBM and XGBoostLSS win in terms of computational efficiency. CatBoost sometimes improves predictive performance, especially in the presence of high cardinality categorical variables, common in actuarial science. The fully interpretable EGBM achieves competitive predictive performance compared to the black box algorithms considered. We find that there is no trade-off between model adequacy and predictive accuracy: both are achievable simultaneously.
Authors: Huiyu Duan, Qiang Hu, Jiarui Wang, Liu Yang, Zitong Xu, Lu Liu, Xiongkuo Min, Chunlei Cai, Tianxiao Ye, Xiaoyun Zhang, Guangtao Zhai
Abstract: The rapid growth of user-generated content (UGC) videos has produced an urgent need for effective video quality assessment (VQA) algorithms to monitor video quality and guide optimization and recommendation procedures. However, current VQA models generally only give an overall rating for a UGC video, which lacks fine-grained labels for serving video processing and recommendation applications. To address the challenges and promote the development of UGC videos, we establish the first large-scale Fine-grained Video quality assessment Database, termed FineVD, which comprises 6104 UGC videos with fine-grained quality scores and descriptions across multiple dimensions. Based on this database, we propose a Fine-grained Video Quality assessment (FineVQ) model to learn the fine-grained quality of UGC videos, with the capabilities of quality rating, quality scoring, and quality attribution. Extensive experimental results demonstrate that our proposed FineVQ can produce fine-grained video-quality results and achieve state-of-the-art performance on FineVD and other commonly used UGC-VQA datasets.
Authors: Phi Vu Tran
Abstract: While modern visual recognition systems have made significant advancements, many continue to struggle with the open problem of learning from few exemplars. This paper focuses on the task of object detection in the setting where object classes follow a natural long-tailed distribution. Existing methods for long-tailed detection resort to external ImageNet labels to augment the low-shot training instances. However, such dependency on a large labeled database has limited utility in practical scenarios. We propose a versatile and scalable approach to leverage optional unlabeled images, which are easy to collect without the burden of human annotations. Our SimLTD framework is straightforward and intuitive, and consists of three simple steps: (1) pre-training on abundant head classes; (2) transfer learning on scarce tail classes; and (3) fine-tuning on a sampled set of both head and tail classes. Our approach can be viewed as an improved head-to-tail model transfer paradigm without the added complexities of meta-learning or knowledge distillation, as was required in past research. By harnessing supplementary unlabeled images, without extra image labels, SimLTD establishes new record results on the challenging LVIS v1 benchmark across both supervised and semi-supervised settings.
Authors: Lars van der Laan, David Hubbard, Allen Tran, Nathan Kallus, Aur\'elien Bibaut
Abstract: Estimating long-term causal effects from short-term data is essential for decision-making in healthcare, economics, and industry, where long-term follow-up is often infeasible. Markov Decision Processes (MDPs) offer a principled framework for modeling outcomes as sequences of states, actions, and rewards over time. We introduce a semiparametric extension of Double Reinforcement Learning (DRL) for statistically efficient, model-robust inference on linear functionals of the Q-function, such as policy values, in infinite-horizon, time-homogeneous MDPs. By imposing semiparametric structure on the Q-function, our method relaxes the strong state overlap assumptions required by fully nonparametric approaches, improving efficiency and stability. To address computational and robustness challenges of minimax nuisance estimation, we develop a novel debiased plug-in estimator based on isotonic Bellman calibration, which integrates fitted Q-iteration with an isotonic regression step. This procedure leverages the Q-function as a data-driven dimension reduction, debiases all linear functionals of interest simultaneously, and enables nonparametric inference without explicit nuisance function estimation. Bellman calibration generalizes isotonic calibration to MDPs and may be of independent interest for prediction in reinforcement learning. Finally, we show that model selection for the Q-function incurs only second-order bias and extend the adaptive debiased machine learning (ADML) framework to MDPs for data-driven learning of semiparametric structure.
Authors: BG Tong
Abstract: Objective: This study explores a novel deep learning approach for EEG analysis and perceptual state guidance, inspired by Level of Detail (LOD) theory. The goal is to improve perceptual state identification accuracy and advance personalized psychological therapy. Methods: Portable EEG devices and music rhythm signals were used for data collection. LOD theory was applied to dynamically adjust EEG signal processing, extracting core perceptual features. A Unity-based software system integrated EEG data with audio materials. The deep learning model combined a CNN for feature extraction and classification, and a DQN for reinforcement learning to optimize rhythm adjustments. Results: The CNN achieved 94.05% accuracy in perceptual state classification. The DQN guided subjects to target states with a 92.45% success rate, averaging 13.2 rhythm cycles. However, only 50% of users reported psychological alignment with the target state, indicating room for improvement. Discussion: The results validate the potential of LOD-based EEG biofeedback. Limitations include dataset source, label subjectivity, and reward function optimization. Future work will expand to diverse subjects, incorporate varied musical elements, and refine reward functions for better generalization and personalization.
Authors: Saikat Roy, Mahmoud Mostapha, Radu Miron, Matt Holbrook, Mariappan Nadar
Abstract: Plug-and-play approaches to solving inverse problems such as restoration and super-resolution have recently benefited from Diffusion-based generative priors for natural as well as medical images. However, solutions often use the standard albeit computationally intensive route of training and inferring with the whole image on the diffusion prior. While patch-based approaches to evaluating diffusion priors in plug-and-play methods have received some interest, they remain an open area of study. In this work, we explore the feasibility of the usage of patches for training and inference of a diffusion prior on MRI images. We explore the minor adaptation necessary for artifact avoidance, the performance and the efficiency of memory usage of patch-based methods as well as the adaptability of whole image training to patch-based evaluation - evaluating across multiple plug-and-play methods, tasks and datasets.
Authors: Amogh Joshi, Sourav Sanyal, Kaushik Roy
Abstract: The integration of human-intuitive interactions into autonomous systems has been limited. Traditional Natural Language Processing (NLP) systems struggle with context and intent understanding, severely restricting human-robot interaction. Recent advancements in Large Language Models (LLMs) have transformed this dynamic, allowing for intuitive and high-level communication through speech and text, and bridging the gap between human commands and robotic actions. Additionally, autonomous navigation has emerged as a central focus in robotics research, with artificial intelligence (AI) increasingly being leveraged to enhance these systems. However, existing AI-based navigation algorithms face significant challenges in latency-critical tasks where rapid decision-making is critical. Traditional frame-based vision systems, while effective for high-level decision-making, suffer from high energy consumption and latency, limiting their applicability in real-time scenarios. Neuromorphic vision systems, combining event-based cameras and spiking neural networks (SNNs), offer a promising alternative by enabling energy-efficient, low-latency navigation. Despite their potential, real-world implementations of these systems, particularly on physical platforms such as drones, remain scarce. In this work, we present Neuro-LIFT, a real-time neuromorphic navigation framework implemented on a Parrot Bebop2 quadrotor. Leveraging an LLM for natural language processing, Neuro-LIFT translates human speech into high-level planning commands which are then autonomously executed using event-based neuromorphic vision and physics-driven planning. Our framework demonstrates its capabilities in navigating in a dynamic environment, avoiding obstacles, and adapting to human instructions in real-time.
Authors: Yuanchen Yuan, Jin Cheng, N\'uria Armengol Urp\'i, Stelian Coros
Abstract: Enabling legged robots to perform non-prehensile loco-manipulation is crucial for enhancing their versatility. Learning behaviors such as whole-body object pushing often requires sophisticated planning strategies or extensive task-specific reward shaping, especially in unstructured environments. In this work, we present CAIMAN, a practical reinforcement learning framework that encourages the agent to gain control over other entities in the environment. CAIMAN leverages causal action influence as an intrinsic motivation objective, allowing legged robots to efficiently acquire object pushing skills even under sparse task rewards. We employ a hierarchical control strategy, combining a low-level locomotion module with a high-level policy that generates task-relevant velocity commands and is trained to maximize the intrinsic reward. To estimate causal action influence, we learn the dynamics of the environment by integrating a kinematic prior with data collected during training.We empirically demonstrate CAIMAN's superior sample efficiency and adaptability to diverse scenarios in simulation, as well as its successful transfer to real-world systems without further fine-tuning.
Authors: Tairan He, Jiawei Gao, Wenli Xiao, Yuanhang Zhang, Zi Wang, Jiashun Wang, Zhengyi Luo, Guanqi He, Nikhil Sobanbab, Chaoyi Pan, Zeji Yi, Guannan Qu, Kris Kitani, Jessica Hodgins, Linxi "Jim" Fan, Yuke Zhu, Changliu Liu, Guanya Shi
Abstract: Humanoid robots hold the potential for unparalleled versatility in performing human-like, whole-body skills. However, achieving agile and coordinated whole-body motions remains a significant challenge due to the dynamics mismatch between simulation and the real world. Existing approaches, such as system identification (SysID) and domain randomization (DR) methods, often rely on labor-intensive parameter tuning or result in overly conservative policies that sacrifice agility. In this paper, we present ASAP (Aligning Simulation and Real-World Physics), a two-stage framework designed to tackle the dynamics mismatch and enable agile humanoid whole-body skills. In the first stage, we pre-train motion tracking policies in simulation using retargeted human motion data. In the second stage, we deploy the policies in the real world and collect real-world data to train a delta (residual) action model that compensates for the dynamics mismatch. Then, ASAP fine-tunes pre-trained policies with the delta action model integrated into the simulator to align effectively with real-world dynamics. We evaluate ASAP across three transfer scenarios: IsaacGym to IsaacSim, IsaacGym to Genesis, and IsaacGym to the real-world Unitree G1 humanoid robot. Our approach significantly improves agility and whole-body coordination across various dynamic motions, reducing tracking error compared to SysID, DR, and delta dynamics learning baselines. ASAP enables highly agile motions that were previously difficult to achieve, demonstrating the potential of delta action learning in bridging simulation and real-world dynamics. These results suggest a promising sim-to-real direction for developing more expressive and agile humanoids.
Authors: Hichem Ammar Khodja, Fr\'ed\'eric B\'echet, Quentin Brabant, Alexis Nasr, Gw\'enol\'e Lecorv\'e
Abstract: This paper explores the robustness of language models (LMs) to variations in the temporal context within factual knowledge. It examines whether LMs can correctly associate a temporal context with a past fact valid over a defined period, by asking them to differentiate correct from incorrect contexts. The accuracy of LMs is analyzed along two dimensions: the distance of the incorrect context from the validity period and the granularity of the context. To this end, a dataset called TimeStress is introduced, enabling the evaluation of 18 diverse LMs. Results reveal that the best LM achieves perfect accuracy for only 6% of the studied facts, with critical errors that humans would not make. This work highlights the limitations of current LMs in temporal representation. We provide all data and code for further research.
Authors: Haocheng Xi, Shuo Yang, Yilong Zhao, Chenfeng Xu, Muyang Li, Xiuyu Li, Yujun Lin, Han Cai, Jintao Zhang, Dacheng Li, Jianfei Chen, Ion Stoica, Kurt Keutzer, Song Han
Abstract: Diffusion Transformers (DiTs) dominate video generation but their high computational cost severely limits real-world applicability, usually requiring tens of minutes to generate a few seconds of video even on high-performance GPUs. This inefficiency primarily arises from the quadratic computational complexity of 3D Full Attention with respect to the context length. In this paper, we propose a training-free framework termed Sparse VideoGen (SVG) that leverages the inherent sparsity in 3D Full Attention to boost inference efficiency. We reveal that the attention heads can be dynamically classified into two groups depending on distinct sparse patterns: (1) Spatial Head, where only spatially-related tokens within each frame dominate the attention output, and (2) Temporal Head, where only temporally-related tokens across different frames dominate. Based on this insight, SVG proposes an online profiling strategy to capture the dynamic sparse patterns and predicts the type of attention head. Combined with a novel hardware-efficient tensor layout transformation and customized kernel implementations, SVG achieves up to 2.28x and 2.33x end-to-end speedup on CogVideoX-v1.5 and HunyuanVideo, respectively, while preserving generation quality. Our code is open-sourced and is available at https://github.com/svg-project/Sparse-VideoGen
Authors: Samir M. Perlaza, Gaetan Bisson
Abstract: In this paper, closed-form expressions are presented for the variation of the expectation of a given function due to changes in the probability measure used for the expectation. They unveil interesting connections with Gibbs probability measures, mutual information, and lautum information.
Authors: Rupesh Dulal, Rabin Dulal
Abstract: Identifying the extent of brain tumors is a significant challenge in brain cancer treatment. The main difficulty is in the approximate detection of tumor size. Magnetic resonance imaging (MRI) has become a critical diagnostic tool. However, manually detecting the boundaries of brain tumors from MRI scans is a labor-intensive task that requires extensive expertise. Deep learning and computer-aided detection techniques have led to notable advances in machine learning for this purpose. In this paper, we propose a modified You Only Look Once (YOLOv8) model to accurately detect the tumors within the MRI images. The proposed model replaced the Non-Maximum Suppression (NMS) algorithm with a Real-Time Detection Transformer (RT- DETR) in the detection head. NMS filters out redundant or overlapping bounding boxes in the detected tumors, but they are hand-designed and pre-set. RT-DETR removes hand-designed components. The second improvement was made by replacing the normal convolution block with ghost convolution. Ghost Convolution reduces computational and memory costs while maintaining high accuracy and enabling faster inference, making it ideal for resource-constrained environments and real-time applications. The third improvement was made by introducing a vision transformer block in the backbone of YOLOv8 to extract context-aware features. We used a publicly available dataset of brain tumors in the proposed model. The proposed model performed better than the original YOLOv8 model and also performed better than other object detectors (Faster R- CNN, Mask R-CNN, YOLO, YOLOv3, YOLOv4, YOLOv5, SSD, RetinaNet, EfficientDet, and DETR). The proposed model achieved 0.91 mAP (mean Average Precision)@0.5.
Authors: Hui Lan, Haoge Chang, Eleanor Dillon, Vasilis Syrgkanis
Abstract: We address the problem of estimating heterogeneous treatment effects in panel data, adopting the popular Difference-in-Differences (DiD) framework under the conditional parallel trends assumption. We propose a novel doubly robust meta-learner for the Conditional Average Treatment Effect on the Treated (CATT), reducing the estimation to a convex risk minimization problem involving a set of auxiliary models. Our framework allows for the flexible estimation of the CATT, when conditioning on any subset of variables of interest using generic machine learning. Leveraging Neyman orthogonality, our proposed approach is robust to estimation errors in the auxiliary models. As a generalization to our main result, we develop a meta-learning approach for the estimation of general conditional functionals under covariate shift. We also provide an extension to the instrumented DiD setting with non-compliance. Empirical results demonstrate the superiority of our approach over existing baselines.
Authors: Jaiden Fairoze, Guillermo Ortiz-Jimenez, Mel Vecerik, Somesh Jha, Sven Gowal
Abstract: This work investigates the theoretical boundaries of creating publicly-detectable schemes to enable the provenance of watermarked imagery. Metadata-based approaches like C2PA provide unforgeability and public-detectability. ML techniques offer robust retrieval and watermarking. However, no existing scheme combines robustness, unforgeability, and public-detectability. In this work, we formally define such a scheme and establish its existence. Although theoretically possible, we find that at present, it is intractable to build certain components of our scheme without a leap in deep learning capabilities. We analyze these limitations and propose research directions that need to be addressed before we can practically realize robust and publicly-verifiable provenance.
Authors: Andreas Galanis, Leslie Ann Goldberg, Xusheng Zhang
Abstract: Consider a $k$-SAT formula $\Phi$ where every variable appears at most $d$ times, and let $\sigma$ be a satisfying assignment of $\Phi$ sampled proportionally to $e^{\beta m(\sigma)}$ where $m(\sigma)$ is the number of variables set to true and $\beta$ is a real parameter. Given $\Phi$ and $\sigma$, can we learn the value of $\beta$ efficiently? This problem falls into a recent line of works about single-sample ("one-shot") learning of Markov random fields. The $k$-SAT setting we consider here was recently studied by Galanis, Kandiros, and Kalavasis (SODA'24) where they showed that single-sample learning is possible when roughly $d\leq 2^{k/6.45}$ and impossible when $d\geq (k+1) 2^{k-1}$. Crucially, for their impossibility results they used the existence of unsatisfiable instances which, aside from the gap in $d$, left open the question of whether the feasibility threshold for one-shot learning is dictated by the satisfiability threshold of $k$-SAT formulas of bounded degree. Our main contribution is to answer this question negatively. We show that one-shot learning for $k$-SAT is infeasible well below the satisfiability threshold; in fact, we obtain impossibility results for degrees $d$ as low as $k^2$ when $\beta$ is sufficiently large, and bootstrap this to small values of $\beta$ when $d$ scales exponentially with $k$, via a probabilistic construction. On the positive side, we simplify the analysis of the learning algorithm and obtain significantly stronger bounds on $d$ in terms of $\beta$. In particular, for the uniform case $\beta\rightarrow 0$ that has been studied extensively in the sampling literature, our analysis shows that learning is possible under the condition $d\lesssim 2^{k/2}$. This is nearly optimal (up to constant factors) in the sense that it is known that sampling a uniformly-distributed satisfying assignment is NP-hard for $d\gtrsim 2^{k/2}$.
Authors: Mark Beliaev, Victor Yang, Madhura Raju, Jiachen Sun, Xinghai Hu
Abstract: In this study, we tackle industry challenges in video content classification by exploring and optimizing GPT-based models for zero-shot classification across seven critical categories of video quality. We contribute a novel approach to improving GPT's performance through prompt optimization and policy refinement, demonstrating that simplifying complex policies significantly reduces false negatives. Additionally, we introduce a new decomposition-aggregation-based prompt engineering technique, which outperforms traditional single-prompt methods. These experiments, conducted on real industry problems, show that thoughtful prompt design can substantially enhance GPT's performance without additional finetuning, offering an effective and scalable solution for improving video classification.
Authors: Huayi Wang, Zirui Wang, Junli Ren, Qingwei Ben, Tao Huang, Weinan Zhang, Jiangmiao Pang
Abstract: Traversing risky terrains with sparse footholds poses a significant challenge for humanoid robots, requiring precise foot placements and stable locomotion. Existing learning-based approaches often struggle on such complex terrains due to sparse foothold rewards and inefficient learning processes. To address these challenges, we introduce BeamDojo, a reinforcement learning (RL) framework designed for enabling agile humanoid locomotion on sparse footholds. BeamDojo begins by introducing a sampling-based foothold reward tailored for polygonal feet, along with a double critic to balancing the learning process between dense locomotion rewards and sparse foothold rewards. To encourage sufficient trial-and-error exploration, BeamDojo incorporates a two-stage RL approach: the first stage relaxes the terrain dynamics by training the humanoid on flat terrain while providing it with task-terrain perceptive observations, and the second stage fine-tunes the policy on the actual task terrain. Moreover, we implement a onboard LiDAR-based elevation map to enable real-world deployment. Extensive simulation and real-world experiments demonstrate that BeamDojo achieves efficient learning in simulation and enables agile locomotion with precise foot placement on sparse footholds in the real world, maintaining a high success rate even under significant external disturbances.
Authors: Sheng-Yu Huang, Zi-Ting Chou, Yu-Chiang Frank Wang
Abstract: When performing 3D inpainting using novel-view rendering methods like Neural Radiance Field (NeRF) or 3D Gaussian Splatting (3DGS), how to achieve texture and geometry consistency across camera views has been a challenge. In this paper, we propose a framework of 3D Gaussian Inpainting with Depth-Guided Cross-View Consistency (3DGIC) for cross-view consistent 3D inpainting. Guided by the rendered depth information from each training view, our 3DGIC exploits background pixels visible across different views for updating the inpainting mask, allowing us to refine the 3DGS for inpainting purposes.Through extensive experiments on benchmark datasets, we confirm that our 3DGIC outperforms current state-of-the-art 3D inpainting methods quantitatively and qualitatively.
Authors: Annabelle Michael Carrell, Albert Gong, Abhishek Shetty, Raaz Dwivedi, Lester Mackey
Abstract: The goal in thinning is to summarize a dataset using a small set of representative points. Remarkably, sub-Gaussian thinning algorithms like Kernel Halving and Compress can match the quality of uniform subsampling while substantially reducing the number of summary points. However, existing guarantees cover only a restricted range of distributions and kernel-based quality measures and suffer from pessimistic dimension dependence. To address these deficiencies, we introduce a new low-rank analysis of sub-Gaussian thinning that applies to any distribution and any kernel, guaranteeing high-quality compression whenever the kernel or data matrix is approximately low-rank. To demonstrate the broad applicability of the techniques, we design practical sub-Gaussian thinning approaches that improve upon the best known guarantees for approximating attention in transformers, accelerating stochastic gradient training through reordering, and distinguishing distributions in near-linear time.
Authors: Xialin He, Runpei Dong, Zixuan Chen, Saurabh Gupta
Abstract: Automatic fall recovery is a crucial prerequisite before humanoid robots can be reliably deployed. Hand-designing controllers for getting up is difficult because of the varied configurations a humanoid can end up in after a fall and the challenging terrains humanoid robots are expected to operate on. This paper develops a learning framework to produce controllers that enable humanoid robots to get up from varying configurations on varying terrains. Unlike previous successful applications of learning to humanoid locomotion, the getting-up task involves complex contact patterns (which necessitates accurately modeling of the collision geometry) and sparser rewards. We address these challenges through a two-phase approach that induces a curriculum. The first stage focuses on discovering a good getting-up trajectory under minimal constraints on smoothness or speed / torque limits. The second stage then refines the discovered motions into deployable (i.e. smooth and slow) motions that are robust to variations in initial configuration and terrains. We find these innovations enable a real-world G1 humanoid robot to get up from two main situations that we considered: a) lying face up and b) lying face down, both tested on flat, deformable, slippery surfaces and slopes (e.g., sloppy grass and snowfield). This is one of the first successful demonstrations of learned getting-up policies for human-sized humanoid robots in the real world.
Authors: Bryan L. M. de Oliveira, Luana G. B. Martins, Bruno Brand\~ao, Luckeciano C. Melo
Abstract: Large language models excel at following explicit instructions, but they often struggle with ambiguous or incomplete user requests, defaulting to verbose, generic responses instead of seeking clarification. We introduce InfoQuest, a multi-turn chat benchmark designed to evaluate how dialogue agents handle hidden context in open-ended user requests. This benchmark presents intentionally ambiguous scenarios that require models to engage in information-seeking dialogue by asking clarifying questions before providing appropriate responses. Our evaluation of both open and closed models reveals that, while proprietary models generally perform better, all current assistants struggle to gather critical information effectively. They often require multiple turns to infer user intent and frequently default to generic responses without proper clarification. We provide a systematic methodology for generating diverse scenarios and evaluating models' information-seeking capabilities, which can be leveraged to automatically generate data for self-improvement. We also offer insights into the current limitations of language models in handling ambiguous requests through multi-turn interactions.
Authors: Yingheng Tang, Wenbin Xu, Jie Cao, Weilu Gao, Steve Farrell, Benjamin Erichson, Michael W. Mahoney, Andy Nonaka, Zhi Yao
Abstract: Understanding and predicting the properties of inorganic materials is crucial for accelerating advancements in materials science and driving applications in energy, electronics, and beyond. Integrating material structure data with language-based information through multi-modal large language models (LLMs) offers great potential to support these efforts by enhancing human-AI interaction. However, a key challenge lies in integrating atomic structures at full resolution into LLMs. In this work, we introduce MatterChat, a versatile structure-aware multi-modal LLM that unifies material structural data and textual inputs into a single cohesive model. MatterChat employs a bridging module to effectively align a pretrained machine learning interatomic potential with a pretrained LLM, reducing training costs and enhancing flexibility. Our results demonstrate that MatterChat significantly improves performance in material property prediction and human-AI interaction, surpassing general-purpose LLMs such as GPT-4. We also demonstrate its usefulness in applications such as more advanced scientific reasoning and step-by-step material synthesis.
Authors: Jingqiu Ding, Yiding Hua, Lucas Slot, David Steurer
Abstract: We investigate implications of the (extended) low-degree conjecture (recently formalized in [MW23]) in the context of the symmetric stochastic block model. Assuming the conjecture holds, we establish that no polynomial-time algorithm can weakly recover community labels below the Kesten-Stigum (KS) threshold. In particular, we rule out polynomial-time estimators that, with constant probability, achieve correlation with the true communities that is significantly better than random. Whereas, above the KS threshold, polynomial-time algorithms are known to achieve constant correlation with the true communities with high probability[Mas14,AS15]. To our knowledge, we provide the first rigorous evidence for the sharp transition in recovery rate for polynomial-time algorithms at the KS threshold. Notably, under a stronger version of the low-degree conjecture, our lower bound remains valid even when the number of blocks diverges. Furthermore, our results provide evidence of a computational-to-statistical gap in learning the parameters of stochastic block models. In contrast to prior work, which either (i) rules out polynomial-time algorithms for hypothesis testing with 1-o(1) success probability [Hopkins18, BBK+21a] under the low-degree conjecture, or (ii) rules out low-degree polynomials for learning the edge connection probability matrix [LG23], our approach provides stronger lower bounds on the recovery and learning problem. Our proof combines low-degree lower bounds from [Hopkins18, BBK+21a] with graph splitting and cross-validation techniques. In order to rule out general recovery algorithms, we employ the correlation preserving projection method developed in [HS17].
Authors: Zhaoxing Li, Vahid Yazdanpanah, Jindi Wang, Wen Gu, Lei Shi, Alexandra I. Cristea, Sarah Kiden, Sebastian Stein
Abstract: The integration of AI in education offers significant potential to enhance learning efficiency. Large Language Models (LLMs), such as ChatGPT, Gemini, and Llama, allow students to query a wide range of topics, providing unprecedented flexibility. However, LLMs face challenges, such as handling varying content relevance and lack of personalization. To address these challenges, we propose TutorLLM, a personalized learning recommender LLM system based on Knowledge Tracing (KT) and Retrieval-Augmented Generation (RAG). The novelty of TutorLLM lies in its unique combination of KT and RAG techniques with LLMs, which enables dynamic retrieval of context-specific knowledge and provides personalized learning recommendations based on the student's personal learning state. Specifically, this integration allows TutorLLM to tailor responses based on individual learning states predicted by the Multi-Features with Latent Relations BERT-based KT (MLFBK) model and to enhance response accuracy with a Scraper model. The evaluation includes user assessment questionnaires and performance metrics, demonstrating a 10% improvement in user satisfaction and a 5\% increase in quiz scores compared to using general LLMs alone.
Authors: Xin Zhang, Liangxiu Han, Stephen White, Saad Hassan, Philip A Kalra, James Ritchie, Carl Diver, Jennie Shorley
Abstract: Acute Coronary Syndromes (ACS), including ST-segment elevation myocardial infarctions (STEMI) and non-ST-segment elevation myocardial infarctions (NSTEMI), remain a leading cause of mortality worldwide. Traditional cardiovascular risk scores rely primarily on clinical data, often overlooking environmental influences like air pollution that significantly impact heart health. Moreover, integrating complex time-series environmental data with clinical records is challenging. We introduce TabulaTime, a multimodal deep learning framework that enhances ACS risk prediction by combining clinical risk factors with air pollution data. TabulaTime features three key innovations: First, it integrates time-series air pollution data with clinical tabular data to improve prediction accuracy. Second, its PatchRWKV module automatically extracts complex temporal patterns, overcoming limitations of traditional feature engineering while maintaining linear computational complexity. Third, attention mechanisms enhance interpretability by revealing interactions between clinical and environmental factors. Experimental results show that TabulaTime improves prediction accuracy by over 20% compared to conventional models such as CatBoost, Random Forest, and LightGBM, with air pollution data alone contributing over a 10% improvement. Feature importance analysis identifies critical predictors including previous angina, systolic blood pressure, PM10, and NO2. Overall, TabulaTime bridges clinical and environmental insights, supporting personalized prevention strategies and informing public health policies to mitigate ACS risk.
Authors: Ziwei Wang, Siyang Li, Xiaoqing Chen, Wei Li, Dongrui Wu
Abstract: Electroencephalography (EEG)-based brain-computer interfaces (BCIs) enable neural interaction by decoding brain activity for external communication. Motor imagery (MI) decoding has received significant attention due to its intuitive mechanism. However, most existing models rely on single-stream architectures and overlook the multi-view nature of EEG signals, leading to limited performance and generalization. We propose a multi-view contrastive network (MVCNet), a dual-branch architecture that parallelly integrates CNN and Transformer models to capture both local spatial-temporal features and global temporal dependencies. To enhance the informativeness of training data, MVCNet incorporates a unified augmentation pipeline across time, frequency, and spatial domains. Two contrastive modules are further introduced: a cross-view contrastive module that enforces consistency of original and augmented views, and a cross-model contrastive module that aligns features extracted from both branches. Final representations are fused and jointly optimized by contrastive and classification losses. Experiments on five public MI datasets across three scenarios demonstrate that MVCNet consistently outperforms seven state-of-the-art MI decoding networks, highlighting its effectiveness and generalization ability. MVCNet provides a robust solution for MI decoding by integrating multi-view information and dual-branch modeling, contributing to the development of more reliable BCI systems.
Authors: Sumei Fan, Deyun Zhang, Yue Wang, Shijia Geng, Kun Lu, Meng Sang, Weilun Xu, Haixue Wang, Qinghao Zhao, Chuandong Cheng, Peng Wang, Shenda Hong
Abstract: Home-based single-lead AI-ECG devices have enabled continuous, real-world cardiac monitoring. However, the accuracy of parameter calculations from single-lead AI-ECG algorithm remains to be fully validated, which is critical for conditions such as Long QT Syndrome (LQTS) and First-Degree Atrioventricular Block (AVBI). In this multicenter study, we assessed FeatureDB, an ECG measurements computation algorithm, in the context of single-lead monitoring using three annotated datasets: PTB-XL+ (n=21,354), CSE (n=105), and HeartVoice-ECG-lite (n=369). FeatureDB showed strong correlation with standard ECG machines (12SL and Uni-G) in key measurements (PR, QRS, QT, QTc), and high agreement confirmed by Bland-Altman analysis. In detecting LQTS (AUC=0.786) and AVBI (AUC=0.684), FeatureDB demonstrated diagnostic performance comparable to commercial ECG systems (12SL: 0.859/0.716; Uni-G: 0.817/0.605), significantly outperforming ECGDeli (0.501/0.569). Notably, FeatureDB can operate locally on resource-limited devices, facilitating use in low-connectivity settings. These findings confirm the clinical reliability of FeatureDB for single-lead ECG diagnostics and highlight its potential to bridge traditional ECG diagnostics with wearable technology for scalable cardiovascular monitoring and early intervention.
Authors: Bram van Dartel, Marc Damie, Florian Hahn
Abstract: Among all privacy attacks against Machine Learning (ML), membership inference attacks (MIA) attracted the most attention. In these attacks, the attacker is given an ML model and a data point, and they must infer whether the data point was used for training. The attacker also has an auxiliary dataset to tune their inference algorithm. Attack papers commonly simulate setups in which the attacker's and the target's datasets are sampled from the same distribution. This setting is convenient to perform experiments, but it rarely holds in practice. ML literature commonly starts with similar simplifying assumptions (i.e., "i.i.d." datasets), and later generalizes the results to support heterogeneous data distributions. Similarly, our work makes a first step in the generalization of the MIA evaluation to heterogeneous data. First, we design a metric to measure the heterogeneity between any pair of tabular data distributions. This metric provides a continuous scale to analyze the phenomenon. Second, we compare two methodologies to simulate a data heterogeneity between the target and the attacker. These setups provide opposite performances: 90% attack accuracy vs. 50% (i.e., random guessing). Our results show that the MIA accuracy depends on the experimental setup; and even if research on MIA considers heterogeneous data setups, we have no standardized baseline of how to simulate it. The lack of such a baseline for MIA experiments poses a significant challenge to risk assessments in real-world machine learning scenarios.
Authors: Moo Jin Kim, Chelsea Finn, Percy Liang
Abstract: Recent vision-language-action models (VLAs) build upon pretrained vision-language models and leverage diverse robot datasets to demonstrate strong task execution, language following ability, and semantic generalization. Despite these successes, VLAs struggle with novel robot setups and require fine-tuning to achieve good performance, yet how to most effectively fine-tune them is unclear given many possible strategies. In this work, we study key VLA adaptation design choices such as different action decoding schemes, action representations, and learning objectives for fine-tuning, using OpenVLA as our representative base model. Our empirical analysis informs an Optimized Fine-Tuning (OFT) recipe that integrates parallel decoding, action chunking, a continuous action representation, and a simple L1 regression-based learning objective to altogether improve inference efficiency, policy performance, and flexibility in the model's input-output specifications. We propose OpenVLA-OFT, an instantiation of this recipe, which sets a new state of the art on the LIBERO simulation benchmark, significantly boosting OpenVLA's average success rate across four task suites from 76.5% to 97.1% while increasing action generation throughput by 26$\times$. In real-world evaluations, our fine-tuning recipe enables OpenVLA to successfully execute dexterous, high-frequency control tasks on a bimanual ALOHA robot and outperform other VLAs ($\pi_0$ and RDT-1B) fine-tuned using their default recipes, as well as strong imitation learning policies trained from scratch (Diffusion Policy and ACT) by up to 15% (absolute) in average success rate. We release code for OFT and pretrained model checkpoints at https://openvla-oft.github.io/.
Authors: Saman Khamesian, Asiful Arefeen, Stephanie M. Carpenter, Hassan Ghasemzadeh
Abstract: Maintaining a balanced diet is essential for overall health, yet many individuals struggle with meal planning due to nutritional complexity, time constraints, and lack of dietary knowledge. Personalized food recommendations can help address these challenges by tailoring meal plans to individual preferences, habits, and dietary restrictions. However, existing dietary recommendation systems often lack adaptability, fail to consider real-world constraints such as food ingredient availability, and require extensive user input, making them impractical for sustainable and scalable daily use. To address these limitations, we introduce NutriGen, a framework based on large language models (LLM) designed to generate personalized meal plans that align with user-defined dietary preferences and constraints. By building a personalized nutrition database and leveraging prompt engineering, our approach enables LLMs to incorporate reliable nutritional references like the USDA nutrition database while maintaining flexibility and ease-of-use. We demonstrate that LLMs have strong potential in generating accurate and user-friendly food recommendations, addressing key limitations in existing dietary recommendation systems by providing structured, practical, and scalable meal plans. Our evaluation shows that Llama 3.1 8B and GPT-3.5 Turbo achieve the lowest percentage errors of 1.55\% and 3.68\%, respectively, producing meal plans that closely align with user-defined caloric targets while minimizing deviation and improving precision. Additionally, we compared the performance of DeepSeek V3 against several established models to evaluate its potential in personalized nutrition planning.
Authors: Ivan Sviridov, Konstantin Egorov
Abstract: Cardiovascular diseases (CVDs) are disorders impacting the heart and circulatory system. These disorders are the foremost and continuously escalating cause of mortality worldwide. One of the main tasks when working with CVDs is analyzing and identifying pathologies on a 12-lead electrocardiogram (ECG) with a standard 10-second duration. Using machine learning (ML) in automatic ECG analysis increases CVD diagnostics' availability, speed, and accuracy. However, the most significant difficulty in developing ML models is obtaining a sufficient training dataset. Due to the limitations of medical data usage, such as expensiveness, errors, the ambiguity of labels, imbalance of classes, and privacy issues, utilizing synthetic samples depending on specific pathologies bypasses these restrictions and improves algorithm quality. Existing solutions for the conditional generation of ECG signals are mainly built on Generative Adversarial Networks (GANs), and only a few papers consider the architectures based on Variational Autoencoders (VAEs), showing comparable results in recent works. This paper proposes the publicly available conditional Nouveau VAE model for ECG signal generation (cNVAE-ECG), which produces high-resolution ECGs with multiple pathologies. We provide an extensive comparison of the proposed model on various practical downstream tasks, including transfer learning scenarios showing an area under the receiver operating characteristic (AUROC) increase up to 2% surpassing GAN-like competitors.
Authors: Jingyuan Xue, Longfei Wei, Fang Sheng, Jianfei Zhang
Abstract: Battery degradation significantly impacts the reliability and efficiency of energy storage systems, particularly in electric vehicles and industrial applications. Predicting the remaining useful life (RUL) of lithium-ion batteries is crucial for optimizing maintenance schedules, reducing costs, and improving safety. Traditional RUL prediction methods often struggle with nonlinear degradation patterns and uncertainty quantification. To address these challenges, we propose a hybrid survival analysis framework integrating survival data reconstruction, survival model learning, and survival probability estimation. Our approach transforms battery voltage time series into time-to-failure data using path signatures. The multiple Cox-based survival models and machine-learning-based methods, such as DeepHit and MTLR, are learned to predict battery failure-free probabilities over time. Experiments conducted on the Toyota battery and NASA battery datasets demonstrate the effectiveness of our approach, achieving high time-dependent AUC and concordance index (C-Index) while maintaining a low integrated Brier score. The data and source codes for this work are available to the public at https://github.com/thinkxca/rul.
Authors: Lahiru Akmeemana, Chamodya Attanayake, Husni Faiz, Sandareka Wickramanayake
Abstract: The transition to microservices has revolutionized software architectures, offering enhanced scalability and modularity. However, the distributed and dynamic nature of microservices introduces complexities in ensuring system reliability, making anomaly detection crucial for maintaining performance and functionality. Anomalies stemming from network and performance issues must be swiftly identified and addressed. Existing anomaly detection techniques often rely on statistical models or machine learning methods that struggle with the high-dimensional, interdependent data inherent in microservice applications. Current techniques and available datasets predominantly focus on system traces and logs, limiting their ability to support advanced detection models. This paper addresses these gaps by introducing the RS-Anomic dataset generated using the open-source RobotShop microservice application. The dataset captures multivariate performance metrics and response times under normal and anomalous conditions, encompassing ten types of anomalies. We propose a novel anomaly detection model called Graph Attention and LSTM-based Microservice Anomaly Detection (GAL-MAD), leveraging Graph Attention and Long Short-Term Memory architectures to capture spatial and temporal dependencies in microservices. We utilize SHAP values to localize anomalous services and identify root causes to enhance explainability. Experimental results demonstrate that GAL-MAD outperforms state-of-the-art models on the RS-Anomic dataset, achieving higher accuracy and recall across varying anomaly rates. The explanations provide actionable insights into service anomalies, which benefits system administrators.
Authors: Jingyi Wang, Duanfeng Chu, Zejian Deng, Liping Lu, Jinxiang Wang, Chen Sun
Abstract: To address the challenge of insufficient interactivity and behavioral diversity in autonomous driving decision-making, this paper proposes a Cognitive Hierarchical Agent for Reasoning and Motion Stylization (CHARMS). By leveraging Level-k game theory, CHARMS captures human-like reasoning patterns through a two-stage training pipeline comprising reinforcement learning pretraining and supervised fine-tuning. This enables the resulting models to exhibit diverse and human-like behaviors, enhancing their decision-making capacity and interaction fidelity in complex traffic environments. Building upon this capability, we further develop a scenario generation framework that utilizes the Poisson cognitive hierarchy theory to control the distribution of vehicles with different driving styles through Poisson and binomial sampling. Experimental results demonstrate that CHARMS is capable of both making intelligent driving decisions as an ego vehicle and generating diverse, realistic driving scenarios as environment vehicles. The code for CHARMS is released at https://github.com/chuduanfeng/CHARMS.
Authors: Yasmine Mustafa, Mohamed Elmahallawy, Tie Luo
Abstract: Alzheimer's disease (AD) leads to progressive cognitive decline, making early detection crucial for effective intervention. While deep learning models have shown high accuracy in AD diagnosis, their lack of interpretability limits clinical trust and adoption. This paper introduces a novel pre-model approach leveraging Jacobian Maps (JMs) within a multi-modal framework to enhance explainability and trustworthiness in AD detection. By capturing localized brain volume changes, JMs establish meaningful correlations between model predictions and well-known neuroanatomical biomarkers of AD. We validate JMs through experiments comparing a 3D CNN trained on JMs versus on traditional preprocessed data, which demonstrates superior accuracy. We also employ 3D Grad-CAM analysis to provide both visual and quantitative insights, further showcasing improved interpretability and diagnostic reliability.
Authors: Zhanwei Wang, Qunsong Zeng, Haotian Zheng, Kaibin Huang
Abstract: One of the key missions of sixth-generation (6G) mobile networks is to deploy large-scale artificial intelligence (AI) models at the network edge to provide remote-inference services for edge devices. The resultant platform, known as edge inference, will support a wide range of Internet-of-Things applications, such as autonomous driving, industrial automation, and augmented reality. Given the mission-critical and time-sensitive nature of these tasks, it is essential to design edge inference systems that are both reliable and capable of meeting stringent end-to-end (E2E) latency constraints. Existing studies, which primarily focus on communication reliability as characterized by channel outage probability, may fail to guarantee E2E performance, specifically in terms of E2E inference accuracy and latency. To address this limitation, we propose a theoretical framework that introduces and mathematically characterizes the inference outage (InfOut) probability, which quantifies the likelihood that the E2E inference accuracy falls below a target threshold. Under an E2E latency constraint, this framework establishes a fundamental tradeoff between communication overhead (i.e., uploading more sensor observations) and inference reliability as quantified by the InfOut probability. To find a tractable way to optimize this tradeoff, we derive accurate surrogate functions for InfOut probability by applying a Gaussian approximation to the distribution of the received discriminant gain. Experimental results demonstrate the superiority of the proposed design over conventional communication-centric approaches in terms of E2E inference reliability.
Authors: Anna Goldie, Azalia Mirhoseini, Hao Zhou, Irene Cai, Christopher D. Manning
Abstract: Reinforcement learning has been shown to improve the performance of large language models. However, traditional approaches like RLHF or RLAIF treat the problem as single-step. As focus shifts toward more complex reasoning and agentic tasks, language models must take multiple steps of text generation, reasoning and environment interaction before generating a solution. We propose a synthetic data generation and RL methodology targeting multi-step optimization scenarios. This approach, called Step-Wise Reinforcement Learning (SWiRL), iteratively generates multi-step reasoning and tool use data, and then learns from that data. It employs a simple step-wise decomposition that breaks each multi-step trajectory into multiple sub-trajectories corresponding to each action by the original model. It then applies synthetic data filtering and RL optimization on these sub-trajectories. We evaluated SWiRL on a number of multi-step tool use, question answering, and mathematical reasoning tasks. Our experiments show that SWiRL outperforms baseline approaches by 21.5%, 12.3%, 14.8%, 11.1%, and 15.3% in relative accuracy on GSM8K, HotPotQA, CofCA, MuSiQue, and BeerQA, respectively. Excitingly, the approach exhibits generalization across tasks: for example, training only on HotPotQA (text question-answering) improves zero-shot performance on GSM8K (a math dataset) by a relative 16.9%.
Authors: Osvaldo Simeone, Sangwoo Park, Matteo Zecchin
Abstract: AI is poised to revolutionize telecommunication networks by boosting efficiency, automation, and decision-making. However, the black-box nature of most AI models introduces substantial risk, possibly deterring adoption by network operators. These risks are not addressed by the current prevailing deployment strategy, which typically follows a best-effort train-and-deploy paradigm. This paper reviews conformal calibration, a general framework that moves beyond the state of the art by adopting computationally lightweight, advanced statistical tools that offer formal reliability guarantees without requiring further training or fine-tuning. Conformal calibration encompasses pre-deployment calibration via uncertainty quantification or hyperparameter selection; online monitoring to detect and mitigate failures in real time; and counterfactual post-deployment performance analysis to address "what if" diagnostic questions after deployment. By weaving conformal calibration into the AI model lifecycle, network operators can establish confidence in black-box AI models as a dependable enabling technology for wireless systems.
Authors: Hsin-Yi Lin, Huan-Hsin Tseng, Samuel Yen-Chi Chen, Shinjae Yoo
Abstract: Conventional Variational Quantum Circuits (VQCs) for Quantum Machine Learning typically rely on a fixed Hermitian observable, often built from Pauli operators. Inspired by the Heisenberg picture, we propose an adaptive non-local measurement framework that substantially increases the model complexity of the quantum circuits. Our introduction of dynamical Hermitian observables with evolving parameters shows that optimizing VQC rotations corresponds to tracing a trajectory in the observable space. This viewpoint reveals that standard VQCs are merely a special case of the Heisenberg representation. Furthermore, we show that properly incorporating variational rotations with non-local observables enhances qubit interaction and information mixture, admitting flexible circuit designs. Two non-local measurement schemes are introduced, and numerical simulations on classification tasks confirm that our approach outperforms conventional VQCs, yielding a more powerful and resource-efficient approach as a Quantum Neural Network.
Authors: Siyuan Liang, Jiayang Liu, Jiecheng Zhai, Tianmeng Fang, Rongcheng Tu, Aishan Liu, Xiaochun Cao, Dacheng Tao
Abstract: The rapid development of generative artificial intelligence has made text to video models essential for building future multimodal world simulators. However, these models remain vulnerable to jailbreak attacks, where specially crafted prompts bypass safety mechanisms and lead to the generation of harmful or unsafe content. Such vulnerabilities undermine the reliability and security of simulation based applications. In this paper, we propose T2VShield, a comprehensive and model agnostic defense framework designed to protect text to video models from jailbreak threats. Our method systematically analyzes the input, model, and output stages to identify the limitations of existing defenses, including semantic ambiguities in prompts, difficulties in detecting malicious content in dynamic video outputs, and inflexible model centric mitigation strategies. T2VShield introduces a prompt rewriting mechanism based on reasoning and multimodal retrieval to sanitize malicious inputs, along with a multi scope detection module that captures local and global inconsistencies across time and modalities. The framework does not require access to internal model parameters and works with both open and closed source systems. Extensive experiments on five platforms show that T2VShield can reduce jailbreak success rates by up to 35 percent compared to strong baselines. We further develop a human centered audiovisual evaluation protocol to assess perceptual safety, emphasizing the importance of visual level defense in enhancing the trustworthiness of next generation multimodal simulators.
Authors: Seyed Shayan Daneshvar, Da Tan, Shaowei Wang, Carson Leung
Abstract: Various deep learning (DL) methods have recently been utilized to detect software vulnerabilities. Real-world software vulnerability datasets are rare and hard to acquire, as there is no simple metric for classifying vulnerability. Such datasets are heavily imbalanced, and none of the current datasets are considered huge for DL models. To tackle these problems, a recent work has tried to augment the dataset using the source code and generate realistic single-statement vulnerabilities, which is not quite practical and requires manual checking of the generated vulnerabilities. In this paper, we aim to explore the augmentation of vulnerabilities at the representation level to help current models learn better, which has never been done before to the best of our knowledge. We implement and evaluate five augmentation techniques that augment the embedding of the data and have recently been used for code search, which is a completely different software engineering task. We also introduced a conditioned version of those augmentation methods, which ensures the augmentation does not change the vulnerable section of the vector representation. We show that such augmentation methods can be helpful and increase the F1-score by up to 9.67%, yet they cannot beat Random Oversampling when balancing datasets, which increases the F1-score by 10.82%.
Authors: Md Abdul Baset Sarker, Art Nguyen, Sigmond Kukla, Kevin Fite, Masudul H. Imtiaz
Abstract: This paper introduces a novel AI vision-enabled pediatric prosthetic hand designed to assist children aged 10-12 with upper limb disabilities. The prosthesis features an anthropomorphic appearance, multi-articulating functionality, and a lightweight design that mimics a natural hand, making it both accessible and affordable for low-income families. Using 3D printing technology and integrating advanced machine vision, sensing, and embedded computing, the prosthetic hand offers a low-cost, customizable solution that addresses the limitations of current myoelectric prostheses. A micro camera is interfaced with a low-power FPGA for real-time object detection and assists with precise grasping. The onboard DL-based object detection and grasp classification models achieved accuracies of 96% and 100% respectively. In the force prediction, the mean absolute error was found to be 0.018. The features of the proposed prosthetic hand can thus be summarized as: a) a wrist-mounted micro camera for artificial sensing, enabling a wide range of hand-based tasks; b) real-time object detection and distance estimation for precise grasping; and c) ultra-low-power operation that delivers high performance within constrained power and resource limits.
Authors: Chen Xu, Qiang Wang, Lijun Sun
Abstract: Variational Autoencoders (VAEs) typically rely on a probabilistic decoder with a predefined likelihood, most commonly an isotropic Gaussian, to model the data conditional on latent variables. While convenient for optimization, this choice often leads to likelihood misspecification, resulting in blurry reconstructions and poor data fidelity, especially for high-dimensional data such as images. In this work, we propose EnVAE, a novel likelihood-free generative framework that has a deterministic decoder and employs the energy score--a proper scoring rule--to build the reconstruction loss. This enables likelihood-free inference without requiring explicit parametric density functions. To address the computational inefficiency of the energy score, we introduce a fast variant, FEnVAE, based on the local smoothness of the decoder and the sharpness of the posterior distribution of latent variables. This yields an efficient single-sample training objective that integrates seamlessly into existing VAE pipelines with minimal overhead. Empirical results on standard benchmarks demonstrate that EnVAE achieves superior reconstruction and generation quality compared to likelihood-based baselines. Our framework offers a general, scalable, and statistically principled alternative for flexible and nonparametric distribution learning in generative modeling.
Authors: Haochen Wang, Zhiwei Shi, Chengxi Zhu, Yafei Qiao, Cheng Zhang, Fan Yang, Pengjie Ren, Lan Lu, Dong Xuan
Abstract: Learning-based methods, such as imitation learning (IL) and reinforcement learning (RL), can produce excel control policies over challenging agile robot tasks, such as sports robot. However, no existing work has harmonized learning-based policy with model-based methods to reduce training complexity and ensure the safety and stability for agile badminton robot control. In this paper, we introduce Hamlet, a novel hybrid control system for agile badminton robots. Specifically, we propose a model-based strategy for chassis locomotion which provides a base for arm policy. We introduce a physics-informed "IL+RL" training framework for learning-based arm policy. In this train framework, a model-based strategy with privileged information is used to guide arm policy training during both IL and RL phases. In addition, we train the critic model during IL phase to alleviate the performance drop issue when transitioning from IL to RL. We present results on our self-engineered badminton robot, achieving 94.5% success rate against the serving machine and 90.7% success rate against human players. Our system can be easily generalized to other agile mobile manipulation tasks such as agile catching and table tennis. Our project website: https://dreamstarring.github.io/HAMLET/.
Authors: Bingye Zhou, Caiyang Yu
Abstract: Neural Architecture Search (NAS) has gained widespread attention for its transformative potential in deep learning model design. However, the vast and complex search space of NAS leads to significant computational and time costs. Neural Architecture Generation (NAG) addresses this by reframing NAS as a generation problem, enabling the precise generation of optimal architectures for specific tasks. Despite its promise, mainstream methods like diffusion models face limitations in global search capabilities and are still hindered by high computational and time demands. To overcome these challenges, we propose Evolutionary Diffusion-based Neural Architecture Generation (EDNAG), a novel approach that achieves efficient and training-free architecture generation. EDNAG leverages evolutionary algorithms to simulate the denoising process in diffusion models, using fitness to guide the transition from random Gaussian distributions to optimal architecture distributions. This approach combines the strengths of evolutionary strategies and diffusion models, enabling rapid and effective architecture generation. Extensive experiments demonstrate that EDNAG achieves state-of-the-art (SOTA) performance in architecture optimization, with an improvement in accuracy of up to 10.45%. Furthermore, it eliminates the need for time-consuming training and boosts inference speed by an average of 50 times, showcasing its exceptional efficiency and effectiveness.