Authors: Bernhard Jaeger, Daniel Dauner, Jens Bei{\ss}wenger, Simon Gerstenecker, Kashyap Chitta, Andreas Geiger
Abstract: We investigate reinforcement learning (RL) for privileged planning in autonomous driving. State-of-the-art approaches for this task are rule-based, but these methods do not scale to the long tail. RL, on the other hand, is scalable and does not suffer from compounding errors like imitation learning. Contemporary RL approaches for driving use complex shaped rewards that sum multiple individual rewards, \eg~progress, position, or orientation rewards. We show that PPO fails to optimize a popular version of these rewards when the mini-batch size is increased, which limits the scalability of these approaches. Instead, we propose a new reward design based primarily on optimizing a single intuitive reward term: route completion. Infractions are penalized by terminating the episode or multiplicatively reducing route completion. We find that PPO scales well with higher mini-batch sizes when trained with our simple reward, even improving performance. Training with large mini-batch sizes enables efficient scaling via distributed data parallelism. We scale PPO to 300M samples in CARLA and 500M samples in nuPlan with a single 8-GPU node. The resulting model achieves 64 DS on the CARLA longest6 v2 benchmark, outperforming other RL methods with more complex rewards by a large margin. Requiring only minimal adaptations from its use in CARLA, the same method is the best learning-based approach on nuPlan. It scores 91.3 in non-reactive and 90.6 in reactive traffic on the Val14 benchmark while being an order of magnitude faster than prior work.
Authors: A. J Miller, Fangzhou Yu, Michael Brauckmann, Farbod Farshidian
Abstract: This work presents an overview of the technical details behind a high performance reinforcement learning policy deployment with the Spot RL Researcher Development Kit for low level motor access on Boston Dynamics Spot. This represents the first public demonstration of an end to end end reinforcement learning policy deployed on Spot hardware with training code publicly available through Nvidia IsaacLab and deployment code available through Boston Dynamics. We utilize Wasserstein Distance and Maximum Mean Discrepancy to quantify the distributional dissimilarity of data collected on hardware and in simulation to measure our sim2real gap. We use these measures as a scoring function for the Covariance Matrix Adaptation Evolution Strategy to optimize simulated parameters that are unknown or difficult to measure from Spot. Our procedure for modeling and training produces high quality reinforcement learning policies capable of multiple gaits, including a flight phase. We deploy policies capable of over 5.2ms locomotion, more than triple Spots default controller maximum speed, robustness to slippery surfaces, disturbance rejection, and overall agility previously unseen on Spot. We detail our method and release our code to support future work on Spot with the low level API.
Authors: Karmanbir Batth, Krish Sethi, Aly Shariff, Leo Shi, Hetul Patel
Abstract: In this paper, we explore the Transformer based architectures for reinforcement learning in both online and offline settings within the Doom game environment. Our investigation focuses on two primary approaches: Deep Transformer Q- learning Networks (DTQN) for online learning and Decision Transformers (DT) for offline reinforcement learning. DTQN leverages the sequential modelling capabilities of Transformers to enhance Q-learning in partially observable environments,while Decision Transformers repurpose sequence modelling techniques to enable offline agents to learn from past trajectories without direct interaction with the environment. We conclude that while Transformers might have performed well in Atari games, more traditional methods perform better than Transformer based method in both the settings in the VizDoom environment.
Authors: Luiz Antonio Nicolau Anghinoni, Gustavo Weber Denardin, Jadson Castro Gertrudes, Dalcimar Casanova, Jefferson Tales Oliva
Abstract: Epilepsy, affecting approximately 50 million people globally, is characterized by abnormal brain activity and remains challenging to treat. The diagnosis of epilepsy relies heavily on electroencephalogram (EEG) data, where specialists manually analyze epileptiform patterns across pre-ictal, ictal, post-ictal, and interictal periods. However, the manual analysis of EEG signals is prone to variability between experts, emphasizing the need for automated solutions. Although previous studies have explored preprocessing techniques and machine learning approaches for seizure detection, there is a gap in understanding how the representation of EEG data (time, frequency, or time-frequency domains) impacts the predictive performance of deep learning models. This work addresses this gap by systematically comparing deep neural networks trained on EEG data in these three domains. Through the use of statistical tests, we identify the optimal data representation and model architecture for epileptic seizure detection. The results demonstrate that frequency-domain data achieves detection metrics exceeding 97\%, providing a robust foundation for more accurate and reliable seizure detection systems.
Authors: Mert Sonmezer, Seyda Ertekin
Abstract: Long-term time series forecasting plays a pivotal role in various real-world applications. Despite recent advancements and the success of different architectures, forecasting is often challenging due to non-stationary nature of the real-world data, which frequently exhibit distribution shifts and temporal changes in statistical properties like mean and variance over time. Previous studies suggest that this inherent variability complicates forecasting, limiting the performance of many models by leading to loss of non-stationarity and resulting in over-stationarization (Liu, Wu, Wang and Long, 2022). To address this challenge, we introduce a novel architecture, ChoronoAdaptive Network (CANet), inspired by style-transfer techniques. The core of CANet is the Non-stationary Adaptive Normalization module, seamlessly integrating the Style Blending Gate and Adaptive Instance Normalization (AdaIN) (Huang and Belongie, 2017). The Style Blending Gate preserves and reintegrates non-stationary characteristics, such as mean and standard deviation, by blending internal and external statistics, preventing over-stationarization while maintaining essential temporal dependencies. Coupled with AdaIN, which dynamically adapts the model to statistical changes, this approach enhances predictive accuracy under non-stationary conditions. CANet also employs multi-resolution patching to handle short-term fluctuations and long-term trends, along with Fourier analysis-based adaptive thresholding to reduce noise. A Stacked Kronecker Product Layer further optimizes the model's efficiency while maintaining high performance. Extensive experiments on real-world datasets validate CANet's superiority over state-of-the-art methods, achieving a 42% reduction in MSE and a 22% reduction in MAE. The source code is publicly available at https://github.com/mertsonmezer/CANet.
Authors: Mateo Espinosa Zarlenga, Gabriele Dominici, Pietro Barbiero, Zohreh Shams, Mateja Jamnik
Abstract: In this paper, we investigate how concept-based models (CMs) respond to out-of-distribution (OOD) inputs. CMs are interpretable neural architectures that first predict a set of high-level concepts (e.g., stripes, black) and then predict a task label from those concepts. In particular, we study the impact of concept interventions (i.e., operations where a human expert corrects a CM's mispredicted concepts at test time) on CMs' task predictions when inputs are OOD. Our analysis reveals a weakness in current state-of-the-art CMs, which we term leakage poisoning, that prevents them from properly improving their accuracy when intervened on for OOD inputs. To address this, we introduce MixCEM, a new CM that learns to dynamically exploit leaked information missing from its concepts only when this information is in-distribution. Our results across tasks with and without complete sets of concept annotations demonstrate that MixCEMs outperform strong baselines by significantly improving their accuracy for both in-distribution and OOD samples in the presence and absence of concept interventions.
Authors: Fatemeh Vares, Brittany Johnson
Abstract: Deep Neural Networks (DNNs) often rely on statistical correlations rather than causal reasoning, limiting their robustness and interpretability. While testing methods can identify failures, effective debugging and repair remain challenging. This paper explores causal inference as an approach primarily for DNN repair, leveraging causal debugging, counterfactual analysis, and structural causal models (SCMs) to identify and correct failures. We discuss in what ways these techniques support fairness, adversarial robustness, and backdoor mitigation by providing targeted interventions. Finally, we discuss key challenges, including scalability, generalization, and computational efficiency, and outline future directions for integrating causality-driven interventions to enhance DNN reliability.
Authors: Liangzu Peng, Ren\'e Vidal
Abstract: Continual learning is an emerging subject in machine learning that aims to solve multiple tasks presented sequentially to the learner without forgetting previously learned tasks. Recently, many deep learning based approaches have been proposed for continual learning, however the mathematical foundations behind existing continual learning methods remain underdeveloped. On the other hand, adaptive filtering is a classic subject in signal processing with a rich history of mathematically principled methods. However, its role in understanding the foundations of continual learning has been underappreciated. In this tutorial, we review the basic principles behind both continual learning and adaptive filtering, and present a comparative analysis that highlights multiple connections between them. These connections allow us to enhance the mathematical foundations of continual learning based on existing results for adaptive filtering, extend adaptive filtering insights using existing continual learning methods, and discuss a few research directions for continual learning suggested by the historical developments in adaptive filtering.
Authors: Aditya S Ellendula, Chandrajit Bajaj
Abstract: We present a dynamic self-balancing octree data structure that enables efficient neighborhood maintenance in evolving metric spaces, a key challenge in modern machine learning systems. Many learning and generative models operate as dynamical systems whose representations evolve during training, requiring fast, adaptive spatial organization. Our two-parameter octree supports logarithmic-time updates and queries, eliminating the need for costly full rebuilds as data distributions shift. We demonstrate its effectiveness in four areas: (1) accelerating Stein variational gradient descent by supporting more particles with lower overhead; (2) enabling real-time, incremental KNN classification with logarithmic complexity; (3) facilitating efficient, dynamic indexing and retrieval for retrieval-augmented generation; and (4) improving sample efficiency by jointly optimizing input and latent spaces. Across all applications, our approach yields exponential speedups while preserving accuracy, particularly in high-dimensional spaces where maintaining adaptive spatial structure is critical.
Authors: Nooshin Yousefzadeh, Rahul Sengupta, Sanjay Ranka
Abstract: Urban congestion at signalized intersections leads to significant delays, economic losses, and increased emissions. Existing deep learning models often lack spatial generalizability, rely on complex architectures, and struggle with real-time deployment. To address these limitations, we propose the Temporal Graph-based Digital Twin (TGDT), a scalable framework that integrates Temporal Convolutional Networks and Attentional Graph Neural Networks for dynamic, direction-aware traffic modeling and assessment at urban corridors. TGDT estimates key Measures of Effectiveness (MOEs) for traffic flow optimization at both the intersection level (e.g., queue length, waiting time) and the corridor level (e.g., traffic volume, travel time). Its modular architecture and sequential optimization scheme enable easy extension to any number of intersections and MOEs. The model outperforms state-of-the-art baselines by accurately producing high-dimensional, concurrent multi-output estimates. It also demonstrates high robustness and accuracy across diverse traffic conditions, including extreme scenarios, while relying on only a minimal set of traffic features. Fully parallelized, TGDT can simulate over a thousand scenarios within a matter of seconds, offering a cost-effective, interpretable, and real-time solution for traffic signal optimization.
Authors: Emiliano Penaloza, Tianyue H. Zhan, Laurent Charlin, Mateo Espinosa Zarlenga
Abstract: Concept Bottleneck Models (CBMs) propose to enhance the trustworthiness of AI systems by constraining their decisions on a set of human understandable concepts. However, CBMs typically assume that datasets contains accurate concept labels an assumption often violated in practice, which we show can significantly degrade performance (by 25% in some cases). To address this, we introduce the Concept Preference Optimization (CPO) objective, a new loss function based on Direct Preference Optimization, which effectively mitigates the negative impact of concept mislabeling on CBM performance. We provide an analysis on some key properties of the CPO objective showing it directly optimizes for the concept's posterior distribution, and contrast it against Binary Cross Entropy (BCE) where we show CPO is inherently less sensitive to concept noise. We empirically confirm our analysis finding that CPO consistently outperforms BCE in three real world datasets with and without added label noise.
Authors: Zhongtian Chen, Daniel Murfet
Abstract: We develop a geometric account of sequence modelling that links patterns in the data to measurable properties of the loss landscape in transformer networks. First, we cast conditional sequence distributions into a Hilbert-space framework and apply tensor decompositions to identify their principal modes. Truncating the small-amplitude modes yields an effective data distribution that preserves dominant structure while discarding statistical detail. Second, we show theoretically that Local Learning Coefficient (LLC) estimates are insensitive to modes below a data-dependent threshold. Consequently, the LLC calculated in practice characterises the geometry of the effective rather than the true distribution. This insight clarifies why reliable LLC estimates can be obtained even when a network parameter is not a strict minimiser of the population loss, and it highlights how the inverse temperature in SGLD acts as a resolution dial on the landscape structure.
Authors: Konstantin Sch\"urholt, L\'eo Meynent, Yefan Zhou, Haiquan Lu, Yaoqing Yang, Damian Borth
Abstract: Using the weights of trained Neural Network (NN) models as data modality has recently gained traction as a research field - dubbed Weight Space Learning (WSL). Multiple recent works propose WSL methods to analyze models, evaluate methods, or synthesize weights. Weight space learning methods require populations of trained models as datasets for development and evaluation. However, existing collections of models - called `model zoos' - are unstructured or follow a rudimentary definition of diversity. In parallel, work rooted in statistical physics has identified phases and phase transitions in NN models. Models are homogeneous within the same phase but qualitatively differ from one phase to another. We combine the idea of `model zoos' with phase information to create a controlled notion of diversity in populations. We introduce 12 large-scale zoos that systematically cover known phases and vary over model architecture, size, and datasets. These datasets cover different modalities, such as computer vision, natural language processing, and scientific ML. For every model, we compute loss landscape metrics and validate full coverage of the phases. With this dataset, we provide the community with a resource with a wide range of potential applications for WSL and beyond. Evidence suggests the loss landscape phase plays a role in applications such as model training, analysis, or sparsification. We demonstrate this in an exploratory study of the downstream methods like transfer learning or model weights averaging.
Authors: Xiaolu Chen, Chenghao Huang, Yanru Zhang, Hao Wang
Abstract: The rapid expansion of distributed photovoltaic (PV) installations worldwide, many being behind-the-meter systems, has significantly challenged energy management and grid operations, as unobservable PV generation further complicates the supply-demand balance. Therefore, estimating this generation from net load, known as PV disaggregation, is critical. Given privacy concerns and the need for large training datasets, federated learning becomes a promising approach, but statistical heterogeneity, arising from geographical and behavioral variations among prosumers, poses new challenges to PV disaggregation. To overcome these challenges, a privacy-preserving distributed PV disaggregation framework is proposed using Personalized Federated Learning (PFL). The proposed method employs a two-level framework that combines local and global modeling. At the local level, a transformer-based PV disaggregation model is designed to generate solar irradiance embeddings for representing local PV conditions. A novel adaptive local aggregation mechanism is adopted to mitigate the impact of statistical heterogeneity on the local model, extracting a portion of global information that benefits the local model. At the global level, a central server aggregates information uploaded from multiple data centers, preserving privacy while enabling cross-center knowledge sharing. Experiments on real-world data demonstrate the effectiveness of this proposed framework, showing improved accuracy and robustness compared to benchmark methods.
Authors: Vignesh Balaji, Christos Kozyrakis, Gal Chechik, Haggai Maron
Abstract: Graph Neural Networks (GNNs) enable learning on realworld graphs and mini-batch training has emerged as the de facto standard for training GNNs because it can scale to very large graphs and improve convergence. Current mini-batch construction policies largely ignore efficiency considerations of GNN training. Specifically, existing mini-batching techniques employ randomization schemes to improve accuracy and convergence. However, these randomization schemes are often agnostic to the structural properties of the graph (for eg. community structure), resulting in highly irregular memory access patterns during GNN training that make suboptimal use of on-chip GPU caches. On the other hand, while deterministic mini-batching based solely on graph structure delivers fast runtime performance, the lack of randomness compromises both the final model accuracy and training convergence speed. In this paper, we present Community-structure-aware Randomized Mini-batching (COMM-RAND), a novel methodology that bridges the gap between the above extremes. COMM-RAND allows practitioners to explore the space between pure randomness and pure graph structural awareness during mini-batch construction, leading to significantly more efficient GNN training with similar accuracy. We evaluated COMM-RAND across four popular graph learning benchmarks. COMM-RAND cuts down GNN training time by up to 2.76x (1.8x on average) while achieving an accuracy that is within 1.79% points (0.42% on average) compared to popular random mini-batching approaches.
Authors: Shota Deguchi, Mitsuteru Asai
Abstract: Physics-informed neural networks have attracted significant attention in scientific machine learning for their capability to solve forward and inverse problems governed by partial differential equations. However, the accuracy of PINN solutions is often limited by the treatment of boundary conditions. Conventional penalty-based methods, which incorporate boundary conditions as penalty terms in the loss function, cannot guarantee exact satisfaction of the given boundary conditions and are highly sensitive to the choice of penalty parameters. This paper demonstrates that distance functions, specifically R-functions, can be leveraged to enforce boundary conditions, overcoming these limitations. R-functions provide normalized distance fields, enabling accurate representation of boundary geometries, including non-convex domains, and facilitating various types of boundary conditions. We extend this distance function-based boundary condition imposition method to inverse problems using PINNs and introduce an adaptive weight tuning technique to ensure reliable and efficient inverse analysis. We demonstrate the efficacy of the method through several numerical experiments. Numerical results show that the proposed method solves inverse problems more accurately and efficiently than penalty-based methods, even in the presence of complex non-convex geometries. This approach offers a reliable and efficient framework for inverse analysis using PINNs, with potential applications across a wide range of engineering problems.
Authors: Jerrin Thomas Panachakel, Pradeep Kumar G., Suryaa Seran, Kanishka Sharma, Ramakrishnan Angarai Ganesan
Abstract: While it is beneficial to objectively determine whether a subject is meditating, most research in the literature reports good results only in a subject-dependent manner. This study aims to distinguish the modified state of consciousness experienced during Rajyoga meditation from the resting state of the brain in a subject-independent manner using EEG data. Three architectures have been proposed and evaluated: The CSP-LDA Architecture utilizes common spatial pattern (CSP) for feature extraction and linear discriminant analysis (LDA) for classification. The CSP-LDA-LSTM Architecture employs CSP for feature extraction, LDA for dimensionality reduction, and long short-term memory (LSTM) networks for classification, modeling the binary classification problem as a sequence learning problem. The SVD-NN Architecture uses singular value decomposition (SVD) to select the most relevant components of the EEG signals and a shallow neural network (NN) for classification. The CSP-LDA-LSTM architecture gives the best performance with 98.2% accuracy for intra-subject classification. The SVD-NN architecture provides significant performance with 96.4\% accuracy for inter-subject classification. This is comparable to the best-reported accuracies in the literature for intra-subject classification. Both architectures are capable of capturing subject-invariant EEG features for effectively classifying the meditative state from the resting state. The high intra-subject and inter-subject classification accuracies indicate these systems' robustness and their ability to generalize across different subjects.
Authors: Dinan Li, Panagiotis Kakosimos
Abstract: The number of electrified powertrains is ever increasing today towards a more sustainable future; thus, it is essential that unwanted failures are prevented, and a reliable operation is secured. Monitoring the internal temperatures of motors and keeping them under their thresholds is an important first step. Conventional modeling methods require expert knowledge and complicated mathematical approaches. With all the data a modern electric drive collects nowadays during the system operation, it is feasible to apply data-driven approaches for estimating thermal behaviors. In this paper, multiple machine-learning methods are investigated on their capability to approximate the temperatures of the stator winding and bearing in induction motors. The explored algorithms vary from linear to neural networks. For this reason, experimental lab data have been captured from a powertrain under predetermined operating conditions. For each approach, a hyperparameter search is then performed to find the optimal configuration. All the models are evaluated by various metrics, and it has been found that neural networks perform satisfactorily even under transient conditions.
Authors: Aniket Dixit, Muhammad Ibrahim Khan, Faizan Ahmed, James Brusey
Abstract: This paper introduces an approach for developing surrogate environments in reinforcement learning (RL) using the Sparse Identification of Nonlinear Dynamics (SINDy) algorithm. We demonstrate the effectiveness of our approach through extensive experiments in OpenAI Gym environments, particularly Mountain Car and Lunar Lander. Our results show that SINDy-based surrogate models can accurately capture the underlying dynamics of these environments while reducing computational costs by 20-35%. With only 75 interactions for Mountain Car and 1000 for Lunar Lander, we achieve state-wise correlations exceeding 0.997, with mean squared errors as low as 3.11e-06 for Mountain Car velocity and 1.42e-06 for LunarLander position. RL agents trained in these surrogate environments require fewer total steps (65,075 vs. 100,000 for Mountain Car and 801,000 vs. 1,000,000 for Lunar Lander) while achieving comparable performance to those trained in the original environments, exhibiting similar convergence patterns and final performance metrics. This work contributes to the field of model-based RL by providing an efficient method for generating accurate, interpretable surrogate environments.
Authors: Caia Costello, Simon Guo, Anna Goldie, Azalia Mirhoseini
Abstract: Large language models (LLMs) have demonstrated strong capabilities in programming and mathematical reasoning tasks, but are constrained by limited high-quality training data. Synthetic data can be leveraged to enhance fine-tuning outcomes, but several factors influence this process, including model size, synthetic data volume, pruning strategy, and number of fine-tuning rounds. We explore these axes and investigate which conditions enable model self-improvement. We introduce the Think, Prune, Train process, a scalable framework that iteratively fine-tunes models on their own reasoning traces, using ground-truth pruning to ensure high-quality training data. This approach yields improved performance: on GSM8K, Gemma2-2B achieves a Pass@1 of 57.6% (from 41.9%), Gemma2-9B reaches 82%, matching LLaMA-3.1-70B, and LLaMA-3.1-70B attains 91%, even surpassing GPT-4o, demonstrating the effectiveness of self-generated reasoning and systematic data selection for improving LLM capabilities.
Authors: Vasily Ilin, Bamdad Hosseini, Jingwei Hu
Abstract: We propose and analyze a deterministic sampling framework using Score-Based Transport Modeling (SBTM) for sampling an unnormalized target density $\pi$. While diffusion generative modeling relies on pre-training the score function $\nabla \log f_t$ using samples from $\pi$, SBTM addresses the more general and challenging setting where only $\nabla \log\pi$ is known. SBTM approximates the Wasserstein gradient flow on KL$(f_t\|\pi)$ by learning the time-varying score $\nabla \log f_t$ on the fly using score matching. The learned score gives immediate access to relative Fisher information, providing a built-in convergence diagnostic. The deterministic trajectories are smooth, interpretable, and free of Brownian-motion noise, while having the same distribution as ULA. We prove that SBTM dissipates relative entropy at the same rate as the exact gradient flow, provided sufficient training. We further extend our framework to annealed dynamics, to handle non log-concave targets. Numerical experiments validate our theoretical findings: SBTM converges at the optimal rate, has smooth trajectories, and is easily integrated with annealed dynamics. We compare to the baselines of ULA and annealed ULA.
Authors: Gissel Velarde, Michael Weichert, Anuj Deshmunkh, Sanjay Deshmane, Anindya Sudhir, Khushboo Sharma, Vaibhav Joshi
Abstract: Most real-world classification problems deal with imbalanced datasets, posing a challenge for Artificial Intelligence (AI), i.e., machine learning algorithms, because the minority class, which is of extreme interest, often proves difficult to be detected. This paper empirically evaluates tree boosting methods' performance given different dataset sizes and class distributions, from perfectly balanced to highly imbalanced. For tabular data, tree-based methods such as XGBoost, stand out in several benchmarks due to detection performance and speed. Therefore, XGBoost and Imbalance-XGBoost are evaluated. After introducing the motivation to address risk assessment with machine learning, the paper reviews evaluation metrics for detection systems or binary classifiers. It proposes a method for data preparation followed by tree boosting methods including hyper-parameter optimization. The method is evaluated on private datasets of 1 thousand (K), 10K and 100K samples on distributions with 50, 45, 25, and 5 percent positive samples. As expected, the developed method increases its recognition performance as more data is given for training and the F1 score decreases as the data distribution becomes more imbalanced, but it is still significantly superior to the baseline of precision-recall determined by the ratio of positives divided by positives and negatives. Sampling to balance the training set does not provide consistent improvement and deteriorates detection. In contrast, classifier hyper-parameter optimization improves recognition, but should be applied carefully depending on data volume and distribution. Finally, the developed method is robust to data variation over time up to some point. Retraining can be used when performance starts deteriorating.
Authors: Xiaofan Wei, Binyan Zhang
Abstract: Graph contrastive learning (GCL) has garnered significant attention recently since it learns complex structural information from graphs through self-supervised learning manner. However, prevalent GCL models may suffer from performance degradation due to inappropriate contrastive signals. Concretely, they commonly generate augmented views based on random perturbation, which leads to biased essential structures due to the introduction of noise. In addition, they assign equal weight to both hard and easy sample pairs, thereby ignoring the difference in importance of the sample pairs. To address these issues, this study proposes a novel Contrastive Signal Generative Framework for Accurate Graph Learning (CSG2L) with the following two-fold ideas: a) building a singular value decomposition (SVD)-directed augmented module (SVD-aug) to obtain the global interactions as well as avoiding the random noise perturbation; b) designing a local-global dependency learning module (LGDL) with an adaptive reweighting strategy which can differentiate the effects of hard and easy sample pairs. Extensive experiments on benchmark datasets demonstrate that the proposed CSG2L outperforms the state-of-art baselines. Moreover, CSG2L is compatible with a variety of GNNs.
Authors: Mathieu Petitbois, R\'emy Portelas, Sylvain Lamprier, Ludovic Denoyer
Abstract: Imitation Learning (IL) techniques aim to replicate human behaviors in specific tasks. While IL has gained prominence due to its effectiveness and efficiency, traditional methods often focus on datasets collected from experts to produce a single efficient policy. Recently, extensions have been proposed to handle datasets of diverse behaviors by mainly focusing on learning transition-level diverse policies or on performing entropy maximization at the trajectory level. While these methods may lead to diverse behaviors, they may not be sufficient to reproduce the actual diversity of demonstrations or to allow controlled trajectory generation. To overcome these drawbacks, we propose a different method based on two key features: a) Temporal Consistency that ensures consistent behaviors across entire episodes and not just at the transition level as well as b) Controllability obtained by constructing a latent space of behaviors that allows users to selectively activate specific behaviors based on their requirements. We compare our approach to state-of-the-art methods over a diverse set of tasks and environments. Project page: https://mathieu-petitbois.github.io/projects/swr/
Authors: Yvonne Jenniges, Maike Sonnewald, Sebastian Maneth, Are Olsen, Boris P. Koch
Abstract: Defining ocean regions and water masses helps to understand marine processes and can serve downstream-tasks such as defining marine protected areas. However, such definitions are often a result of subjective decisions potentially producing misleading, unreproducible results. Here, the aim was to objectively define regions of the North Atlantic. For this, a data-driven, systematic machine learning approach was applied to generate and validate ocean clusters employing external, internal and relative validation techniques. About 300 million measured salinity, temperature, and oxygen, nitrate, phosphate and silicate concentration values served as input for various clustering methods (KMeans, agglomerative Ward, and Density-Based Spatial Clustering of Applications with Noise (DBSCAN)). Uniform Manifold Approximation and Projection (UMAP) emphasised (dis-)similarities in the data while reducing dimensionality. Based on a systematic validation of the considered clustering methods and their hyperparameters, the results showed that UMAP-DBSCAN best represented the data. To address stochastic variability, 100 UMAP-DBSCAN clustering runs were conducted and aggregated using Native Emergent Manifold Interrogation (NEMI), producing a final set of 321 clusters. Reproducibility was evaluated by calculating the ensemble overlap (88.81 +- 1.8%) and the mean grid cell-wise uncertainty estimated by NEMI (15.49 +- 20%). The presented clustering results agreed very well with common water mass definitions. This study revealed a more detailed regionalization compared to previous concepts such as the Longhurst provinces. The applied method is objective, efficient and reproducible and will support future research focusing on biogeochemical differences and changes in oceanic regions.
Authors: Gissel Velarde, Pedro Branez, Alejandro Bueno, Rodrigo Heredia, Mateo Lopez-Ledezma
Abstract: This paper introduces an open-source and reproducible implementation of Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) Networks for time series forecasting. We evaluated LSTM and GRU networks because of their performance reported in related work. We describe our method and its results on two datasets. The first dataset is the S&P BSE BANKEX, composed of stock time series (closing prices) of ten financial institutions. The second dataset, called Activities, comprises ten synthetic time series resembling weekly activities with five days of high activity and two days of low activity. We report Root Mean Squared Error (RMSE) between actual and predicted values, as well as Directional Accuracy (DA). We show that a single time series from a dataset can be used to adequately train the networks if the sequences in the dataset contain patterns that repeat, even with certain variation, and are properly processed. For 1-step ahead and 20-step ahead forecasts, LSTM and GRU networks significantly outperform a baseline on the Activities dataset. The baseline simply repeats the last available value. On the stock market dataset, the networks perform just like the baseline, possibly due to the nature of these series. We release the datasets used as well as the implementation with all experiments performed to enable future comparisons and to make our research reproducible.
Authors: Stefano Sossi-Rojas, Gissel Velarde, Damian Zieba
Abstract: Bitcoin is one of the cryptocurrencies that is gaining more popularity in recent years. Previous studies have shown that closing price alone is not enough to forecast stock market series. We introduce a new set of time series and demonstrate that a subset is necessary to improve directional accuracy based on a machine learning ensemble. In our experiments, we study which time series and machine learning algorithms deliver the best results. We found that the most relevant time series that contribute to improving directional accuracy are Open, High and Low, with the largest contribution of Low in combination with an ensemble of Gated Recurrent Unit network and a baseline forecast. The relevance of other Bitcoin-related features that are not price-related is negligible. The proposed method delivers similar performance to the state-of-the-art when observing directional accuracy.
Authors: Simon Lucey
Abstract: We generalize the connection between activation function and spline regression/smoothing and characterize how this choice may influence spectral bias within a 1D shallow network. We then demonstrate how gradient descent (GD) can be reinterpreted as a shrinkage operator that masks the singular values of a neural network's Jacobian. Viewed this way, GD implicitly selects the number of frequency components to retain, thereby controlling the spectral bias. An explicit relationship is proposed between the choice of GD hyperparameters (learning rate & number of iterations) and bandwidth (the number of active components). GD regularization is shown to be effective only with monotonic activation functions. Finally, we highlight the utility of non-monotonic activation functions (sinc, Gaussian) as iteration-efficient surrogates for spectral bias.
Authors: Rapha\"el Barboni (\'ENS-PSL), Gabriel Peyr\'e (CNRS,\'ENS-PSL), Fran\c{c}ois-Xavier Vialard (LIGM)
Abstract: We study the convergence of gradient methods for the training of mean-field single hidden layer neural networks with square loss. Observing this is a separable non-linear least-square problem which is linear w.r.t. the outer layer's weights, we consider a Variable Projection (VarPro) or two-timescale learning algorithm, thereby eliminating the linear variables and reducing the learning problem to the training of the feature distribution. Whereas most convergence rates or the training of neural networks rely on a neural tangent kernel analysis where features are fixed, we show such a strategy enables provable convergence rates for the sampling of a teacher feature distribution. Precisely, in the limit where the regularization strength vanishes, we show that the dynamic of the feature distribution corresponds to a weighted ultra-fast diffusion equation. Relying on recent results on the asymptotic behavior of such PDEs, we obtain guarantees for the convergence of the trained feature distribution towards the teacher feature distribution in a teacher-student setup.
Authors: He Shanxuan, Lin Zuhong, Yu Bolun, Gao Xu, Long Biao, Yao Jingjing
Abstract: Accurate prediction of lithium-ion battery lifespan is vital for ensuring operational reliability and reducing maintenance costs in applications like electric vehicles and smart grids. This study presents a hybrid learning framework for precise battery lifespan prediction, integrating dynamic multi-source data fusion with a stacked ensemble (SE) modeling approach. By leveraging heterogeneous datasets from the National Aeronautics and Space Administration (NASA), Center for Advanced Life Cycle Engineering (CALCE), MIT-Stanford-Toyota Research Institute (TRC), and nickel cobalt aluminum (NCA) chemistries, an entropy-based dynamic weighting mechanism mitigates variability across heterogeneous datasets. The SE model combines Ridge regression, long short-term memory (LSTM) networks, and eXtreme Gradient Boosting (XGBoost), effectively capturing temporal dependencies and nonlinear degradation patterns. It achieves a mean absolute error (MAE) of 0.0058, root mean square error (RMSE) of 0.0092, and coefficient of determination (R2) of 0.9839, outperforming established baseline models with a 46.2% improvement in R2 and an 83.2% reduction in RMSE. Shapley additive explanations (SHAP) analysis identifies differential discharge capacity (Qdlin) and temperature of measurement (Temp_m) as critical aging indicators. This scalable, interpretable framework enhances battery health management, supporting optimized maintenance and safety across diverse energy storage systems, thereby contributing to improved battery health management in energy storage systems.
Authors: Rong Cheng, Jinyi Liu, YAN ZHENG, Fei Ni, Jiazhen Du, Hangyu Mao, Fuzheng Zhang, Bo Wang, Jianye HAO
Abstract: Multi-Hop Question Answering (MHQA) tasks permeate real-world applications, posing challenges in orchestrating multi-step reasoning across diverse knowledge domains. While existing approaches have been improved with iterative retrieval, they still struggle to identify and organize dynamic knowledge. To address this, we propose DualRAG, a synergistic dual-process framework that seamlessly integrates reasoning and retrieval. DualRAG operates through two tightly coupled processes: Reasoning-augmented Querying (RaQ) and progressive Knowledge Aggregation (pKA). They work in concert: as RaQ navigates the reasoning path and generates targeted queries, pKA ensures that newly acquired knowledge is systematically integrated to support coherent reasoning. This creates a virtuous cycle of knowledge enrichment and reasoning refinement. Through targeted fine-tuning, DualRAG preserves its sophisticated reasoning and retrieval capabilities even in smaller-scale models, demonstrating its versatility and core advantages across different scales. Extensive experiments demonstrate that this dual-process approach substantially improves answer accuracy and coherence, approaching, and in some cases surpassing, the performance achieved with oracle knowledge access. These results establish DualRAG as a robust and efficient solution for complex multi-hop reasoning tasks.
Authors: Andrea Quintanilla, Johan Van Horebeek
Abstract: Given the high computational complexity of decision tree estimation, classical methods construct a tree by adding one node at a time in a recursive way. To facilitate promoting fairness, we propose a fairness criterion local to the tree nodes. We prove how it is related to the Statistical Parity criterion, popular in the Algorithmic Fairness literature, and show how to incorporate it into standard recursive tree estimation algorithms. We present a tree estimation algorithm called Constrained Logistic Regression Tree (C-LRT), which is a modification of the standard CART algorithm using locally linear classifiers and imposing restrictions as done in Constrained Logistic Regression. Finally, we evaluate the performance of trees estimated with C-LRT on datasets commonly used in the Algorithmic Fairness literature, using various classification and fairness metrics. The results confirm that C-LRT successfully allows to control and balance accuracy and fairness.
Authors: Prajwal Chauhan, Salah Eddine Choutri, Mohamed Ghattassi, Nader Masmoudi, Saif Eddin Jabari
Abstract: This paper investigates the limitations of neural operators in learning solutions for a Hughes model, a first-order hyperbolic conservation law system for crowd dynamics. The model couples a Fokker-Planck equation representing pedestrian density with a Hamilton-Jacobi-type (eikonal) equation. This Hughes model belongs to the class of nonlinear hyperbolic systems that often exhibit complex solution structures, including shocks and discontinuities. In this study, we assess the performance of three state-of-the-art neural operators (Fourier Neural Operator, Wavelet Neural Operator, and Multiwavelet Neural Operator) in various challenging scenarios. Specifically, we consider (1) discontinuous and Gaussian initial conditions and (2) diverse boundary conditions, while also examining the impact of different numerical schemes. Our results show that these neural operators perform well in easy scenarios with fewer discontinuities in the initial condition, yet they struggle in complex scenarios with multiple initial discontinuities and dynamic boundary conditions, even when trained specifically on such complex samples. The predicted solutions often appear smoother, resulting in a reduction in total variation and a loss of important physical features. This smoothing behavior is similar to issues discussed by Daganzo (1995), where models that introduce artificial diffusion were shown to miss essential features such as shock waves in hyperbolic systems. These results suggest that current neural operator architectures may introduce unintended regularization effects that limit their ability to capture transport dynamics governed by discontinuities. They also raise concerns about generalizing these methods to traffic applications where shock preservation is essential.
Authors: Garrett Baker, George Wang, Jesse Hoogland, Daniel Murfet
Abstract: We develop a linear response framework for interpretability that treats a neural network as a Bayesian statistical mechanical system. A small, controlled perturbation of the data distribution, for example shifting the Pile toward GitHub or legal text, induces a first-order change in the posterior expectation of an observable localized on a chosen component of the network. The resulting susceptibility can be estimated efficiently with local SGLD samples and factorizes into signed, per-token contributions that serve as attribution scores. Building a set of perturbations (probes) yields a response matrix whose low-rank structure separates functional modules such as multigram and induction heads in a 3M-parameter transformer. Susceptibilities link local learning coefficients from singular learning theory with linear-response theory, and quantify how local loss landscape geometry deforms under shifts in the data distribution.
Authors: Richard Oliver Lane
Abstract: Probabilities or confidence values produced by artificial intelligence (AI) and machine learning (ML) models often do not reflect their true accuracy, with some models being under or over confident in their predictions. For example, if a model is 80% sure of an outcome, is it correct 80% of the time? Probability calibration metrics measure the discrepancy between confidence and accuracy, providing an independent assessment of model calibration performance that complements traditional accuracy metrics. Understanding calibration is important when the outputs of multiple systems are combined, for assurance in safety or business-critical contexts, and for building user trust in models. This paper provides a comprehensive review of probability calibration metrics for classifier and object detection models, organising them according to a number of different categorisations to highlight their relationships. We identify 82 major metrics, which can be grouped into four classifier families (point-based, bin-based, kernel or curve-based, and cumulative) and an object detection family. For each metric, we provide equations where available, facilitating implementation and comparison by future researchers.
Authors: Simon Hakenes, Tobias Glasmachers
Abstract: This paper addresses the challenge of navigation in large, visually complex environments with sparse rewards. We propose a method that uses object-oriented macro actions grounded in a topological map, allowing a simple Deep Q-Network (DQN) to learn effective navigation policies. The agent builds a map by detecting objects from RGBD input and selecting discrete macro actions that correspond to navigating to these objects. This abstraction drastically reduces the complexity of the underlying reinforcement learning problem and enables generalization to unseen environments. We evaluate our approach in a photorealistic 3D simulation and show that it significantly outperforms a random baseline under both immediate and terminal reward conditions. Our results demonstrate that topological structure and macro-level abstraction can enable sample-efficient learning even from pixel data.
Authors: Marco Turzi, Siamak Mehrkanoon
Abstract: Weather forecasting is essential for facilitating diverse socio-economic activity and environmental conservation initiatives. Deep learning techniques are increasingly being explored as complementary approaches to Numerical Weather Prediction (NWP) models, offering potential benefits such as reduced complexity and enhanced adaptability in specific applications. This work presents a novel design, Small Shuffled Attention UNet (SSA-UNet), which enhances SmaAt-UNet's architecture by including a shuffle channeling mechanism to optimize performance and diminish complexity. To assess its efficacy, this architecture and its reduced variant are examined and trained on two datasets: a Dutch precipitation dataset from 2016 to 2019, and a French cloud cover dataset containing radar images from 2017 to 2018. Three output configurations of the proposed architecture are evaluated, yielding outputs of 1, 6, and 12 precipitation maps, respectively. To better understand how this model operates and produces its predictions, a gradient-based approach called Grad-CAM is used to analyze the outputs generated. The analysis of heatmaps generated by Grad-CAM facilitated the identification of regions within the input maps that the model considers most informative for generating its predictions. The implementation of SSA-UNet can be found on our Github\footnote{\href{https://github.com/MarcoTurzi/SSA-UNet}{https://github.com/MarcoTurzi/SSA-UNet}}
URLs: https://github.com/MarcoTurzi/SSA-UNet, https://github.com/MarcoTurzi/SSA-UNet
Authors: Anh-Duy Pham, Olivier Basole Kashongwe, Martin Atzmueller, Tim R\"omer
Abstract: Balancing performance and interpretability in multivariate time series classification is a significant challenge due to data complexity and high dimensionality. This paper introduces PHeatPruner, a method integrating persistent homology and sheaf theory to address these challenges. Persistent homology facilitates the pruning of up to 45% of the applied variables while maintaining or enhancing the accuracy of models such as Random Forest, CatBoost, XGBoost, and LightGBM, all without depending on posterior probabilities or supervised optimization algorithms. Concurrently, sheaf theory contributes explanatory vectors that provide deeper insights into the data's structural nuances. The approach was validated using the UEA Archive and a mastitis detection dataset for dairy cows. The results demonstrate that PHeatPruner effectively preserves model accuracy. Furthermore, our results highlight PHeatPruner's key features, i.e. simplifying complex data and offering actionable insights without increasing processing time or complexity. This method bridges the gap between complexity reduction and interpretability, suggesting promising applications in various fields.
Authors: Roya Nasiri
Abstract: The biases in artificial intelligence (AI) models can lead to automated decision-making processes that discriminate against groups and/or individuals based on sensitive properties such as gender and race. While there are many studies on diagnosing and mitigating biases in various AI models, there is little research on individual fairness in Graph Neural Networks (GNNs). Unlike traditional models, which treat data features independently and overlook their inter-relationships, GNNs are designed to capture graph-based structure where nodes are interconnected. This relational approach enables GNNs to model complex dependencies, but it also means that biases can propagate through these connections, complicating the detection and mitigation of individual fairness violations. This PhD project aims to develop a testing framework to assess and ensure individual fairness in GNNs. It first systematically reviews the literature on individual fairness, categorizing existing approaches to define, measure, test, and mitigate model biases, creating a taxonomy of individual fairness. Next, the project will develop a framework for testing and ensuring fairness in GNNs by adapting and extending current fairness testing and mitigation techniques. The framework will be evaluated through industrial case studies, focusing on graph-based large language models.
Authors: Irshad A. Meer, Bruno H\"ormann, Mustafa Ozger, Fabien Geyer, Alberto Viseras, Dominic Schupke, Cicek Cavdar
Abstract: The integration of unmanned aerial vehicles (UAVs) into cellular networks presents significant mobility management challenges, primarily due to frequent handovers caused by probabilistic line-of-sight conditions with multiple ground base stations (BSs). To tackle these challenges, reinforcement learning (RL)-based methods, particularly deep Q-networks (DQN), have been employed to optimize handover decisions dynamically. However, a major drawback of these learning-based approaches is their black-box nature, which limits interpretability in the decision-making process. This paper introduces an explainable AI (XAI) framework that incorporates Shapley Additive Explanations (SHAP) to provide deeper insights into how various state parameters influence handover decisions in a DQN-based mobility management system. By quantifying the impact of key features such as reference signal received power (RSRP), reference signal received quality (RSRQ), buffer status, and UAV position, our approach enhances the interpretability and reliability of RL-based handover solutions. To validate and compare our framework, we utilize real-world network performance data collected from UAV flight trials. Simulation results show that our method provides intuitive explanations for policy decisions, effectively bridging the gap between AI-driven models and human decision-makers.
Authors: Danial Dervovic, Michael Cashmore
Abstract: Missing data in supervised learning is well-studied, but the specific issue of missing labels during model evaluation has been overlooked. Ignoring samples with missing values, a common solution, can introduce bias, especially when data is Missing Not At Random (MNAR). We propose a multiple imputation technique for evaluating classifiers using metrics such as precision, recall, and ROC-AUC. This method not only offers point estimates but also a predictive distribution for these quantities when labels are missing. We empirically show that the predictive distribution's location and shape are generally correct, even in the MNAR regime. Moreover, we establish that this distribution is approximately Gaussian and provide finite-sample convergence bounds. Additionally, a robustness proof is presented, confirming the validity of the approximation under a realistic error model.
Authors: Marina Andric, Mauro Dragoni
Abstract: Length of hospital stay is a critical metric for assessing healthcare quality and optimizing hospital resource management. This study aims to identify factors influencing LoS within the Italian healthcare context, using a dataset of hospitalization records from over 60 healthcare facilities in the Piedmont region, spanning from 2020 to 2023. We explored a variety of features, including patient characteristics, comorbidities, admission details, and hospital-specific factors. Significant correlations were found between LoS and features such as age group, comorbidity score, admission type, and the month of admission. Machine learning models, specifically CatBoost and Random Forest, were used to predict LoS. The highest R2 score, 0.49, was achieved with CatBoost, demonstrating good predictive performance.
Authors: Rabanus Derr, Jessie Finocchiaro, Robert C. Williamson
Abstract: Fueled by discussions around "trustworthiness" and algorithmic fairness, calibration of predictive systems has regained scholars attention. The vanilla definition and understanding of calibration is, simply put, on all days on which the rain probability has been predicted to be p, the actual frequency of rain days was p. However, the increased attention has led to an immense variety of new notions of "calibration." Some of the notions are incomparable, serve different purposes, or imply each other. In this work, we provide two accounts which motivate calibration: self-realization of forecasted properties and precise estimation of incurred losses of the decision makers relying on forecasts. We substantiate the former via the reflection principle and the latter by actuarial fairness. For both accounts we formulate prototypical definitions via properties $\Gamma$ of outcome distributions, e.g., the mean or median. The prototypical definition for self-realization, which we call $\Gamma$-calibration, is equivalent to a certain type of swap regret under certain conditions. These implications are strongly connected to the omniprediction learning paradigm. The prototypical definition for precise loss estimation is a modification of decision calibration adopted from Zhao et al. [73]. For binary outcome sets both prototypical definitions coincide under appropriate choices of reference properties. For higher-dimensional outcome sets, both prototypical definitions can be subsumed by a natural extension of the binary definition, called distribution calibration with respect to a property. We conclude by commenting on the role of groupings in both accounts of calibration often used to obtain multicalibration. In sum, this work provides a semantic map of calibration in order to navigate a fragmented terrain of notions and definitions.
Authors: Vinicius L S Silva, Pablo Salinas, Claire E Heaney, Matthew Jackson, Christopher C Pain
Abstract: We propose a novel type of nonlinear solver acceleration for systems of nonlinear partial differential equations (PDEs) that is based on online/adaptive learning. It is applied in the context of multiphase flow in porous media. The proposed method rely on four pillars: (i) dimensionless numbers as input parameters for the machine learning model, (ii) simplified numerical model (two-dimensional) for the offline training, (iii) dynamic control of a nonlinear solver tuning parameter (numerical relaxation), (iv) and online learning for real-time improvement of the machine learning model. This strategy decreases the number of nonlinear iterations by dynamically modifying a single global parameter, the relaxation factor, and by adaptively learning the attributes of each numerical model on-the-run. Furthermore, this work performs a sensitivity study in the dimensionless parameters (machine learning features), assess the efficacy of various machine learning models, demonstrate a decrease in nonlinear iterations using our method in more intricate, realistic three-dimensional models, and fully couple a machine learning model into an open-source multiphase flow simulator achieving up to 85\% reduction in computational time.
Authors: Christopher B\"ulte, Yusuf Sale, Timo L\"ohr, Paul Hofman, Gitta Kutyniok, Eyke H\"ullermeier
Abstract: Uncertainty quantification (UQ) is crucial in machine learning, yet most (axiomatic) studies of uncertainty measures focus on classification, leaving a gap in regression settings with limited formal justification and evaluations. In this work, we introduce a set of axioms to rigorously assess measures of aleatoric, epistemic, and total uncertainty in supervised regression. By utilizing a predictive exponential family, we can generalize commonly used approaches for uncertainty representation and corresponding uncertainty measures. More specifically, we analyze the widely used entropy- and variance-based measures regarding limitations and challenges. Our findings provide a principled foundation for UQ in regression, offering theoretical insights and practical guidelines for reliable uncertainty assessment.
Authors: Kun He, Zijian Song, Shuoxi Zhang, John E. Hopcroft
Abstract: Class-Incremental Learning (CIL) is a critical capability for real-world applications, enabling learning systems to adapt to new tasks while retaining knowledge from previous ones. Recent advancements in pre-trained models (PTMs) have significantly advanced the field of CIL, demonstrating superior performance over traditional methods. However, understanding how features evolve and are distributed across incremental tasks remains an open challenge. In this paper, we propose a novel approach to modeling feature evolution in PTM-based CIL through the lens of neural collapse (NC), a striking phenomenon observed in the final phase of training, which leads to a well-separated, equiangular feature space. We explore the connection between NC and CIL effectiveness, showing that aligning feature distributions with the NC geometry enhances the ability to capture the dynamic behavior of continual learning. Based on this insight, we introduce Neural Collapse-inspired Pre-Trained Model-based CIL (NCPTM-CIL), a method that dynamically adjusts the feature space to conform to the elegant NC structure, thereby enhancing the continual learning process. Extensive experiments demonstrate that NCPTM-CIL outperforms state-of-the-art methods across four benchmark datasets. Notably, when initialized with ViT-B/16-IN1K, NCPTM-CIL surpasses the runner-up method by 6.73% on VTAB, 1.25% on CIFAR-100, and 2.5% on OmniBenchmark.
Authors: Tewodros Alemu Ayall, Andy Li, Matthew Beddows, Milan Markovic, Georgios Leontidis
Abstract: Due to rapid population growth globally, digitally-enabled agricultural sectors are crucial for sustainable food production and making informed decisions about resource management for farmers and various stakeholders. The deployment of Internet of Things (IoT) technologies that collect real-time observations of various environmental (e.g., temperature, humidity, etc.) and operational factors (e.g., irrigation) influencing production is often seen as a critical step to enable additional novel downstream tasks, such as AI-based yield forecasting. However, since AI models require large amounts of data, this creates practical challenges in a real-world dynamic farm setting where IoT observations would need to be collected over a number of seasons. In this study, we deployed IoT sensors in strawberry production polytunnels for two growing seasons to collect environmental data, including water usage, external and internal temperature, external and internal humidity, soil moisture, soil temperature, and photosynthetically active radiation. The sensor observations were combined with manually provided yield records spanning a period of four seasons. To bridge the gap of missing IoT observations for two additional seasons, we propose an AI-based backcasting approach to generate synthetic sensor observations using historical weather data from a nearby weather station and the existing polytunnel observations. We built an AI-based yield forecasting model to evaluate our approach using the combination of real and synthetic observations. Our results demonstrated that incorporating synthetic data improved yield forecasting accuracy, with models incorporating synthetic data outperforming those trained only on historical yield, weather records, and real sensor data.
Authors: Hiroki Naganuma, Xinzhi Zhang, Man-Chung Yue, Ioannis Mitliagkas, Philipp A. Witte, Russell J. Hewett, Yin Tat Lee
Abstract: Following AI scaling trends, frontier models continue to grow in size and continue to be trained on larger datasets. Training these models requires huge investments in exascale computational resources, which has in turn driven development of distributed deep learning methods. Data parallelism is an essential approach to speed up training, but it requires frequent global communication between workers, which can bottleneck training at the largest scales. In this work, we propose a method called Pseudo-Asynchronous Local SGD (PALSGD) to improve the efficiency of data-parallel training. PALSGD is an extension of Local SGD (Stich, 2018) and DiLoCo (Douillard et al., 2023), designed to further reduce communication frequency by introducing a pseudo-synchronization mechanism. PALSGD allows the use of longer synchronization intervals compared to standard Local SGD. Despite the reduced communication frequency, the pseudo-synchronization approach ensures that model consistency is maintained, leading to performance results comparable to those achieved with more frequent synchronization. Furthermore, we provide a theoretical analysis of PALSGD, establishing its convergence and deriving its convergence rate. This analysis offers insights into the algorithm's behavior and performance guarantees. We evaluated PALSGD on image classification and language modeling tasks. Our results show that PALSGD achieves better performance in less time compared to existing methods like Distributed Data Parallel (DDP), and DiLoCo. Notably, PALSGD trains 18.4% faster than DDP on ImageNet-1K with ResNet-50, 24.4% faster than DDP on TinyStories with GPT-Neo125M, and 21.1% faster than DDP on TinyStories with GPT-Neo-8M.
Authors: Sanjeev Raja, Martin \v{S}\'ipka, Michael Psenka, Tobias Kreiman, Michal Pavelka, Aditi S. Krishnapriyan
Abstract: Transition path sampling (TPS), which involves finding probable paths connecting two points on an energy landscape, remains a challenge due to the complexity of real-world atomistic systems. Current machine learning approaches use expensive, task-specific, and data-free training procedures, limiting their ability to benefit from recent advances in atomistic machine learning, such as high-quality datasets and large-scale pre-trained models. In this work, we address TPS by interpreting candidate paths as trajectories sampled from stochastic dynamics induced by the learned score function of pre-trained generative models, specifically denoising diffusion and flow matching. Under these dynamics, finding high-likelihood transition paths becomes equivalent to minimizing the Onsager-Machlup (OM) action functional. This enables us to repurpose pre-trained generative models for TPS in a zero-shot manner, in contrast with bespoke, task-specific TPS models trained in previous work. We demonstrate our approach on varied molecular systems, obtaining diverse, physically realistic transition pathways and generalizing beyond the pre-trained model's original training dataset. Our method can be easily incorporated into new generative models, making it practically relevant as models continue to scale and improve with increased data availability.
Authors: Han Zhang, Hao Zhou, Medhat Elsayed, Majid Bavand, Raimundas Gaigalas, Yigit Ozcan, Melike Erol-Kantarci
Abstract: Federated learning (FL) is a promising technique for learning-based functions in wireless networks, thanks to its distributed implementation capability. On the other hand, distributed learning may increase the risk of exposure to malicious attacks where attacks on a local model may spread to other models by parameter exchange. Meanwhile, such attacks can be hard to detect due to the dynamic wireless environment, especially considering local models can be heterogeneous with non-independent and identically distributed (non-IID) data. Therefore, it is critical to evaluate the effect of malicious attacks and develop advanced defense techniques for FL-enabled wireless networks. In this work, we introduce a federated deep reinforcement learning-based cell sleep control scenario that enhances the energy efficiency of the network. We propose multiple intelligent attacks targeting the learning-based approach and we propose defense methods to mitigate such attacks. In particular, we have designed two attack models, generative adversarial network (GAN)-enhanced model poisoning attack and regularization-based model poisoning attack. As a counteraction, we have proposed two defense schemes, autoencoder-based defense, and knowledge distillation (KD)-enabled defense. The autoencoder-based defense method leverages an autoencoder to identify the malicious participants and only aggregate the parameters of benign local models during the global aggregation, while KD-based defense protects the model from attacks by controlling the knowledge transferred between the global model and local models.
Authors: Yixiao Wang
Abstract: Imitation learning holds the promise of equipping robots with versatile skills by learning from expert demonstrations. However, policies trained on finite datasets often struggle to generalize beyond the training distribution. In this work, we present a unified perspective on the generalization capability of imitation learning, grounded in both information theorey and data distribution property. We first show that the generalization gap can be upper bounded by (i) the conditional information bottleneck on intermediate representations and (ii) the mutual information between the model parameters and the training dataset. This characterization provides theoretical guidance for designing effective training strategies in imitation learning, particularly in determining whether to freeze, fine-tune, or train large pretrained encoders (e.g., vision-language models or vision foundation models) from scratch to achieve better generalization. Furthermore, we demonstrate that high conditional entropy from input to output induces a flatter likelihood landscape, thereby reducing the upper bound on the generalization gap. In addition, it shortens the stochastic gradient descent (SGD) escape time from sharp local minima, which may increase the likelihood of reaching global optima under fixed optimization budgets. These insights explain why imitation learning often exhibits limited generalization and underscore the importance of not only scaling the diversity of input data but also enriching the variability of output labels conditioned on the same input.
Authors: Dhadkan Shrestha, Lincoln Bhattarai
Abstract: This paper explores the use of an extended neuroevolutionary approach, based on NeuroEvolution of Augmenting Topologies (NEAT), for autonomous robots in dynamic environments associated with hazardous tasks like firefighting, urban search-and-rescue (USAR), and industrial inspections. Building on previous research, it expands the simulation environment to larger and more complex settings, demonstrating NEAT's adaptability across different applications. By integrating recent advancements in NEAT and reinforcement learning, the study uses modern simulation frameworks for realism and hybrid algorithms for optimization. Experimental results show that NEAT-evolved controllers achieve success rates comparable to state-of-the-art deep reinforcement learning methods, with superior structural adaptability. The agents reached ~80% success in outdoor tests, surpassing baseline models. The paper also highlights the benefits of transfer learning among tasks and evaluates the effectiveness of NEAT in complex 3D navigation. Contributions include evaluating NEAT for diverse autonomous applications and discussing real-world deployment considerations, emphasizing the approach's potential as an alternative or complement to deep reinforcement learning in autonomous navigation tasks.
Authors: Ze Yang, Yihong Jin, Juntian Liu, Xinhe Xu, Yihan Zhang, Shuyang Ji
Abstract: The rapidly evolving cloud platforms and the escalating complexity of network traffic demand proper network traffic monitoring and anomaly detection to ensure network security and performance. This paper introduces a large language model (LLM)-based network traffic monitoring and anomaly detection system. In addition to existing models such as autoencoders and decision trees, we harness the power of large language models for processing sequence data from network traffic, which allows us a better capture of underlying complex patterns, as well as slight fluctuations in the dataset. We show for a given detection task, the need for a hybrid model that incorporates the attention mechanism of the transformer architecture into a supervised learning framework in order to achieve better accuracy. A pre-trained large language model analyzes and predicts the probable network traffic, and an anomaly detection layer that considers temporality and context is added. Moreover, we present a novel transfer learning-based methodology to enhance the model's effectiveness to quickly adapt to unknown network structures and adversarial conditions without requiring extensive labeled datasets. Actual results show that the designed model outperforms traditional methods in detection accuracy and computational efficiency, effectively identify various network anomalies such as zero-day attacks and traffic congestion pattern, and significantly reduce the false positive rate.
Authors: Anirudhan Badrinath, Alex Yang, Kousik Rajesh, Prabhat Agarwal, Jaewon Yang, Haoyu Chen, Jiajing Xu, Charles Rosenberg
Abstract: Representation learning, a task of learning latent vectors to represent entities, is a key task in improving search and recommender systems in web applications. Various representation learning methods have been developed, including graph-based approaches for relationships among entities, sequence-based methods for capturing the temporal evolution of user activities, and content-based models for leveraging text and visual content. However, the development of a unifying framework that integrates these diverse techniques to support multiple applications remains a significant challenge. This paper presents OmniSage, a large-scale representation framework that learns universal representations for a variety of applications at Pinterest. OmniSage integrates graph neural networks with content-based models and user sequence models by employing multiple contrastive learning tasks to effectively process graph data, user sequence data, and content signals. To support the training and inference of OmniSage, we developed an efficient infrastructure capable of supporting Pinterest graphs with billions of nodes. The universal representations generated by OmniSage have significantly enhanced user experiences on Pinterest, leading to an approximate 2.5% increase in sitewide repins (saves) across five applications. This paper highlights the impact of unifying representation learning methods, and we will open source the OmniSage code by the time of publication.
Authors: Mohaddeseh Chegini, Ali Mahloojifar
Abstract: The Computer_Aided Diagnosis (CAD) systems facilitate accurate diagnosis of diseases. The development of CADs by leveraging third generation neural network, namely, Spiking Neural Network (SNN), is essential to utilize of the benefits of SNNs, such as their event_driven processing, parallelism, low power consumption, and the ability to process sparse temporal_spatial information. However, Deep SNN as a deep learning model faces challenges with unreliability. To deal with unreliability challenges due to inability to quantify the uncertainty of the predictions, we proposed a deep Bayesian Convolutional Spiking Neural Network based_CADs with uncertainty_aware module. In this study, the Monte Carlo Dropout method as Bayesian approximation is used as an uncertainty quantification method. This method was applied to several medical image classification tasks. Our experimental results demonstrate that our proposed model is accurate and reliable and will be a proper alternative to conventional deep learning for medical image classification.
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.
Authors: Eviatar Bach, Ricardo Baptista, Edoardo Calvello, Bohan Chen, Andrew Stuart
Abstract: The filtering distribution in hidden Markov models evolves according to the law of a mean-field model in state--observation space. The ensemble Kalman filter (EnKF) approximates this mean-field model with an ensemble of interacting particles, employing a Gaussian ansatz for the joint distribution of the state and observation at each observation time. These methods are robust, but the Gaussian ansatz limits accuracy. This shortcoming is addressed by approximating the mean-field evolution using a novel form of neural operator taking probability distributions as input: a Measure Neural Mapping (MNM). A MNM is used to design a novel approach to filtering, the MNM-enhanced ensemble filter (MNMEF), which is defined in both the mean-fieldlimit and for interacting ensemble particle approximations. The ensemble approach uses empirical measures as input to the MNM and is implemented using the set transformer, which is invariant to ensemble permutation and allows for different ensemble sizes. The derivation of methods from a mean-field formulation allows a single parameterization of the algorithm to be deployed at different ensemble sizes. In practice fine-tuning of a small number of parameters, for specific ensemble sizes, further enhances the accuracy of the scheme. The promise of the approach is demonstrated by its superior root-mean-square-error performance relative to leading methods in filtering the Lorenz 96 and Kuramoto-Sivashinsky models.
Authors: Max Muchen Sun, Allison Pinosky, Todd Murphey
Abstract: Ergodic coverage effectively generates exploratory behaviors for embodied agents by aligning the spatial distribution of the agent's trajectory with a target distribution, where the difference between these two distributions is measured by the ergodic metric. However, existing ergodic coverage methods are constrained by the limited set of ergodic metrics available for control synthesis, fundamentally limiting their performance. In this work, we propose an alternative approach to ergodic coverage based on flow matching, a technique widely used in generative inference for efficient and scalable sampling. We formally derive the flow matching problem for ergodic coverage and show that it is equivalent to a linear quadratic regulator problem with a closed-form solution. Our formulation enables alternative ergodic metrics from generative inference that overcome the limitations of existing ones. These metrics were previously infeasible for control synthesis but can now be supported with no computational overhead. Specifically, flow matching with the Stein variational gradient flow enables control synthesis directly over the score function of the target distribution, improving robustness to the unnormalized distributions; on the other hand, flow matching with the Sinkhorn divergence flow enables an optimal transport-based ergodic metric, improving coverage performance on non-smooth distributions with irregular supports. We validate the improved performance and competitive computational efficiency of our method through comprehensive numerical benchmarks and across different nonlinear dynamics. We further demonstrate the practicality of our method through a series of drawing and erasing tasks on a Franka robot.
Authors: Zemin Zheng, Xin Zhou, Jinchi Lv
Abstract: Data reduction with uncertainty quantification plays a key role in various multi-task learning applications, where large numbers of responses and features are present. To this end, a general framework of high-dimensional manifold-based SOFAR inference (SOFARI) was introduced recently in Zheng, Zhou, Fan and Lv (2024) for interpretable multi-task learning inference focusing on the left factor vectors and singular values exploiting the latent singular value decomposition (SVD) structure. Yet, designing a valid inference procedure on the latent right factor vectors is not straightforward from that of the left ones and can be even more challenging due to asymmetry of left and right singular vectors in the response matrix. To tackle these issues, in this paper we suggest a new method of high-dimensional manifold-based SOFAR inference for latent responses (SOFARI-R), where two variants of SOFARI-R are introduced. The first variant deals with strongly orthogonal factors by coupling left singular vectors with the design matrix and then appropriately rescaling them to generate new Stiefel manifolds. The second variant handles the more general weakly orthogonal factors by employing the hard-thresholded SOFARI estimates and delicately incorporating approximation errors into the distribution. Both variants produce bias-corrected estimators for the latent right factor vectors that enjoy asymptotically normal distributions with justified asymptotic variance estimates. We demonstrate the effectiveness of the newly suggested method using extensive simulation studies and an economic application.
Authors: Abrar Fahim, Shamik Dey, Md. Nurul Absur, Md Kamrul Siam, Md. Tahmidul Huque, Jafreen Jafor Godhuli
Abstract: Digital systems find it challenging to keep up with cybersecurity threats. The daily emergence of more than 560,000 new malware strains poses significant hazards to the digital ecosystem. The traditional malware detection methods fail to operate properly and yield high false positive rates with low accuracy of the protection system. This study explores the ways in which malware can be detected using these machine learning (ML) and deep learning (DL) approaches to address those shortcomings. This study also includes a systematic comparison of the performance of some of the widely used ML models, such as random forest, multi-layer perceptron (MLP), and deep neural network (DNN), for determining the effectiveness of the domain of modern malware threat systems. We use a considerable-sized database from Kaggle, which has undergone optimized feature selection and preprocessing to improve model performance. Our finding suggests that the DNN model outperformed the other traditional models with the highest training accuracy of 99.92% and an almost perfect AUC score. Furthermore, the feature selection and preprocessing can help improve the capabilities of detection. This research makes an important contribution by analyzing the performance of the model on the performance metrics and providing insight into the effectiveness of the advanced detection techniques to build more robust and more reliable cybersecurity solutions against the growing malware threats.
Authors: Raza Ul Mustafa, Sesha Dassanayake
Abstract: The Quality of Experience (QoE) is the users satisfaction while streaming a video session over an over-the-top (OTT) platform like YouTube. QoE of YouTube reflects the smooth streaming session without any buffering and quality shift events. One of the most important factors nowadays affecting QoE of YouTube is frequent shifts from higher to lower resolutions and vice versa. These shifts ensure a smooth streaming session; however, it might get a lower mean opinion score. For instance, dropping from 1080p to 480p during a video can preserve continuity but might reduce the viewers enjoyment. Over time, OTT platforms are looking for alternative ways to boost user experience instead of relying on traditional Quality of Service (QoS) metrics such as bandwidth, latency, and throughput. As a result, we look into the relationship between quality shifting in YouTube streaming sessions and the channel metrics RSRP, RSRQ, and SNR. Our findings state that these channel metrics positively correlate with shifts. Thus, in real-time, OTT can only rely on them to predict video streaming sessions into lower- and higher-resolution categories, thus providing more resources to improve user experience. Using traditional Machine Learning (ML) classifiers, we achieved an accuracy of 77-percent, while using only RSRP, RSRQ, and SNR. In the era of 5G and beyond, where ultra-reliable, low-latency networks promise enhanced streaming capabilities, the proposed methodology can be used to improve OTT services.
Authors: Josua Spisak, Sergiu Tcaci Popescu, Stefan Wermter, Matej Hoffmann, J. Kevin O'Regan
Abstract: We present a computational model of the mechanisms that may determine infants' behavior in the "mobile paradigm". This paradigm has been used in developmental psychology to explore how infants learn the sensory effects of their actions. In this paradigm, a mobile (an articulated and movable object hanging above an infant's crib) is connected to one of the infant's limbs, prompting the infant to preferentially move that "connected" limb. This ability to detect a "sensorimotor contingency" is considered to be a foundational cognitive ability in development. To understand how infants learn sensorimotor contingencies, we built a model that attempts to replicate infant behavior. Our model incorporates a neural network, action-outcome prediction, exploration, motor noise, preferred activity level, and biologically-inspired motor control. We find that simulations with our model replicate the classic findings in the literature showing preferential movement of the connected limb. An interesting observation is that the model sometimes exhibits a burst of movement after the mobile is disconnected, casting light on a similar occasional finding in infants. In addition to these general findings, the simulations also replicate data from two recent more detailed studies using a connection with the mobile that was either gradual or all-or-none. A series of ablation studies further shows that the inclusion of mechanisms of action-outcome prediction, exploration, motor noise, and biologically-inspired motor control was essential for the model to correctly replicate infant behavior. This suggests that these components are also involved in infants' sensorimotor learning.
Authors: Ahod Alghuried, Abdulaziz Alghamdi, Ali Alkinoon, Soohyeon Choi, Manar Mohaisen, David Mohaisen
Abstract: Phishing detection on Ethereum has increasingly leveraged advanced machine learning techniques to identify fraudulent transactions. However, limited attention has been given to understanding the effectiveness of feature selection strategies and the role of graph-based models in enhancing detection accuracy. In this paper, we systematically examine these issues by analyzing and contrasting explicit transactional features and implicit graph-based features, both experimentally and analytically. We explore how different feature sets impact the performance of phishing detection models, particularly in the context of Ethereum's transactional network. Additionally, we address key challenges such as class imbalance and dataset composition and their influence on the robustness and precision of detection methods. Our findings demonstrate the advantages and limitations of each feature type, while also providing a clearer understanding of how feature affect model resilience and generalization in adversarial environments.
Authors: Kaiyuan Tang, Siyuan Yao, Chaoli Wang
Abstract: In volume visualization, users can interactively explore the three-dimensional data by specifying color and opacity mappings in the transfer function (TF) or adjusting lighting parameters, facilitating meaningful interpretation of the underlying structure. However, rendering large-scale volumes demands powerful GPUs and high-speed memory access for real-time performance. While existing novel view synthesis (NVS) methods offer faster rendering speeds with lower hardware requirements, the visible parts of a reconstructed scene are fixed and constrained by preset TF settings, significantly limiting user exploration. This paper introduces inverse volume rendering via Gaussian splatting (iVR-GS), an innovative NVS method that reduces the rendering cost while enabling scene editing for interactive volume exploration. Specifically, we compose multiple iVR-GS models associated with basic TFs covering disjoint visible parts to make the entire volumetric scene visible. Each basic model contains a collection of 3D editable Gaussians, where each Gaussian is a 3D spatial point that supports real-time scene rendering and editing. We demonstrate the superior reconstruction quality and composability of iVR-GS against other NVS solutions (Plenoxels, CCNeRF, and base 3DGS) on various volume datasets. The code is available at https://github.com/TouKaienn/iVR-GS.
Authors: Yinlong Dai, Robert Ramirez Sanchez, Ryan Jeronimus, Shahabedin Sagheb, Cara M. Nunez, Heramb Nemlekar, Dylan P. Losey
Abstract: Today's robots learn new tasks by imitating human examples. However, this standard approach to visual imitation learning is fundamentally limited: the robot observes what the human does, but not why the human chooses those behaviors. Without understanding the features that factor into the human's decisions, robot learners often misinterpret the data and fail to perform the task when the environment changes. We therefore propose a shift in perspective: instead of asking human teachers just to show what actions the robot should take, we also enable humans to indicate task-relevant features using markers and language prompts. Our proposed algorithm, CIVIL, leverages this augmented data to filter the robot's visual observations and extract a feature representation that causally informs human actions. CIVIL then applies these causal features to train a transformer-based policy that emulates human behaviors without being confused by visual distractors. Our simulations, real-world experiments, and user study demonstrate that robots trained with CIVIL can learn from fewer human demonstrations and perform better than state-of-the-art baselines, especially in previously unseen scenarios. See videos at our project website: https://civil2025.github.io
Authors: Kaiyuan Tan, Peilun Li, Jun Wang, Thomas Beckers
Abstract: The ability to predict trajectories of surrounding agents and obstacles is a crucial component in many robotic applications. Data-driven approaches are commonly adopted for state prediction in scenarios where the underlying dynamics are unknown. However, the performance, reliability, and uncertainty of data-driven predictors become compromised when encountering out-of-distribution observations relative to the training data. In this paper, we introduce a Plug-and-Play Physics-Informed Machine Learning (PnP-PIML) framework to address this challenge. Our method employs conformal prediction to identify outlier dynamics and, in that case, switches from a nominal predictor to a physics-consistent model, namely distributed Port-Hamiltonian systems (dPHS). We leverage Gaussian processes to model the energy function of the dPHS, enabling not only the learning of system dynamics but also the quantification of predictive uncertainty through its Bayesian nature. In this way, the proposed framework produces reliable physics-informed predictions even for the out-of-distribution scenarios.
Authors: Chang Xiao, Brenda Yang
Abstract: Generative conversational interfaces powered by large language models (LLMs) typically stream output token-by-token at a rate determined by computational budget, often neglecting actual human reading speeds and the cognitive load associated with the content. This mismatch frequently leads to inefficient use of computational resources. For example, in cloud-based services, streaming content faster than users can read appears unnecessary, resulting in wasted computational resources and potential delays for other users, particularly during peak usage periods. To address this issue, we propose an adaptive streaming method that dynamically adjusts the pacing of LLM streaming output in real-time based on inferred cognitive load. Our approach estimates the cognitive load associated with streaming content and strategically slows down the stream during complex or information-rich segments, thereby freeing computational resources for other users. Our statistical analysis of computational savings, combined with crowdsourced user studies, provides insights into the trade-offs between service efficiency and user satisfaction, demonstrating that our method can significantly reduce computational consumption up to 16.8\%. This context-aware computational resource management strategy presents a practical framework for enhancing system efficiency in cloud-based conversational AI interfaces without compromising user experience.
Authors: Yazan Otoum, Amiya Nayak
Abstract: With the rapid digitalization of healthcare systems, there has been a substantial increase in the generation and sharing of private health data. Safeguarding patient information is essential for maintaining consumer trust and ensuring compliance with legal data protection regulations. Machine learning is critical in healthcare, supporting personalized treatment, early disease detection, predictive analytics, image interpretation, drug discovery, efficient operations, and patient monitoring. It enhances decision-making, accelerates research, reduces errors, and improves patient outcomes. In this paper, we utilize machine learning methodologies, including differential privacy and federated learning, to develop privacy-preserving models that enable healthcare stakeholders to extract insights without compromising individual privacy. Differential privacy introduces noise to data to guarantee statistical privacy, while federated learning enables collaborative model training across decentralized datasets. We explore applying these technologies to Heart Disease Data, demonstrating how they preserve privacy while delivering valuable insights and comprehensive analysis. Our results show that using a federated learning model with differential privacy achieved a test accuracy of 85%, ensuring patient data remained secure and private throughout the process.
Authors: Hanrui Wang, Shuo Wang, Chun-Shien Lu, Isao Echizen
Abstract: Facial recognition technology poses significant privacy risks, as it relies on biometric data that is inherently sensitive and immutable if compromised. To mitigate these concerns, face recognition systems convert raw images into embeddings, traditionally considered privacy-preserving. However, model inversion attacks pose a significant privacy threat by reconstructing these private facial images, making them a crucial tool for evaluating the privacy risks of face recognition systems. Existing methods usually require training individual generators for each target model, a computationally expensive process. In this paper, we propose DiffUMI, a training-free diffusion-driven universal model inversion attack for face recognition systems. DiffUMI is the first approach to apply a diffusion model for unconditional image generation in model inversion. Unlike other methods, DiffUMI is universal, eliminating the need for training target-specific generators. It operates within a fixed framework and pretrained diffusion model while seamlessly adapting to diverse target identities and models. DiffUMI breaches privacy-preserving face recognition systems with state-of-the-art success, demonstrating that an unconditional diffusion model, coupled with optimized adversarial search, enables efficient and high-fidelity facial reconstruction. Additionally, we introduce a novel application of out-of-domain detection (OODD), marking the first use of model inversion to distinguish non-face inputs from face inputs based solely on embeddings.
Authors: Sourav Chatterjee, Timothy Sudijono
Abstract: One of the enduring problems surrounding neural networks is to identify the factors that differentiate them from traditional statistical models. We prove a pair of results which distinguish feedforward neural networks among parametric models at the population level, for regression tasks. Firstly, we prove that for any pair of random variables $(X,Y)$, neural networks always learn a nontrivial relationship between $X$ and $Y$, if one exists. Secondly, we prove that for reasonable smooth parametric models, under local and global identifiability conditions, there exists a nontrivial $(X,Y)$ pair for which the parametric model learns the constant predictor $\mathbb{E}[Y]$. Together, our results suggest that a lack of identifiability distinguishes neural networks among the class of smooth parametric models.
Authors: Wataru Kawakami, Keita Suzuki, Junichiro Iwasawa
Abstract: Large Language Models (LLMs) show potential in medicine, yet clinical adoption is hindered by concerns over factual accuracy, language-specific limitations (e.g., Japanese), and critically, their reliability when required to generate reasoning explanations -- a prerequisite for trust. This paper introduces Preferred-MedLLM-Qwen-72B, a 72B-parameter model optimized for the Japanese medical domain to achieve both high accuracy and stable reasoning. We employ a two-stage fine-tuning process on the Qwen2.5-72B base model: first, Continued Pretraining (CPT) on a comprehensive Japanese medical corpus instills deep domain knowledge. Second, Reasoning Preference Optimization (RPO), a preference-based method, enhances the generation of reliable reasoning pathways while preserving high answer accuracy. Evaluations on the Japanese Medical Licensing Exam benchmark (IgakuQA) show Preferred-MedLLM-Qwen-72B achieves state-of-the-art performance (0.868 accuracy), surpassing strong proprietary models like GPT-4o (0.866). Crucially, unlike baseline or CPT-only models which exhibit significant accuracy degradation (up to 11.5\% and 3.8\% respectively on IgakuQA) when prompted for explanations, our model maintains its high accuracy (0.868) under such conditions. This highlights RPO's effectiveness in stabilizing reasoning generation. This work underscores the importance of optimizing for reliable explanations alongside accuracy. We release the Preferred-MedLLM-Qwen-72B model weights to foster research into trustworthy LLMs for specialized, high-stakes applications.
Authors: Muhammad Mubashar, Shireen Kudukkil Manchingal, Fabio Cuzzolin
Abstract: Large Language Models (LLMs) are known to produce very high-quality tests and responses to our queries. But how much can we trust this generated text? In this paper, we study the problem of uncertainty quantification in LLMs. We propose a novel Random-Set Large Language Model (RSLLM) approach which predicts finite random sets (belief functions) over the token space, rather than probability vectors as in classical LLMs. In order to allow so efficiently, we also present a methodology based on hierarchical clustering to extract and use a budget of "focal" subsets of tokens upon which the belief prediction is defined, rather than using all possible collections of tokens, making the method scalable yet effective. RS-LLMs encode the epistemic uncertainty induced in their generation process by the size and diversity of its training set via the size of the credal sets associated with the predicted belief functions. The proposed approach is evaluated on CoQA and OBQA datasets using Llama2-7b, Mistral-7b and Phi-2 models and is shown to outperform the standard model in both datasets in terms of correctness of answer while also showing potential in estimating the second level uncertainty in its predictions and providing the capability to detect when its hallucinating.
Authors: Natansh Mathur, Brian Coyle, Nishant Jain, Snehal Raj, Akshat Tandon, Jasper Simon Krauser, Rainer Stoessel
Abstract: Identification of defects or anomalies in 3D objects is a crucial task to ensure correct functionality. In this work, we combine Bayesian learning with recent developments in quantum and quantum-inspired machine learning, specifically orthogonal neural networks, to tackle this anomaly detection problem for an industrially relevant use case. Bayesian learning enables uncertainty quantification of predictions, while orthogonality in weight matrices enables smooth training. We develop orthogonal (quantum) versions of 3D convolutional neural networks and show that these models can successfully detect anomalies in 3D objects. To test the feasibility of incorporating quantum computers into a quantum-enhanced anomaly detection pipeline, we perform hardware experiments with our models on IBM's 127-qubit Brisbane device, testing the effect of noise and limited measurement shots.
Authors: Atharva Kulkarni, Yuan Zhang, Joel Ruben Antony Moniz, Xiou Ge, Bo-Hsiang Tseng, Dhivya Piraviperumal, Swabha Swayamdipta, Hong Yu
Abstract: Hallucinations pose a significant obstacle to the reliability and widespread adoption of language models, yet their accurate measurement remains a persistent challenge. While many task- and domain-specific metrics have been proposed to assess faithfulness and factuality concerns, the robustness and generalization of these metrics are still untested. In this paper, we conduct a large-scale empirical evaluation of 6 diverse sets of hallucination detection metrics across 4 datasets, 37 language models from 5 families, and 5 decoding methods. Our extensive investigation reveals concerning gaps in current hallucination evaluation: metrics often fail to align with human judgments, take an overtly myopic view of the problem, and show inconsistent gains with parameter scaling. Encouragingly, LLM-based evaluation, particularly with GPT-4, yields the best overall results, and mode-seeking decoding methods seem to reduce hallucinations, especially in knowledge-grounded settings. These findings underscore the need for more robust metrics to understand and quantify hallucinations, and better strategies to mitigate them.
Authors: Miranda C. N. Cheng, Niki Stratikopoulou
Abstract: Numerical simulations of quantum field theories on lattices serve as a fundamental tool for studying the non-perturbative regime of the theories, where analytic tools often fall short. Challenges arise when one takes the continuum limit or as the system approaches a critical point, especially in the presence of non-trivial topological structures in the theory. Rapid recent advances in machine learning provide a promising avenue for progress in this area. These lecture notes aim to give a brief account of lattice field theories, normalizing flows, and how the latter can be applied to study the former. The notes are based on the lectures given by the first author in various recent research schools.
Authors: Tatsunori Tanaka, Fi Zheng, Kai Sato, Zhifeng Li, Yuanyun Zhang, Shi Li
Abstract: Clinical language models have achieved strong performance on downstream tasks by pretraining on domain specific corpora such as discharge summaries and medical notes. However, most approaches treat the electronic health record as a static document, neglecting the temporally-evolving and causally entwined nature of patient trajectories. In this paper, we introduce a novel temporal entailment pretraining objective for language models in the clinical domain. Our method formulates EHR segments as temporally ordered sentence pairs and trains the model to determine whether a later state is entailed by, contradictory to, or neutral with respect to an earlier state. Through this temporally structured pretraining task, models learn to perform latent clinical reasoning over time, improving their ability to generalize across forecasting and diagnosis tasks. We pretrain on a large corpus derived from MIMIC IV and demonstrate state of the art results on temporal clinical QA, early warning prediction, and disease progression modeling.
Authors: Rob Romijnders, Stefanos Laskaridis, Ali Shahin Shamsabadi, Hamed Haddadi
Abstract: Large Language Models (LLM) are typically trained on vast amounts of data from various sources. Even when designed modularly (e.g., Mixture-of-Experts), LLMs can leak privacy on their sources. Conversely, training such models in isolation arguably prohibits generalization. To this end, we propose a framework, NoEsis, which builds upon the desired properties of modularity, privacy, and knowledge transfer. NoEsis integrates differential privacy with a hybrid two-staged parameter-efficient fine-tuning that combines domain-specific low-rank adapters, acting as experts, with common prompt tokens, acting as a knowledge-sharing backbone. Results from our evaluation on CodeXGLUE showcase that NoEsis can achieve provable privacy guarantees with tangible knowledge transfer across domains, and empirically show protection against Membership Inference Attacks. Finally, on code completion tasks, NoEsis bridges at least 77% of the accuracy gap between the non-shared and the non-private baseline.
Authors: Michel Gokan Khan, Renan Guarese, Fabian Johnson, Xi Vincent Wang, Anders Bergman, Benjamin Edvinsson, Mario Romero, J\'er\'emy Vachier, Jan Kronqvist
Abstract: We introduce PerfCam, an open source Proof-of-Concept (PoC) digital twinning framework that combines camera and sensory data with 3D Gaussian Splatting and computer vision models for digital twinning, object tracking, and Key Performance Indicators (KPIs) extraction in industrial production lines. By utilizing 3D reconstruction and Convolutional Neural Networks (CNNs), PerfCam offers a semi-automated approach to object tracking and spatial mapping, enabling digital twins that capture real-time KPIs such as availability, performance, Overall Equipment Effectiveness (OEE), and rate of conveyor belts in the production line. We validate the effectiveness of PerfCam through a practical deployment within realistic test production lines in the pharmaceutical industry and contribute an openly published dataset to support further research and development in the field. The results demonstrate PerfCam's ability to deliver actionable insights through its precise digital twin capabilities, underscoring its value as an effective tool for developing usable digital twins in smart manufacturing environments and extracting operational analytics.
Authors: Lovro Sindicic, Ivica Kopriva
Abstract: Deep subspace clustering (DSC) algorithms face several challenges that hinder their widespread adoption across variois application domains. First, clustering quality is typically assessed using only the encoder's output layer, disregarding valuable information present in the intermediate layers. Second, most DSC approaches treat representation learning and subspace clustering as independent tasks, limiting their effectiveness. Third, they assume the availability of a held-out dataset for hyperparameter tuning, which is often impractical in real-world scenarios. Fourth, learning termination is commonly based on clustering error monitoring, requiring external labels. Finally, their performance often depends on post-processing techniques that rely on labeled data. To address this limitations, we introduce a novel single-view DSC approach that: (i) minimizes a layer-wise self expression loss using a joint representation matrix; (ii) optimizes a subspace-structured norm to enhance clustering quality; (iii) employs a multi-stage sequential learning framework, consisting of pre-training and fine-tuning, enabling the use of multiple regularization terms without hyperparameter tuning; (iv) incorporates a relative error-based self-stopping mechanism to terminate training without labels; and (v) retains a fixed number of leading coefficients in the learned representation matrix based on prior knowledge. We evaluate the proposed method on six datasets representing faces, digits, and objects. The results show that our method outperforms most linear SC algorithms with careffulyl tuned hyperparameters while maintaining competitive performance with the best performing linear appoaches.
Authors: {\TH}\'orir Hrafn Har{\dh}arson, Hrafn Loftsson, Stef\'an \'Olafsson
Abstract: The integration of language models in the legal domain holds considerable promise for streamlining processes and improving efficiency in managing extensive workloads. However, the specialized terminology, nuanced language, and formal style of legal texts can present substantial challenges. This study examines whether preference-based training techniques, specifically Reinforcement Learning from Human Feedback and Direct Preference Optimization, can enhance models' performance in generating Icelandic legal summaries that align with domain-specific language standards and user preferences. We compare models fine-tuned with preference training to those using conventional supervised learning. Results indicate that preference training improves the legal accuracy of generated summaries over standard fine-tuning but does not significantly enhance the overall quality of Icelandic language usage. Discrepancies between automated metrics and human evaluations further underscore the importance of qualitative assessment in developing language models for the legal domain.
Authors: Jia-Qi Yang, Lei Shi
Abstract: This paper investigates regularized stochastic gradient descent (SGD) algorithms for estimating nonlinear operators from a Polish space to a separable Hilbert space. We assume that the regression operator lies in a vector-valued reproducing kernel Hilbert space induced by an operator-valued kernel. Two significant settings are considered: an online setting with polynomially decaying step sizes and regularization parameters, and a finite-horizon setting with constant step sizes and regularization parameters. We introduce regularity conditions on the structure and smoothness of the target operator and the input random variables. Under these conditions, we provide a dimension-free convergence analysis for the prediction and estimation errors, deriving both expectation and high-probability error bounds. Our analysis demonstrates that these convergence rates are nearly optimal. Furthermore, we present a new technique for deriving bounds with high probability for general SGD schemes, which also ensures almost-sure convergence. Finally, we discuss potential extensions to more general operator-valued kernels and the encoder-decoder framework.
Authors: Raul David Dominguez Sanchez, Xavier Diaz Ortiz, Xingcheng Zhou, Max Peter Ronecker, Michael Karner, Daniel Watzenig, Alois Knoll
Abstract: Railway systems, particularly in Germany, require high levels of automation to address legacy infrastructure challenges and increase train traffic safely. A key component of automation is robust long-range perception, essential for early hazard detection, such as obstacles at level crossings or pedestrians on tracks. Unlike automotive systems with braking distances of ~70 meters, trains require perception ranges exceeding 1 km. This paper presents an deep-learning-based approach for long-range 3D object detection tailored for autonomous trains. The method relies solely on monocular images, inspired by the Faraway-Frustum approach, and incorporates LiDAR data during training to improve depth estimation. The proposed pipeline consists of four key modules: (1) a modified YOLOv9 for 2.5D object detection, (2) a depth estimation network, and (3-4) dedicated short- and long-range 3D detection heads. Evaluations on the OSDaR23 dataset demonstrate the effectiveness of the approach in detecting objects up to 250 meters. Results highlight its potential for railway automation and outline areas for future improvement.
Authors: Nguyen Vu Khai Tam, Cao Huyen My, Vo Nguyen Le Duy
Abstract: Transfer learning (TL) for high-dimensional regression (HDR) is an important problem in machine learning, particularly when dealing with limited sample size in the target task. However, there currently lacks a method to quantify the statistical significance of the relationship between features and the response in TL-HDR settings. In this paper, we introduce a novel statistical inference framework for assessing the reliability of feature selection in TL-HDR, called PTL-SI (Post-TL Statistical Inference). The core contribution of PTL-SI is its ability to provide valid $p$-values to features selected in TL-HDR, thereby rigorously controlling the false positive rate (FPR) at desired significance level $\alpha$ (e.g., 0.05). Furthermore, we enhance statistical power by incorporating a strategic divide-and-conquer approach into our framework. We demonstrate the validity and effectiveness of the proposed PTL-SI through extensive experiments on both synthetic and real-world high-dimensional datasets, confirming its theoretical properties and utility in testing the reliability of feature selection in TL scenarios.
Authors: Petar Labura, Tomislav Antic, Tomislav Capuder
Abstract: The widespread integration of new technologies in low-voltage distribution networks on the consumer side creates the need for distribution system operators to perform advanced real-time calculations to estimate network conditions. In recent years, data-driven models based on machine learning and big data analysis have emerged for calculation purposes, leveraging the information available in large datasets obtained from smart meters and other advanced measurement infrastructure. However, existing data-driven algorithms do not take into account the quality of data collected from smart meters. They lack built-in anomaly detection mechanisms and fail to differentiate anomalies based on whether the value or context of anomalous data instances deviates from the norm. This paper focuses on methods for detecting and mitigating the impact of anomalies on the consumption of active and reactive power datasets. It proposes an anomaly detection framework based on the Isolation Forest machine learning algorithm and Fast Fourier Transform filtering that works in both the time and frequency domain and is unaffected by point anomalies or contextual anomalies of the power consumption data. The importance of integrating anomaly detection methods is demonstrated in the analysis important for distribution networks with a high share of smart meters.
Authors: Surajit Majumder, Paritosh Ranjan, Prodip Roy, Bhuban Padhan
Abstract: This paper introduces decentralized and modular neural network framework designed to enhance the scalability, interpretability, and performance of artificial intelligence (AI) systems. At the heart of this framework is a dynamic switch mechanism that governs the selective activation and training of individual neurons based on input characteristics, allowing neurons to specialize in distinct segments of the data domain. This approach enables neurons to learn from disjoint subsets of data, mimicking biological brain function by promoting task specialization and improving the interpretability of neural network behavior. Furthermore, the paper explores the application of federated learning and decentralized training for real-world AI deployments, particularly in edge computing and distributed environments. By simulating localized training on non-overlapping data subsets, we demonstrate how modular networks can be efficiently trained and evaluated. The proposed framework also addresses scalability, enabling AI systems to handle large datasets and distributed processing while preserving model transparency and interpretability. Finally, we discuss the potential of this approach in advancing the design of scalable, privacy-preserving, and efficient AI systems for diverse applications.
Authors: Ritesh Goru, Shanay Mehta, Prateek Jain
Abstract: Training Large Language Models ( LLMs) to generate explicit reasoning before they produce an answer has been shown to improve their performance across various tasks such as mathematics and coding. However, fine-tuning LLMs on multi-turn reasoning datasets presents a unique challenge: LLMs must generate reasoning tokens that are excluded from subsequent inputs to the LLM. This discrepancy prevents us from processing an entire conversation in a single forward pass-an optimization readily available when we fine-tune on a multi-turn non-reasoning dataset. This paper proposes a novel approach that overcomes this limitation through response token duplication and a custom attention mask that enforces appropriate visibility constraints. Our approach significantly reduces the training time and allows efficient fine-tuning on multi-turn reasoning datasets.
Authors: Qinyu Chen, Chang Gao, Min Liu, Daniele Perrone, Yan Ru Pei, Zuowen Wang, Zhuo Zou, Shihang Tan, Tao Han, Guorui Lu, Zhen Xu, Junyuan Ding, Ziteng Wang, Zongwei Wu, Han Han, Yuliang Wu, Jinze Chen, Wei Zhai, Yang Cao, Zheng-jun Zha, Nuwan Bandara, Thivya Kandappu, Archan Misra, Xiaopeng Lin, Hongxiang Huang, Hongwei Ren, Bojun Cheng, Hoang M. Truong, Vinh-Thuan Ly, Huy G. Tran, Thuan-Phat Nguyen, Tram T. Doan
Abstract: This survey serves as a review for the 2025 Event-Based Eye Tracking Challenge organized as part of the 2025 CVPR event-based vision workshop. This challenge focuses on the task of predicting the pupil center by processing event camera recorded eye movement. We review and summarize the innovative methods from teams rank the top in the challenge to advance future event-based eye tracking research. In each method, accuracy, model size, and number of operations are reported. In this survey, we also discuss event-based eye tracking from the perspective of hardware design.
Authors: Jonathan Kouchly, Ben Finkelshtein, Michael Bronstein, Ron Levie
Abstract: Learning on large graphs presents significant challenges, with traditional Message Passing Neural Networks suffering from computational and memory costs scaling linearly with the number of edges. We introduce the Intersecting Block Graph (IBG), a low-rank factorization of large directed graphs based on combinations of intersecting bipartite components, each consisting of a pair of communities, for source and target nodes. By giving less weight to non-edges, we show how to efficiently approximate any graph, sparse or dense, by a dense IBG. Specifically, we prove a constructive version of the weak regularity lemma, showing that for any chosen accuracy, every graph, regardless of its size or sparsity, can be approximated by a dense IBG whose rank depends only on the accuracy. This dependence of the rank solely on the accuracy, and not on the sparsity level, is in contrast to previous forms of the weak regularity lemma. We present a graph neural network architecture operating on the IBG representation of the graph and demonstrating competitive performance on node classification, spatio-temporal graph analysis, and knowledge graph completion, while having memory and computational complexity linear in the number of nodes rather than edges.
Authors: Christian Pionzewski, Rebecca Rademacher, J\'er\^ome Rutinowski, Antonia Ponikarov, Stephan Matzke, Tim Chilla, Pia Schreynemackers, Alice Kirchheim
Abstract: This contribution explores the impact of synthetic training data usage and the prediction of material wear and aging in the context of re-identification. Different experimental setups and gallery set expanding strategies are tested, analyzing their impact on performance over time for aging re-identification subjects. Using a continuously updating gallery, we were able to increase our mean Rank-1 accuracy by 24%, as material aging was taken into account step by step. In addition, using models trained with 10% artificial training data, Rank-1 accuracy could be increased by up to 13%, in comparison to a model trained on only real-world data, significantly boosting generalized performance on hold-out data. Finally, this work introduces a novel, open-source re-identification dataset, pallet-block-2696. This dataset contains 2,696 images of Euro pallets, taken over a period of 4 months. During this time, natural aging processes occurred and some of the pallets were damaged during their usage. These wear and tear processes significantly changed the appearance of the pallets, providing a dataset that can be used to generate synthetically aged pallets or other wooden materials.
Authors: Prashant Garg, Thiemo Fetzer
Abstract: Using basic health statements authorized by UK and EU registers and 9,100 journalist-vetted public-health assertions on topics such as abortion, COVID-19 and politics from sources ranging from peer-reviewed journals and government advisories to social media and news across the political spectrum, we benchmark six leading large language models from in 21 languages, finding that, despite high accuracy on English-centric textbook claims, performance falls in multiple non-European languages and fluctuates by topic and source, highlighting the urgency of comprehensive multilingual, domain-aware validation before deploying AI in global health communication.
Authors: Yangyang Xu, Kexin Li, Li Yang, You-Wei Wen
Abstract: Tensor Robust Principal Component Analysis (TRPCA) is a fundamental technique for decomposing multi-dimensional data into a low-rank tensor and an outlier tensor, yet existing methods relying on sparse outlier assumptions often fail under structured corruptions. In this paper, we propose a self-guided data augmentation approach that employs adaptive weighting to suppress outlier influence, reformulating the original TRPCA problem into a standard Tensor Principal Component Analysis (TPCA) problem. The proposed model involves an optimization-driven weighting scheme that dynamically identifies and downweights outlier contributions during tensor augmentation. We develop an efficient proximal block coordinate descent algorithm with closed-form updates to solve the resulting optimization problem, ensuring computational efficiency. Theoretical convergence is guaranteed through a framework combining block coordinate descent with majorization-minimization principles. Numerical experiments on synthetic and real-world datasets, including face recovery, background subtraction, and hyperspectral denoising, demonstrate that our method effectively handles various corruption patterns. The results show the improvements in both accuracy and computational efficiency compared to state-of-the-art methods.
Authors: Maodong Li, Jiying Zhang, Bin Feng, Wenqi Zeng, Dechin Chen, Zhijun Pan, Yu Li, Zijing Liu, Yi Isaac Yang
Abstract: Drug-protein binding and dissociation dynamics are fundamental to understanding molecular interactions in biological systems. While many tools for drug-protein interaction studies have emerged, especially artificial intelligence (AI)-based generative models, predictive tools on binding/dissociation kinetics and dynamics are still limited. We propose a novel research paradigm that combines molecular dynamics (MD) simulations, enhanced sampling, and AI generative models to address this issue. We propose an enhanced sampling strategy to efficiently implement the drug-protein dissociation process in MD simulations and estimate the free energy surface (FES). We constructed a program pipeline of MD simulations based on this sampling strategy, thus generating a dataset including 26,612 drug-protein dissociation trajectories containing about 13 million frames. We named this dissociation dynamics dataset DD-13M and used it to train a deep equivariant generative model UnbindingFlow, which can generate collision-free dissociation trajectories. The DD-13M database and UnbindingFlow model represent a significant advancement in computational structural biology, and we anticipate its broad applicability in machine learning studies of drug-protein interactions. Our ongoing efforts focus on expanding this methodology to encompass a broader spectrum of drug-protein complexes and exploring novel applications in pathway prediction.
Authors: Tiankai Hang, Jianmin Bao, Fangyun Wei, Dong Chen
Abstract: Autoregressive models have demonstrated remarkable success in sequential data generation, particularly in NLP, but their extension to continuous-domain image generation presents significant challenges. Recent work, the masked autoregressive model (MAR), bypasses quantization by modeling per-token distributions in continuous spaces using a diffusion head but suffers from slow inference due to the high computational cost of the iterative denoising process. To address this, we propose the Fast AutoRegressive model (FAR), a novel framework that replaces MAR's diffusion head with a lightweight shortcut head, enabling efficient few-step sampling while preserving autoregressive principles. Additionally, FAR seamlessly integrates with causal Transformers, extending them from discrete to continuous token generation without requiring architectural modifications. Experiments demonstrate that FAR achieves $2.3\times$ faster inference than MAR while maintaining competitive FID and IS scores. This work establishes the first efficient autoregressive paradigm for high-fidelity continuous-space image generation, bridging the critical gap between quality and scalability in visual autoregressive modeling.
Authors: Hongyu Wang, Shuming Ma, Furu Wei
Abstract: Efficient deployment of 1-bit Large Language Models (LLMs) is hindered by activation outliers, which complicate quantization to low bit-widths. We introduce BitNet v2, a novel framework enabling native 4-bit activation quantization for 1-bit LLMs. To tackle outliers in attention and feed-forward network activations, we propose H-BitLinear, a module applying an online Hadamard transformation prior to activation quantization. This transformation smooths sharp activation distributions into more Gaussian-like forms, suitable for low-bit representation. Experiments show BitNet v2 trained from scratch with 8-bit activations matches BitNet b1.58 performance. Crucially, BitNet v2 achieves minimal performance degradation when trained with native 4-bit activations, significantly reducing memory footprint and computational cost for batched inference.
Authors: KimiTeam, Ding Ding, Zeqian Ju, Yichong Leng, Songxiang Liu, Tong Liu, Zeyu Shang, Kai Shen, Wei Song, Xu Tan, Heyi Tang, Zhengtao Wang, Chu Wei, Yifei Xin, Xinran Xu, Jianwei Yu, Yutao Zhang, Xinyu Zhou, Y. Charles, Jun Chen, Yanru Chen, Yulun Du, Weiran He, Zhenxing Hu, Guokun Lai, Qingcheng Li, Yangyang Liu, Weidong Sun, Jianzhou Wang, Yuzhi Wang, Yuefeng Wu, Yuxin Wu, Dongchao Yang, Hao Yang, Ying Yang, Zhilin Yang, Aoxiong Yin, Ruibin Yuan, Yutong Zhang, Zaida Zhou
Abstract: We present Kimi-Audio, an open-source audio foundation model that excels in audio understanding, generation, and conversation. We detail the practices in building Kimi-Audio, including model architecture, data curation, training recipe, inference deployment, and evaluation. Specifically, we leverage a 12.5Hz audio tokenizer, design a novel LLM-based architecture with continuous features as input and discrete tokens as output, and develop a chunk-wise streaming detokenizer based on flow matching. We curate a pre-training dataset that consists of more than 13 million hours of audio data covering a wide range of modalities including speech, sound, and music, and build a pipeline to construct high-quality and diverse post-training data. Initialized from a pre-trained LLM, Kimi-Audio is continual pre-trained on both audio and text data with several carefully designed tasks, and then fine-tuned to support a diverse of audio-related tasks. Extensive evaluation shows that Kimi-Audio achieves state-of-the-art performance on a range of audio benchmarks including speech recognition, audio understanding, audio question answering, and speech conversation. We release the codes, model checkpoints, as well as the evaluation toolkits in https://github.com/MoonshotAI/Kimi-Audio.
Authors: Vinay Kanakeri, Aritra Mitra
Abstract: We consider the problem of system identification of partially observed linear time-invariant (LTI) systems. Given input-output data, we provide non-asymptotic guarantees for identifying the system parameters under general heavy-tailed noise processes. Unlike previous works that assume Gaussian or sub-Gaussian noise, we consider significantly broader noise distributions that are required to admit only up to the second moment. For this setting, we leverage tools from robust statistics to propose a novel system identification algorithm that exploits the idea of boosting. Despite the much weaker noise assumptions, we show that our proposed algorithm achieves sample complexity bounds that nearly match those derived under sub-Gaussian noise. In particular, we establish that our bounds retain a logarithmic dependence on the prescribed failure probability. Interestingly, we show that such bounds can be achieved by requiring just a finite fourth moment on the excitatory input process.
Authors: Milad Sefidgaran, Abdellatif Zaidi, Piotr Krasnowski
Abstract: We study the problem of distributed multi-view representation learning. In this problem, $K$ agents observe each one distinct, possibly statistically correlated, view and independently extracts from it a suitable representation in a manner that a decoder that gets all $K$ representations estimates correctly the hidden label. In the absence of any explicit coordination between the agents, a central question is: what should each agent extract from its view that is necessary and sufficient for a correct estimation at the decoder? In this paper, we investigate this question from a generalization error perspective. First, we establish several generalization bounds in terms of the relative entropy between the distribution of the representations extracted from training and "test" datasets and a data-dependent symmetric prior, i.e., the Minimum Description Length (MDL) of the latent variables for all views and training and test datasets. Then, we use the obtained bounds to devise a regularizer; and investigate in depth the question of the selection of a suitable prior. In particular, we show and conduct experiments that illustrate that our data-dependent Gaussian mixture priors with judiciously chosen weights lead to good performance. For single-view settings (i.e., $K=1$), our experimental results are shown to outperform existing prior art Variational Information Bottleneck (VIB) and Category-Dependent VIB (CDVIB) approaches. Interestingly, we show that a weighted attention mechanism emerges naturally in this setting. Finally, for the multi-view setting, we show that the selection of the joint prior as a Gaussians product mixture induces a Gaussian mixture marginal prior for each marginal view and implicitly encourages the agents to extract and output redundant features, a finding which is somewhat counter-intuitive.
Authors: Stefano Markidis, Jonah Ekelund, Luca Pennati, Andong Hu, Ivy Peng
Abstract: Geomagnetic storms are large-scale disturbances of the Earth's magnetosphere driven by solar wind interactions, posing significant risks to space-based and ground-based infrastructure. The Disturbance Storm Time (Dst) index quantifies geomagnetic storm intensity by measuring global magnetic field variations. This study applies symbolic regression to derive data-driven equations describing the temporal evolution of the Dst index. We use historical data from the NASA OMNIweb database, including solar wind density, bulk velocity, convective electric field, dynamic pressure, and magnetic pressure. The PySR framework, an evolutionary algorithm-based symbolic regression library, is used to identify mathematical expressions linking dDst/dt to key solar wind. The resulting models include a hierarchy of complexity levels and enable a comparison with well-established empirical models such as the Burton-McPherron-Russell and O'Brien-McPherron models. The best-performing symbolic regression models demonstrate superior accuracy in most cases, particularly during moderate geomagnetic storms, while maintaining physical interpretability. Performance evaluation on historical storm events includes the 2003 Halloween Storm, the 2015 St. Patrick's Day Storm, and a 2017 moderate storm. The results provide interpretable, closed-form expressions that capture nonlinear dependencies and thresholding effects in Dst evolution.
Authors: Giuseppe Loffredo, Elvira Romano, Fabrizio MAturo
Abstract: Functional survival models are key tools for analyzing time-to-event data with complex predictors, such as functional or high-dimensional inputs. Despite their predictive strength, these models often lack interpretability, which limits their value in practical decision-making and risk analysis. This study investigates two key survival models: the Functional Survival Tree (FST) and the Functional Random Survival Forest (FRSF). It introduces novel methods and tools to enhance the interpretability of FST models and improve the explainability of FRSF ensembles. Using both real and simulated datasets, the results demonstrate that the proposed approaches yield efficient, easy-to-understand decision trees that accurately capture the underlying decision-making processes of the model ensemble.
Authors: Zilan Cheng, Zhongjian Wang, Li-Lian Wang, Mejdi Azaiez
Abstract: In this paper, we introduce Proper Orthogonal Decomposition Neural Operators (PODNO) for solving partial differential equations (PDEs) dominated by high-frequency components. Building on the structure of Fourier Neural Operators (FNO), PODNO replaces the Fourier transform with (inverse) orthonormal transforms derived from the Proper Orthogonal Decomposition (POD) method to construct the integral kernel. Due to the optimality of POD basis, the PODNO has potential to outperform FNO in both accuracy and computational efficiency for high-frequency problems. From analysis point of view, we established the universality of a generalization of PODNO, termed as Generalized Spectral Operator (GSO). In addition, we evaluate PODNO's performance numerically on dispersive equations such as the Nonlinear Schrodinger (NLS) equation and the Kadomtsev-Petviashvili (KP) equation.
Authors: Julius von K\"ugelgen, Jakob Ketterer, Xinwei Shen, Nicolai Meinshausen, Jonas Peters
Abstract: We consider the problem of modelling the effects of unseen perturbations such as gene knockdowns or drug combinations on low-level measurements such as RNA sequencing data. Specifically, given data collected under some perturbations, we aim to predict the distribution of measurements for new perturbations. To address this challenging extrapolation task, we posit that perturbations act additively in a suitable, unknown embedding space. More precisely, we formulate the generative process underlying the observed data as a latent variable model, in which perturbations amount to mean shifts in latent space and can be combined additively. Unlike previous work, we prove that, given sufficiently diverse training perturbations, the representation and perturbation effects are identifiable up to affine transformation, and use this to characterize the class of unseen perturbations for which we obtain extrapolation guarantees. To estimate the model from data, we propose a new method, the perturbation distribution autoencoder (PDAE), which is trained by maximising the distributional similarity between true and predicted perturbation distributions. The trained model can then be used to predict previously unseen perturbation distributions. Empirical evidence suggests that PDAE compares favourably to existing methods and baselines at predicting the effects of unseen perturbations.
Authors: Joshua Engels, David D. Baek, Subhash Kantamneni, Max Tegmark
Abstract: Scalable oversight, the process by which weaker AI systems supervise stronger ones, has been proposed as a key strategy to control future superintelligent systems. However, it is still unclear how scalable oversight itself scales. To address this gap, we propose a framework that quantifies the probability of successful oversight as a function of the capabilities of the overseer and the system being overseen. Specifically, our framework models oversight as a game between capability-mismatched players; the players have oversight-specific and deception-specific Elo scores that are a piecewise-linear function of their general intelligence, with two plateaus corresponding to task incompetence and task saturation. We validate our framework with a modified version of the game Nim and then apply it to four oversight games: "Mafia", "Debate", "Backdoor Code" and "Wargames". For each game, we find scaling laws that approximate how domain performance depends on general AI system capability (using Chatbot Arena Elo as a proxy for general capability). We then build on our findings in a theoretical study of Nested Scalable Oversight (NSO), a process in which trusted models oversee untrusted stronger models, which then become the trusted models in the next step. We identify conditions under which NSO succeeds and derive numerically (and in some cases analytically) the optimal number of oversight levels to maximize the probability of oversight success. In our numerical examples, the NSO success rate is below 52% when overseeing systems that are 400 Elo points stronger than the baseline overseer, and it declines further for overseeing even stronger systems.
Authors: Gwen Yidou Weng, Benjie Wang, Guy Van den Broeck
Abstract: As large language models (LMs) advance, there is an increasing need to control their outputs to align with human values (e.g., detoxification) or desired attributes (e.g., personalization, topic). However, autoregressive models focus on next-token predictions and struggle with global properties that require looking ahead. Existing solutions either tune or post-train LMs for each new attribute - expensive and inflexible - or approximate the Expected Attribute Probability (EAP) of future sequences by sampling or training, which is slow and unreliable for rare attributes. We introduce TRACE (Tractable Probabilistic Reasoning for Adaptable Controllable gEneration), a novel framework that efficiently computes EAP and adapts to new attributes through tractable probabilistic reasoning and lightweight control. TRACE distills a Hidden Markov Model (HMM) from an LM and pairs it with a small classifier to estimate attribute probabilities, enabling exact EAP computation over the HMM's predicted futures. This EAP is then used to reweigh the LM's next-token probabilities for globally compliant continuations. Empirically, TRACE achieves state-of-the-art results in detoxification with only 10% decoding overhead, adapts to 76 low-resource personalized LLMs within seconds, and seamlessly extends to composite attributes.
Authors: Anna Katariina Wisakanto, Joe Rogero, Avyay M. Casheekar, Richard Mallah
Abstract: Modern general-purpose artificial intelligence (AI) systems present an urgent risk management challenge, as their rapidly evolving capabilities and potential for catastrophic harm outpace our ability to reliably assess their risks. Current methods often rely on selective testing and undocumented assumptions about risk priorities, frequently failing to make a serious attempt at assessing the set of pathways through which Al systems pose direct or indirect risks to society and the biosphere. This paper introduces the probabilistic risk assessment (PRA) for AI framework, adapting established PRA techniques from high-reliability industries (e.g., nuclear power, aerospace) for the new challenges of advanced AI. The framework guides assessors in identifying potential risks, estimating likelihood and severity, and explicitly documenting evidence, underlying assumptions, and analyses at appropriate granularities. The framework's implementation tool synthesizes the results into a risk report card with aggregated risk estimates from all assessed risks. This systematic approach integrates three advances: (1) Aspect-oriented hazard analysis provides systematic hazard coverage guided by a first-principles taxonomy of AI system aspects (e.g. capabilities, domain knowledge, affordances); (2) Risk pathway modeling analyzes causal chains from system aspects to societal impacts using bidirectional analysis and incorporating prospective techniques; and (3) Uncertainty management employs scenario decomposition, reference scales, and explicit tracing protocols to structure credible projections with novelty or limited data. Additionally, the framework harmonizes diverse assessment methods by integrating evidence into comparable, quantified absolute risk estimates for critical decisions. We have implemented this as a workbook tool for AI developers, evaluators, and regulators, available on the project website.
Authors: Ahmad AbdulKader, Kareem Nassar, Mohamed El-Geish, Daniel Galvez, Chetan Patil
Abstract: We propose using cascaded classifiers for a keyword spotting (KWS) task on narrow-band (NB), 8kHz audio acquired in non-IID environments -- a more challenging task than most state-of-the-art KWS systems face. We present a model that incorporates Deep Neural Networks (DNNs), cascading, multiple-feature representations, and multiple-instance learning. The cascaded classifiers handle the task's class imbalance and reduce power consumption on computationally-constrained devices via early termination. The KWS system achieves a false negative rate of 6% at an hourly false positive rate of 0.75
Authors: Ce Ju, Cuntai Guan
Abstract: Recent progress in geometric deep learning has drawn increasing attention from the machine learning community toward domain adaptation on symmetric positive definite (SPD) manifolds, especially for neuroimaging data that often suffer from distribution shifts across sessions. These data, typically represented as covariance matrices of brain signals, inherently lie on SPD manifolds due to their symmetry and positive definiteness. However, conventional domain adaptation methods often overlook this geometric structure when applied directly to covariance matrices, which can result in suboptimal performance. To address this issue, we introduce a new geometric deep learning framework that combines optimal transport theory with the geometry of SPD manifolds. Our approach aligns data distributions while respecting the manifold structure, effectively reducing both marginal and conditional discrepancies. We validate our method on three cross-session brain computer interface datasets, KU, BNCI2014001, and BNCI2015001, where it consistently outperforms baseline approaches while maintaining the intrinsic geometry of the data. We also provide quantitative results and visualizations to better illustrate the behavior of the learned embeddings.
Authors: Xuan Rao, Bo Zhao, Derong Liu
Abstract: In neural architecture search (NAS) methods based on latent space optimization (LSO), a deep generative model is trained to embed discrete neural architectures into a continuous latent space. In this case, different optimization algorithms that operate in the continuous space can be implemented to search neural architectures. However, the optimization of latent variables is challenging for gradient-based LSO since the mapping from the latent space to the architecture performance is generally non-convex. To tackle this problem, this paper develops a convexity regularized latent space optimization (CR-LSO) method, which aims to regularize the learning process of latent space in order to obtain a convex architecture performance mapping. Specifically, CR-LSO trains a graph variational autoencoder (G-VAE) to learn the continuous representations of discrete architectures. Simultaneously, the learning process of latent space is regularized by the guaranteed convexity of input convex neural networks (ICNNs). In this way, the G-VAE is forced to learn a convex mapping from the architecture representation to the architecture performance. Hereafter, the CR-LSO approximates the performance mapping using the ICNN and leverages the estimated gradient to optimize neural architecture representations. Experimental results on three popular NAS benchmarks show that CR-LSO achieves competitive evaluation results in terms of both computational complexity and architecture performance.
Authors: Milin Zhang, Mohammad Abdi, Jonathan Ashdown, Francesco Restuccia
Abstract: Distributed deep neural networks (DNNs) have been shown to reduce the computational burden of mobile devices and decrease the end-to-end inference latency in edge computing scenarios. While distributed DNNs have been studied, to the best of our knowledge, the resilience of distributed DNNs to adversarial action remains an open problem. In this paper, we fill the existing research gap by rigorously analyzing the robustness of distributed DNNs against adversarial action. We cast this problem in the context of information theory and rigorously proved that (i) the compressed latent dimension improves the robustness but also affect task-oriented performance; and (ii) the deeper splitting point enhances the robustness but also increases the computational burden. These two trade-offs provide a novel perspective to design robust distributed DNN. To test our theoretical findings, we perform extensive experimental analysis by considering 6 different DNN architectures, 6 different approaches for distributed DNN and 10 different adversarial attacks using the ImageNet-1K dataset.
Authors: Borja Aizpurua, Samuel Palmer, Roman Orus
Abstract: In this paper we show how tensor networks help in developing explainability of machine learning algorithms. Specifically, we develop an unsupervised clustering algorithm based on Matrix Product States (MPS) and apply it in the context of a real use-case of adversary-generated threat intelligence. Our investigation proves that MPS rival traditional deep learning models such as autoencoders and GANs in terms of performance, while providing much richer model interpretability. Our approach naturally facilitates the extraction of feature-wise probabilities, Von Neumann Entropy, and mutual information, offering a compelling narrative for classification of anomalies and fostering an unprecedented level of transparency and interpretability, something fundamental to understand the rationale behind artificial intelligence decisions.
Authors: Ossi R\"ais\"a, Antti Honkela
Abstract: Recent studies have highlighted the benefits of generating multiple synthetic datasets for supervised learning, from increased accuracy to more effective model selection and uncertainty estimation. These benefits have clear empirical support, but the theoretical understanding of them is currently very light. We seek to increase the theoretical understanding by deriving bias-variance decompositions for several settings of using multiple synthetic datasets, including differentially private synthetic data. Our theory yields a simple rule of thumb to select the appropriate number of synthetic datasets in the case of mean-squared error and Brier score. We investigate how our theory works in practice with several real datasets, downstream predictors and error metrics. As our theory predicts, multiple synthetic datasets often improve accuracy, while a single large synthetic dataset gives at best minimal improvement, showing that our insights are practically relevant.
Authors: Bram De Cooman, Johan Suykens
Abstract: Model-free reinforcement learning methods lack an inherent mechanism to impose behavioural constraints on the trained policies. Although certain extensions exist, they remain limited to specific types of constraints, such as value constraints with additional reward signals or visitation density constraints. In this work we unify these existing techniques and bridge the gap with classical optimization and control theory, using a generic primal-dual framework for value-based and actor-critic reinforcement learning methods. The obtained dual formulations turn out to be especially useful for imposing additional constraints on the learned policy, as an intrinsic relationship between such dual constraints (or regularization terms) and reward modifications in the primal is revealed. Furthermore, using this framework, we are able to introduce some novel types of constraints, allowing to impose bounds on the policy's action density or on costs associated with transitions between consecutive states and actions. From the adjusted primal-dual optimization problems, a practical algorithm is derived that supports various combinations of policy constraints that are automatically handled throughout training using trainable reward modifications. The proposed $\texttt{DualCRL}$ method is examined in more detail and evaluated under different (combinations of) constraints on two interpretable environments. The results highlight the efficacy of the method, which ultimately provides the designer of such systems with a versatile toolbox of possible policy constraints.
Authors: Ricardo Vinuesa, Paola Cinnella, Jean Rabault, Hossein Azizpour, Stefan Bauer, Bingni W. Brunton, Arne Elofsson, Elias Jarlebring, Hedvig Kjellstrom, Stefano Markidis, David Marlevi, Javier Garcia-Martinez, Steven L. Brunton
Abstract: As modern scientific instruments generate vast amounts of data and the volume of information in the scientific literature continues to grow, machine learning (ML) has become an essential tool for organising, analysing, and interpreting these complex datasets. This paper explores the transformative role of ML in accelerating breakthroughs across a range of scientific disciplines. By presenting key examples -- such as brain mapping and exoplanet detection -- we demonstrate how ML is reshaping scientific research. We also explore different scenarios where different levels of knowledge of the underlying phenomenon are available, identifying strategies to overcome limitations and unlock the full potential of ML. Despite its advances, the growing reliance on ML poses challenges for research applications and rigorous validation of discoveries. We argue that even with these challenges, ML is poised to disrupt traditional methodologies and advance the boundaries of knowledge by enabling researchers to tackle increasingly complex problems. Thus, the scientific community can move beyond the necessary traditional oversimplifications to embrace the full complexity of natural systems, ultimately paving the way for interdisciplinary breakthroughs and innovative solutions to humanity's most pressing challenges.
Authors: Mingqi Yuan, Roger Creus Castanyer, Bo Li, Xin Jin, Wenjun Zeng, Glen Berseth
Abstract: Extrinsic rewards can effectively guide reinforcement learning (RL) agents in specific tasks. However, extrinsic rewards frequently fall short in complex environments due to the significant human effort needed for their design and annotation. This limitation underscores the necessity for intrinsic rewards, which offer auxiliary and dense signals and can enable agents to learn in an unsupervised manner. Although various intrinsic reward formulations have been proposed, their implementation and optimization details are insufficiently explored and lack standardization, thereby hindering research progress. To address this gap, we introduce RLeXplore, a unified, highly modularized, and plug-and-play framework offering reliable implementations of eight state-of-the-art intrinsic reward methods. Furthermore, we conduct an in-depth study that identifies critical implementation details and establishes well-justified standard practices in intrinsically-motivated RL. Our documentation, examples, and source code are available at https://github.com/RLE-Foundation/RLeXplore.
Authors: Vignesh Kothapalli, Tom Tirer
Abstract: A vast amount of literature has recently focused on the "Neural Collapse" (NC) phenomenon, which emerges when training neural network (NN) classifiers beyond the zero training error point. The core component of NC is the decrease in the within-class variability of the network's deepest features, dubbed as NC1. The theoretical works that study NC are typically based on simplified unconstrained features models (UFMs) that mask any effect of the data on the extent of collapse. To address this limitation of UFMs, this paper explores the possibility of analyzing NC1 using kernels associated with shallow NNs. We begin by formulating an NC1 metric as a function of the kernel. Then, we specialize it to the NN Gaussian Process kernel (NNGP) and the Neural Tangent Kernel (NTK), associated with wide networks at initialization and during gradient-based training with a small learning rate, respectively. As a key result, we show that the NTK does not represent more collapsed features than the NNGP for Gaussian data of arbitrary dimensions. This showcases the limitations of data-independent kernels such as NTK in approximating the NC behavior of NNs. As an alternative to NTK, we then empirically explore a recently proposed data-aware Gaussian Process kernel, which generalizes NNGP to model feature learning. We show that this kernel yields lower NC1 than NNGP but may not follow the trends of the shallow NN. Our study demonstrates that adaptivity to data may allow kernel-based analysis of NC, though further advancements in this area are still needed. A nice byproduct of our study is showing both theoretically and empirically that the choice of nonlinear activation function affects NC1 (with ERF yielding lower values than ReLU). The code is available at: https://github.com/kvignesh1420/shallow_nc1
Authors: Lu Li, Tianyu Zhang, Zhiqi Bu, Suyuchen Wang, Huan He, Jie Fu, Yonghui Wu, Jiang Bian, Yong Chen, Yoshua Bengio
Abstract: Model merging has emerged as an effective approach to combine multiple single-task models into a multitask model. This process typically involves computing a weighted average of the model parameters without any additional training. Existing model-merging methods focus on enhancing average task accuracy. However, interference and conflicts between the objectives of different tasks can lead to trade-offs during the merging process. In real-world applications, a set of solutions with various trade-offs can be more informative, helping practitioners make decisions based on diverse preferences. In this paper, we introduce a novel and low-compute algorithm, Model Merging with Amortized Pareto Front (MAP). MAP efficiently identifies a Pareto set of scaling coefficients for merging multiple models, reflecting the trade-offs involved. It amortizes the substantial computational cost of evaluations needed to estimate the Pareto front by using quadratic approximation surrogate models derived from a pre-selected set of scaling coefficients. Experimental results on vision and natural language processing tasks demonstrate that MAP can accurately identify the Pareto front, providing practitioners with flexible solutions to balance competing task objectives. We also introduce Bayesian MAP for scenarios with a relatively low number of tasks and Nested MAP for situations with a high number of tasks, further reducing the computational cost of evaluation.
Authors: Chayne Thrash, Ali Abbasi, Reed Andreas, Parsa Nooralinejad, Soroush Abbasi Koohpayegani, Hamed Pirsiavash, Soheil Kolouri
Abstract: The outstanding performance of large foundational models across diverse tasks, from computer vision to speech and natural language processing, has significantly increased their demand. However, storing and transmitting these models poses significant challenges due to their massive size (e.g., 750GB for Llama 3.1 405B). Recent literature has focused on compressing the original weights or reducing the number of parameters required for fine-tuning these models. These compression methods generally constrain the parameter space, for example, through low-rank reparametrization (e.g., LoRA), pruning, or quantization (e.g., QLoRA) during or after the model training. In this paper, we present a novel model compression method, which we term Manifold-Constrained Neural Compression (MCNC). This method constrains the parameter space to low-dimensional pre-defined and frozen nonlinear manifolds, which effectively cover this space. Given the prevalence of good solutions in over-parameterized deep neural networks, we show that by constraining the parameter space to our proposed manifold, we can identify high-quality solutions while achieving unprecedented compression rates across a wide variety of tasks and architectures. Through extensive experiments in computer vision and natural language processing tasks, we demonstrate that our method significantly outperforms state-of-the-art baselines in terms of compression, accuracy, and/or model reconstruction time. Our code is publicly available at https://github.com/mint-vu/MCNC.
Authors: Lecheng Kong, Jiarui Feng, Hao Liu, Chengsong Huang, Jiaxin Huang, Yixin Chen, Muhan Zhang
Abstract: Foundation models, such as Large Language Models (LLMs) or Large Vision Models (LVMs), have emerged as one of the most powerful tools in the respective fields. However, unlike text and image data, graph data do not have a definitive structure, posing great challenges to developing a Graph Foundation Model (GFM). For example, current attempts at designing general graph models either transform graph data into a language format for LLM-based prediction or still train a GNN model with LLM as an assistant. The former can handle unlimited tasks, while the latter captures graph structure much better -- yet, no existing work can achieve both simultaneously. In this paper, we identify three key desirable properties of a GFM: self-supervised pretraining, fluidity in tasks, and graph awareness. To account for these properties, we extend the conventional language modeling to the graph domain and propose a novel generative graph language model GOFA to solve the problem. The model interleaves randomly initialized GNN layers into a frozen pre-trained LLM so that the semantic and structural modeling abilities are organically combined. GOFA is pre-trained on newly proposed graph-level next-word prediction, question-answering, and structural tasks to obtain the above GFM properties. The pre-trained model is further fine-tuned on downstream tasks to obtain task-solving ability. The fine-tuned model is evaluated on various downstream tasks, demonstrating a strong ability to solve structural and contextual problems in zero-shot scenarios. The code is available at https://github.com/JiaruiFeng/GOFA.
Authors: Bella Finkel, Jose Israel Rodriguez, Chenxi Wu, Thomas Yahl
Abstract: We study the expressive power of deep polynomial neural networks through the geometry of their neurovariety. We introduce the notion of the activation degree threshold of a network architecture to express when the dimension of the neurovariety achieves its theoretical maximum. We prove the existence of the activation degree threshold for all polynomial neural networks without width-one bottlenecks and demonstrate a universal upper bound that is quadratic in the width of largest size. In doing so, we prove the high activation degree conjecture of Kileel, Trager, and Bruna. Certain structured architectures have exceptional activation degree thresholds, making them especially expressive in the sense of their neurovariety dimension. In this direction, we prove that polynomial neural networks with equi-width architectures are maximally expressive by showing their activation degree threshold is one.
Authors: Sungmin Cha, Sungjun Cho, Dasol Hwang, Moontae Lee
Abstract: Large Language Models (LLMs) have demonstrated strong reasoning and memorization capabilities via pretraining on massive textual corpora. However, this poses risk of privacy and copyright violations, highlighting the need for efficient machine unlearning methods that remove sensitive data without retraining from scratch. While Gradient Ascent (GA) is commonly used to unlearn by reducing the likelihood of generating unwanted content, it leads to unstable optimization and catastrophic forgetting of retrained knowledge. We find that combining GA with low-rank adaptation results in poor trade-offs between computational cost and generative performance. To address these challenges, we propose Low-rank Knowledge Unlearning (LoKU), a novel framework that enables robust and efficient unlearning for LLMs. First, we introduce Inverted Hinge Loss, which suppresses unwanted tokens while maintaining fluency by boosting the probability of the next most likely token. Second, we develop a data-adaptive initialization for LoRA adapters via low-rank approximation weighted with relative Fisher information, thereby focusing updates on parameters critical for removing targeted knowledge. Experiments on the Training Data Extraction Challenge dataset using GPT-Neo models as well as on the TOFU benchmark with Phi-1.5B and Llama2-7B models demonstrate that our approach effectively removes sensitive information while maintaining reasoning and generative capabilities with minimal impact. Our implementation can be found in https://github.com/csm9493/efficient-llm-unlearning.
Authors: Christopher Subich
Abstract: This work describes a process for efficiently fine-tuning the GraphCast data-driven forecast model to simulate another analysis system, here the Global Deterministic Prediction System (GDPS) of Environment and Climate Change Canada (ECCC). Using two years of training data (July 2019 -- December 2021) and 37 GPU-days of computation to tune the 37-level, quarter-degree version of GraphCast, the resulting model significantly outperforms both the unmodified GraphCast and operational forecast, showing significant forecast skill in the troposphere over lead times from 1 to 10 days. This fine-tuning is accomplished through abbreviating DeepMind's original training curriculum for GraphCast, relying on a shorter single-step forecast stage to accomplish the bulk of the adaptation work and consolidating the autoregressive stages into separate 12hr, 1d, 2d, and 3d stages with larger learning rates. Additionally, training over 3d forecasts is split into two sub-steps to conserve host memory while maintaining a strong correlation with training over the full period.
Authors: Omar Erak, Omar Alhussein, Wen Tong
Abstract: Task-oriented semantic communication systems have emerged as a promising approach to achieving efficient and intelligent data transmission, where only information relevant to a specific task is communicated. However, existing methods struggle to fully disentangle task-relevant and task-irrelevant information, leading to privacy concerns and subpar performance. To address this, we propose an information-bottleneck method, named CLAD (contrastive learning and adversarial disentanglement). CLAD utilizes contrastive learning to effectively capture task-relevant features while employing adversarial disentanglement to discard task-irrelevant information. Additionally, due to the lack of reliable and reproducible methods to gain insight into the informativeness and minimality of the encoded feature vectors, we introduce a new technique to compute the information retention index (IRI), a comparative metric used as a proxy for the mutual information between the encoded features and the input, reflecting the minimality of the encoded features. The IRI quantifies the minimality and informativeness of the encoded feature vectors across different task-oriented communication techniques. Our extensive experiments demonstrate that CLAD outperforms state-of-the-art baselines in terms of semantic extraction, task performance, privacy preservation, and IRI. CLAD achieves a predictive performance improvement of around 2.5-3%, along with a 77-90% reduction in IRI and a 57-76% decrease in adversarial attribute inference attack accuracy.
Authors: Youngjoon Lee, Jinu Gong, Joonhyuk Kang
Abstract: Federated learning enables edge devices to collaboratively train a global model while maintaining data privacy by keeping data localized. However, the Non-IID nature of data distribution across devices often hinders model convergence and reduces performance. In this paper, we propose a novel plugin for federated optimization techniques that approximates Non-IID data distributions to IID through generative AI-enhanced data augmentation and balanced sampling strategy. Key idea is to synthesize additional data for underrepresented classes on each edge device, leveraging generative AI to create a more balanced dataset across the FL network. Additionally, a balanced sampling approach at the central server selectively includes only the most IID-like devices, accelerating convergence while maximizing the global model's performance. Experimental results validate that our approach significantly improves convergence speed and robustness against data imbalance, establishing a flexible, privacy-preserving FL plugin that is applicable even in data-scarce environments.
Authors: Shi Dong, Xiaobei Niu, Rui Zhong, Zhifeng Wang, Mingzhang Zuo
Abstract: Accurate annotation of educational resources is crucial for effective personalized learning and resource recommendation in online education. However, fine-grained knowledge labels often overlap or share similarities, making it difficult for existing multi-label classification methods to differentiate them. The label distribution imbalance due to sparsity of human annotations further intensifies these challenges. To address these issues, this paper introduces RR2QC, a novel Retrieval Reranking method to multi-label Question Classification by leveraging label semantics and meta-label refinement. First, RR2QC improves the pre-training strategy by utilizing semantic relationships within and across label groups. Second, it introduces a class center learning task to align questions with label semantics during downstream training. Finally, this method decomposes labels into meta-labels and uses a meta-label classifier to rerank the retrieved label sequences. In doing so, RR2QC enhances the understanding and prediction capability of long-tail labels by learning from meta-labels that frequently appear in other labels. Additionally, a mathematical LLM is used to generate solutions for questions, extracting latent information to further refine the model's insights. Experimental results show that RR2QC outperforms existing methods in Precision@K and F1 scores across multiple educational datasets, demonstrating its effectiveness for online education applications. The code and datasets are available at https://github.com/78Erii/RR2QC.
Authors: Haoxin Liu, Chenghao Liu, B. Aditya Prakash
Abstract: Large language models (LLMs), with demonstrated reasoning abilities across multiple domains, are largely underexplored for time-series reasoning (TsR), which is ubiquitous in the real world. In this work, we propose TimerBed, the first comprehensive testbed for evaluating LLMs' TsR performance. Specifically, TimerBed includes stratified reasoning patterns with real-world tasks, comprehensive combinations of LLMs and reasoning strategies, and various supervised models as comparison anchors. We perform extensive experiments with TimerBed, test multiple current beliefs, and verify the initial failures of LLMs in TsR, evidenced by the ineffectiveness of zero shot (ZST) and performance degradation of few shot in-context learning (ICL). Further, we identify one possible root cause: the numerical modeling of data. To address this, we propose a prompt-based solution VL-Time, using visualization-modeled data and language-guided reasoning. Experimental results demonstrate that Vl-Time enables multimodal LLMs to be non-trivial ZST and powerful ICL reasoners for time series, achieving about 140% average performance improvement and 99% average token costs reduction.
Authors: Milin Zhang, Mohammad Abdi, Venkat R. Dasari, Francesco Restuccia
Abstract: Semantic Edge Computing (SEC) and Semantic Communications (SemComs) have been proposed as viable approaches to achieve real-time edge-enabled intelligence in sixth-generation (6G) wireless networks. On one hand, SemCom leverages the strength of Deep Neural Networks (DNNs) to encode and communicate the semantic information only, while making it robust to channel distortions by compensating for wireless effects. Ultimately, this leads to an improvement in the communication efficiency. On the other hand, SEC has leveraged distributed DNNs to divide the computation of a DNN across different devices based on their computational and networking constraints. Although significant progress has been made in both fields, the literature lacks a systematic view to connect both fields. In this work, we fulfill the current gap by unifying the SEC and SemCom fields. We summarize the research problems in these two fields and provide a comprehensive review of the state of the art with a focus on their technical strengths and challenges.
Authors: Cevahir Koprulu, Po-han Li, Tianyu Qiu, Ruihan Zhao, Tyler Westenbroek, David Fridovich-Keil, Sandeep Chinchali, Ufuk Topcu
Abstract: Many continuous control problems can be formulated as sparse-reward reinforcement learning (RL) tasks. In principle, online RL methods can automatically explore the state space to solve each new task. However, discovering sequences of actions that lead to a non-zero reward becomes exponentially more difficult as the task horizon increases. Manually shaping rewards can accelerate learning for a fixed task, but it is an arduous process that must be repeated for each new environment. We introduce a systematic reward-shaping framework that distills the information contained in 1) a task-agnostic prior data set and 2) a small number of task-specific expert demonstrations, and then uses these priors to synthesize dense dynamics-aware rewards for the given task. This supervision substantially accelerates learning in our experiments, and we provide analysis demonstrating how the approach can effectively guide online learning agents to faraway goals.
Authors: Vignesh Tirukkonda, Anirudh Rayas, Gautam Dasarathy
Abstract: Gaussian graphical model selection is an important paradigm with numerous applications, including biological network modeling, financial network modeling, and social network analysis. Traditional approaches assume access to independent and identically distributed (i.i.d) samples, which is often impractical in real-world scenarios. In this paper, we address Gaussian graphical model selection under observations from a more realistic dependent stochastic process known as Glauber dynamics. Glauber dynamics, also called the Gibbs sampler, is a Markov chain that sequentially updates the variables of the underlying model based on the statistics of the remaining model. Such models, aside from frequently being employed to generate samples from complex multivariate distributions, naturally arise in various settings, such as opinion consensus in social networks and clearing/stock-price dynamics in financial networks. In contrast to the extensive body of existing work, we present the first algorithm for Gaussian graphical model selection when data are sampled according to the Glauber dynamics. We provide theoretical guarantees on the computational and statistical complexity of the proposed algorithm's structure learning performance. Additionally, we provide information-theoretic lower bounds on the statistical complexity and show that our algorithm is nearly minimax optimal for a broad class of problems.
Authors: Witold Wydma\'nski, Ulvi Movsum-zada, Jacek Tabor, Marek \'Smieja
Abstract: Although deep learning models have had great success in natural language processing and computer vision, we do not observe comparable improvements in the case of tabular data, which is still the most common data type used in biological, industrial and financial applications. In particular, it is challenging to transfer large-scale pre-trained models to downstream tasks defined on small tabular datasets. To address this, we propose VisTabNet -- a cross-modal transfer learning method, which allows for adapting Vision Transformer (ViT) with pre-trained weights to process tabular data. By projecting tabular inputs to patch embeddings acceptable by ViT, we can directly apply a pre-trained Transformer Encoder to tabular inputs. This approach eliminates the conceptual cost of designing a suitable architecture for processing tabular data, while reducing the computational cost of training the model from scratch. Experimental results on multiple small tabular datasets (less than 1k samples) demonstrate VisTabNet's superiority, outperforming both traditional ensemble methods and recent deep learning models. The proposed method goes beyond conventional transfer learning practice and shows that pre-trained image models can be transferred to solve tabular problems, extending the boundaries of transfer learning. We share our example implementation as a GitHub repository available at https://github.com/wwydmanski/VisTabNet.
Authors: Hy Nguyen, Nguyen Hung Nguyen, Nguyen Linh Bao Nguyen, Srikanth Thudumu, Hung Du, Rajesh Vasa, Kon Mouzakis
Abstract: The Hierarchical Navigable Small World (HNSW) algorithm is widely used for approximate nearest neighbor (ANN) search, leveraging the principles of navigable small-world graphs. However, it faces some limitations. The first is the local optima problem, which arises from the algorithm's greedy search strategy, selecting neighbors based solely on proximity at each step. This often leads to cluster disconnections. The second limitation is that HNSW frequently fails to achieve logarithmic complexity, particularly in high-dimensional datasets, due to the exhaustive traversal through each layer. To address these limitations, we propose a novel algorithm that mitigates local optima and cluster disconnections while enhancing the construction speed, maintaining inference speed. The first component is a dual-branch HNSW structure with LID-based insertion mechanisms, enabling traversal from multiple directions. This improves outlier node capture, enhances cluster connectivity, accelerates construction speed and reduces the risk of local minima. The second component incorporates a bridge-building technique that bypasses redundant intermediate layers, maintaining inference and making up the additional computational overhead introduced by the dual-branch structure. Experiments on various benchmarks and datasets showed that our algorithm outperforms the original HNSW in both accuracy and speed. We evaluated six datasets across Computer Vision (CV), and Natural Language Processing (NLP), showing recall improvements of 18\% in NLP, and up to 30\% in CV tasks while reducing the construction time by up to 20\% and maintaining the inference speed. We did not observe any trade-offs in our algorithm. Ablation studies revealed that LID-based insertion had the greatest impact on performance, followed by the dual-branch structure and bridge-building components.
Authors: Hy Nguyen, Duy Khoa Pham, Srikanth Thudumu, Hung Du, Rajesh Vasa, Kon Mouzakis
Abstract: The widespread use of Multi-layer perceptrons (MLPs) often relies on a fixed activation function (e.g., ReLU, Sigmoid, Tanh) for all nodes within the hidden layers. While effective in many scenarios, this uniformity may limit the networks ability to capture complex data patterns. We argue that employing the same activation function at every node is suboptimal and propose leveraging different activation functions at each node to increase flexibility and adaptability. To achieve this, we introduce Local Control Networks (LCNs), which leverage B-spline functions to enable distinct activation curves at each node. Our mathematical analysis demonstrates the properties and benefits of LCNs over conventional MLPs. In addition, we demonstrate that more complex architectures, such as Kolmogorov-Arnold Networks (KANs), are unnecessary in certain scenarios, and LCNs can be a more efficient alternative. Empirical experiments on various benchmarks and datasets validate our theoretical findings. In computer vision tasks, LCNs achieve marginal improvements over MLPs and outperform KANs by approximately 5\%, while also being more computationally efficient than KANs. In basic machine learning tasks, LCNs show a 1\% improvement over MLPs and a 0.6\% improvement over KANs. For symbolic formula representation tasks, LCNs perform on par with KANs, with both architectures outperforming MLPs. Our findings suggest that diverse activations at the node level can lead to improved performance and efficiency.
Authors: Yufei He, Yuan Sui, Xiaoxin He, Yue Liu, Yifei Sun, Bryan Hooi
Abstract: Existing foundation models, such as CLIP, aim to learn a unified embedding space for multimodal data, enabling a wide range of downstream web-based applications like search, recommendation, and content classification. However, these models often overlook the inherent graph structures in multimodal datasets, where entities and their relationships are crucial. Multimodal graphs (MMGs) represent such graphs where each node is associated with features from different modalities, while the edges capture the relationships between these entities. On the other hand, existing graph foundation models primarily focus on text-attributed graphs (TAGs) and are not designed to handle the complexities of MMGs. To address these limitations, we propose UniGraph2, a novel cross-domain graph foundation model that enables general representation learning on MMGs, providing a unified embedding space. UniGraph2 employs modality-specific encoders alongside a graph neural network (GNN) to learn a unified low-dimensional embedding space that captures both the multimodal information and the underlying graph structure. We propose a new cross-domain multi-graph pre-training algorithm at scale to ensure effective transfer learning across diverse graph domains and modalities. Additionally, we adopt a Mixture of Experts (MoE) component to align features from different domains and modalities, ensuring coherent and robust embeddings that unify the information across modalities. Extensive experiments on a variety of multimodal graph tasks demonstrate that UniGraph2 significantly outperforms state-of-the-art models in tasks such as representation learning, transfer learning, and multimodal generative tasks, offering a scalable and flexible solution for learning on MMGs.
Authors: Siqi Zeng, Yifei He, Weiqiu You, Yifan Hao, Yao-Hung Hubert Tsai, Makoto Yamada, Han Zhao
Abstract: Task vectors, which are derived from the difference between pre-trained and fine-tuned model weights, enable flexible task adaptation and model merging through arithmetic operations such as addition and negation. However, existing approaches often rely on heuristics with limited theoretical support, often leading to performance gaps comparing to direct task fine tuning. Meanwhile, although it is easy to manipulate saved task vectors with arithmetic for different purposes, such compositional flexibility demands high memory usage, especially when dealing with a huge number of tasks, limiting scalability. This work addresses these issues with a theoretically grounded framework that explains task vector arithmetic and introduces the task vector bases framework. Building upon existing task arithmetic literature, our method significantly reduces the memory cost for downstream arithmetic with little effort, while achieving competitive performance and maintaining compositional advantage, providing a practical solution for large-scale task arithmetic. The code is available at https://github.com/uiuctml/TaskVectorBasis.
Authors: Peng Yan, Ahmed Abdulkadir, Gerrit A. Schatte, Giulia Aguzzi, Joonsu Gha, Nikola Pascher, Matthias Rosenthal, Yunlong Gao, Benjamin F. Grewe, Thilo Stadelmann
Abstract: To go from (passive) process monitoring to active process control, an effective AI system must learn about the behavior of the complex system from very limited training data, forming an ad-hoc digital twin with respect to process inputs and outputs that captures the consequences of actions on the process's world. We propose a novel methodology based on learning world models that disentangles process parameters in the learned latent representation, allowing for fine-grained control. Representation learning is driven by the latent factors influencing the processes through contrastive learning within a joint embedding predictive architecture. This makes changes in representations predictable from changes in inputs and vice versa, facilitating interpretability of key factors responsible for process variations, paving the way for effective control actions to keep the process within operational bounds. The effectiveness of our method is validated on the example of plastic injection molding, demonstrating practical relevance in proposing specific control actions for a notoriously unstable process.
Authors: Wuzhou Sun, Siyi Li, Qingxiang Zou, Zixing Liao
Abstract: In some game scenarios, due to the uncertainty of the number of enemy units and the priority of various attributes, the evaluation of the threat level of enemy units as well as the screening has been a challenging research topic, and the core difficulty lies in how to reasonably set the priority of different attributes in order to achieve quantitative evaluation of the threat. In this paper, we innovatively transform the problem of threat assessment into a reinforcement learning problem, and through systematic reinforcement learning training, we successfully construct an efficient neural network evaluator. The evaluator can not only comprehensively integrate the multidimensional attribute features of the enemy, but also effectively combine our state information, thus realizing a more accurate and scientific threat assessment.
Authors: Zhiyuan Ning, Zaitian Wang, Ran Zhang, Ping Xu, Kunpeng Liu, Pengyang Wang, Wei Ju, Pengfei Wang, Yuanchun Zhou, Erik Cambria, Chong Chen
Abstract: Graph clustering aims to divide the graph into different clusters. The recently emerging deep graph clustering approaches are largely built on graph neural networks (GNN). However, GNN is designed for general graph encoding and there is a common issue of representation collapse in existing GNN-based deep graph clustering algorithms. We attribute two main reasons for such issues: (i) the inductive bias of GNN models: GNNs tend to generate similar representations for proximal nodes. Since graphs often contain a non-negligible amount of inter-cluster links, the bias results in error message passing and leads to biased clustering; (ii) the clustering guided loss function: most traditional approaches strive to make all samples closer to pre-learned cluster centers, which causes a degenerate solution assigning all data points to a single label thus making all samples similar and less discriminative. To address these challenges, we investigate graph clustering from a graph cut perspective and propose an innovative and non-GNN-based Deep Cut-informed Graph embedding and Clustering framework, namely DCGC. This framework includes two modules: (i) cut-informed graph encoding; (ii) self-supervised graph clustering via optimal transport. For the encoding module, we derive a cut-informed graph embedding objective to fuse graph structure and attributes by minimizing their joint normalized cut. For the clustering module, we utilize the optimal transport theory to obtain the clustering assignments, which can balance the guidance of "proximity to the pre-learned cluster center". With the above two tailored designs, DCGC is more suitable for the graph clustering task, which can effectively alleviate the problem of representation collapse and achieve better performance. We conduct extensive experiments to demonstrate that our method is simple but effective compared with benchmarks.
Authors: Zhihao Zeng, Ziquan Fang, Yuting Huang, Lu Chen, Yunjun Gao
Abstract: Traffic prediction targets forecasting future traffic conditions using historical traffic data, serving a critical role in urban computing and transportation management. To mitigate the scarcity of traffic data while maintaining data privacy, numerous Federated Traffic Knowledge Transfer (FTT) approaches have been developed, which use transfer learning and federated learning to transfer traffic knowledge from data-rich cities to data-scarce cities, enhancing traffic prediction capabilities for the latter. However, current FTT approaches face challenges such as privacy leakage, cross-city data distribution discrepancies, low data quality, and inefficient knowledge transfer, limiting their privacy protection, effectiveness, robustness, and efficiency in real-world applications. To this end, we propose FedTT, an effective, efficient, and privacy-aware cross-city traffic knowledge transfer framework that transforms the traffic data domain from the data-rich cities and trains traffic models using the transformed data for the data-scarce cities. First, to safeguard data privacy, we propose a traffic secret transmission method that securely transmits and aggregates traffic domain-transformed data from source cities using a lightweight secret aggregation approach. Second, to mitigate the impact of traffic data distribution discrepancies on model performance, we introduce a traffic domain adapter to uniformly transform traffic data from the source cities' domains to that of the target city. Third, to improve traffic data quality, we design a traffic view imputation method to fill in and predict missing traffic data. Finally, to enhance transfer efficiency, FedTT is equipped with a federated parallel training method that enables the simultaneous training of multiple modules. Extensive experiments using 4 real-life datasets demonstrate that FedTT outperforms the 14 state-of-the-art baselines.
Authors: Wuzhou Sun, Siyi Li, Qingxiang Zou, Zixing Liao
Abstract: In various game scenarios, selecting a fixed number of targets from multiple enemy units is an extremely challenging task. This difficulty stems from the complex relationship between the threat levels of enemy units and their feature characteristics, which complicates the design of rule-based evaluators. Moreover, traditional supervised learning methods face the challenge of lacking explicit labels during training when applied to this threat evaluation problem. In this study, we redefine the threat evaluation problem as a reinforcement learning task and introduce an efficient evaluator training algorithm, Eval-PPO, based on the Proximal Policy Optimization (PPO) algorithm. Eval-PPO integrates multidimensional enemy features and the state information of friendly units through systematic training, thereby achieving precise threat assessment. Compared with rule-based methods, Eval-PPO demonstrates a significant improvement in average success rate, with an increase of 17.84%.
Authors: Hisato Komatsu
Abstract: The linear regression (LR) method offers the advantage that optimal parameters can be calculated relatively easily, although its representation capability is limited than that of the deep learning technique. To improve deep reinforcement learning, the Least Squares Deep Q Network (LS-DQN) method was proposed by Levine et al., which combines Deep Q Network (DQN) with LR method. However, the LS-DQN method assumes that the actions are discrete. In this study, we propose the Double Least Squares Deep Deterministic Policy Gradient (DLS-DDPG) method to address this limitation. This method combines the LR method with the Deep Deterministic Policy Gradient (DDPG) technique, one of the representative deep reinforcement learning algorithms for continuous action cases. For the LR update of the critic network, DLS-DDPG uses an algorithm similar to the Fitted Q iteration, the method which LS-DQN adopted. In addition, we calculated the optimal action using the quasi-Newton method and used it as both the agent's action and the training data for the LR update of the actor network. Numerical experiments conducted in MuJoCo environments showed that the proposed method improved performance at least in some tasks, although there are difficulties such as the inability to make the regularization terms small.
Authors: Jiaxiang Cheng, Guoqiang Hu
Abstract: The Cox proportional hazards (CPH) model has been widely applied in survival analysis to estimate relative risks across different subjects given multiple covariates. Traditional CPH models rely on a linear combination of covariates weighted with coefficients as the log-risk function, which imposes a strong and restrictive assumption, limiting generalization. Recent deep learning methods enable non-linear log-risk functions. However, they often lack interpretability due to the end-to-end training mechanisms. The implementation of Kolmogorov-Arnold Networks (KAN) offers new possibilities for extending the CPH model with fully transparent and symbolic non-linear log-risk functions. In this paper, we introduce Generalized Cox Proportional Hazards (GCPH) model, a novel method for survival analysis that leverages KAN to enable a non-linear mapping from covariates to survival outcomes in a fully symbolic manner. GCPH maintains the interpretability of traditional CPH models while allowing for the estimation of non-linear log-risk functions. Experiments conducted on both synthetic data and various public benchmarks demonstrate that GCPH achieves competitive performance in terms of prediction accuracy and exhibits superior interpretability compared to current state-of-the-art methods.
Authors: Tiange Huang, Yongjun Li
Abstract: Unsupervised multivariate time series anomaly detection (UMTSAD) plays a critical role in various domains, including finance, networks, and sensor systems. In recent years, due to the outstanding performance of deep learning in general sequential tasks, many models have been specialized for deep UMTSAD tasks and have achieved impressive results, particularly those based on the Transformer and self-attention mechanisms. However, the sequence anomaly association assumptions underlying these models are often limited to specific predefined patterns and scenarios, such as concentrated or peak anomaly patterns. These limitations hinder their ability to generalize to diverse anomaly situations, especially where the lack of labels poses significant challenges. To address these issues, we propose AMAD, which integrates \textbf{A}uto\textbf{M}asked Attention for UMTS\textbf{AD} scenarios. AMAD introduces a novel structure based on the AutoMask mechanism and an attention mixup module, forming a simple yet generalized anomaly association representation framework. This framework is further enhanced by a Max-Min training strategy and a Local-Global contrastive learning approach. By combining multi-scale feature extraction with automatic relative association modeling, AMAD provides a robust and adaptable solution to UMTSAD challenges. Extensive experimental results demonstrate that the proposed model achieving competitive performance results compared to SOTA benchmarks across a variety of datasets.
Authors: Sebastian Seidel, Uwe M. Borghoff
Abstract: Artificial intelligence (AI) has emerged as a transformative force across industries, driven by advances in deep learning and natural language processing, and fueled by large-scale data and computing resources. Despite its rapid adoption, the opacity of AI systems poses significant challenges to trust and acceptance. This work explores the intersection of connectionist and symbolic approaches to artificial intelligence, focusing on the derivation of interpretable symbolic models, such as decision trees, from feedforward neural networks (FNNs). Decision trees provide a transparent framework for elucidating the operations of neural networks while preserving their functionality. The derivation is presented in a step-by-step approach and illustrated with several examples. A systematic methodology is proposed to bridge neural and symbolic paradigms by exploiting distributed representations in FNNs to identify symbolic components, including fillers, roles, and their interrelationships. The process traces neuron activation values and input configurations across network layers, mapping activations and their underlying inputs to decision tree edges. The resulting symbolic structures effectively capture FNN decision processes and enable scalability to deeper networks through iterative refinement of subpaths for each hidden layer. To validate the theoretical framework, a prototype was developed using Keras .h5-data and emulating TensorFlow within the Java JDK/JavaFX environment. This prototype demonstrates the feasibility of extracting symbolic representations from neural networks, enhancing trust in AI systems, and promoting accountability.
Authors: Vinay Sharma, R\'emi Tanguy Oddon, Pietro Tesini, Jens Ravesloot, Cees Taal, Olga Fink
Abstract: Accurate real-time modeling of multi-body dynamical systems is essential for enabling digital twin applications across industries. While many data-driven approaches aim to learn system dynamics, jointly predicting internal loads and system trajectories remains a key challenge. This dual prediction is especially important for fault detection and predictive maintenance, where internal loads-such as contact forces-act as early indicators of faults, reflecting wear or misalignment before affecting motion. These forces also serve as inputs to degradation models (e.g., crack growth), enabling damage prediction and remaining useful life estimation. We propose Equi-Euler GraphNet, a physics-informed graph neural network (GNN) that simultaneously predicts internal forces and global trajectories in multi-body systems. In this mesh-free framework, nodes represent system components and edges encode interactions. Equi-Euler GraphNet introduces two inductive biases: (1) an equivariant message-passing scheme, interpreting edge messages as interaction forces consistent under Euclidean transformations; and (2) a temporal-aware iterative node update mechanism, based on Euler integration, to capture influence of distant interactions over time. Tailored for cylindrical roller bearings, it decouples ring dynamics from constrained motion of rolling elements. Trained on high-fidelity multiphysics simulations, Equi-Euler GraphNet generalizes beyond the training distribution, accurately predicting loads and trajectories under unseen speeds, loads, and configurations. It outperforms state-of-the-art GNNs focused on trajectory prediction, delivering stable rollouts over thousands of time steps with minimal error accumulation. Achieving up to a 200x speedup over conventional solvers while maintaining comparable accuracy, it serves as an efficient reduced-order model for digital twins, design, and maintenance.
Authors: Xinyu Liang, Hao Wang
Abstract: The scarcity of high-quality residential load data can pose obstacles for decarbonizing the residential sector as well as effective grid planning and operation. The above challenges have motivated research into generating synthetic load data, but existing methods faced limitations in terms of scalability, diversity, and similarity. This paper proposes a Generative Adversarial Network-based Synthetic Residential Load Pattern (RLP-GAN) generation model, a novel weakly-supervised GAN framework, leveraging an over-complete autoencoder to capture dependencies within complex and diverse load patterns and learn household-level data distribution at scale. We incorporate a model weight selection method to address the mode collapse problem and generate load patterns with high diversity. We develop a holistic evaluation method to validate the effectiveness of RLP-GAN using real-world data of 417 households. The results demonstrate that RLP-GAN outperforms state-of-the-art models in capturing temporal dependencies and generating load patterns with higher similarity to real data. Furthermore, we have publicly released the RLP-GAN generated synthetic dataset, which comprises one million synthetic residential load pattern profiles.
Authors: Jingtong Gao, Yewen Li, Shuai Mao, Peng Jiang, Nan Jiang, Yejing Wang, Qingpeng Cai, Fei Pan, Peng Jiang, Kun Gai, Bo An, Xiangyu Zhao
Abstract: Auto-bidding, with its strong capability to optimize bidding decisions within dynamic and competitive online environments, has become a pivotal strategy for advertising platforms. Existing approaches typically employ rule-based strategies or Reinforcement Learning (RL) techniques. However, rule-based strategies lack the flexibility to adapt to time-varying market conditions, and RL-based methods struggle to capture essential historical dependencies and observations within Markov Decision Process (MDP) frameworks. Furthermore, these approaches often face challenges in ensuring strategy adaptability across diverse advertising objectives. Additionally, as offline training methods are increasingly adopted to facilitate the deployment and maintenance of stable online strategies, the issues of documented behavioral patterns and behavioral collapse resulting from training on fixed offline datasets become increasingly significant. To address these limitations, this paper introduces a novel offline Generative Auto-bidding framework with Value-Guided Explorations (GAVE). GAVE accommodates various advertising objectives through a score-based Return-To-Go (RTG) module. Moreover, GAVE integrates an action exploration mechanism with an RTG-based evaluation method to explore novel actions while ensuring stability-preserving updates. A learnable value function is also designed to guide the direction of action exploration and mitigate Out-of-Distribution (OOD) problems. Experimental results on two offline datasets and real-world deployments demonstrate that GAVE outperforms state-of-the-art baselines in both offline evaluations and online A/B tests. By applying the core methods of this framework, we proudly secured first place in the NeurIPS 2024 competition, 'AIGB Track: Learning Auto-Bidding Agents with Generative Models'.
Authors: Xiang Li, Haobing Liu, Jianpeng Qi, Yuan Cao, Guoqing Chao, Yanwei Yu
Abstract: Graph Neural Networks (GNNs) have demonstrated strong performance across various graph-based tasks by effectively capturing relational information between nodes. These models rely on iterative message passing to propagate node features, enabling nodes to aggregate information from their neighbors. Recent research has significantly improved the message-passing mechanism, enhancing GNN scalability on large-scale graphs. However, GNNs still face two main challenges: over-smoothing, where excessive message passing results in indistinguishable node representations, especially in deep networks incorporating high-order neighbors; and scalability issues, as traditional architectures suffer from high model complexity and increased inference time due to redundant information aggregation. This paper proposes a novel framework for large-scale graphs named ScaleGNN that simultaneously addresses both challenges by adaptively fusing multi-level graph features. We first construct neighbor matrices for each order, learning their relative information through trainable weights through an adaptive high-order feature fusion module. This allows the model to selectively emphasize informative high-order neighbors while reducing unnecessary computational costs. Additionally, we introduce a High-order redundant feature masking mechanism based on a Local Contribution Score (LCS), which enables the model to retain only the most relevant neighbors at each order, preventing redundant information propagation. Furthermore, low-order enhanced feature aggregation adaptively integrates low-order and high-order features based on task relevance, ensuring effective capture of both local and global structural information without excessive complexity. Extensive experiments on real-world datasets demonstrate that our approach consistently outperforms state-of-the-art GNN models in both accuracy and computational efficiency.
Authors: Abdesslem Layeb
Abstract: In this paper, we introduce a novel data transformation framework based on Opposition-Based Learning (OBL) to boost the performance of traditional classification algorithms. Originally developed to accelerate convergence in optimization tasks, OBL is leveraged here to generate synthetic opposite samples that enrich the training data and improve decision boundary formation. We explore three OBL variants Global OBL, Class-Wise OBL, and Localized Class-Wise OBL and integrate them with K-Nearest Neighbors (KNN). Extensive experiments conducted on 26 heterogeneous and high-dimensional datasets demonstrate that OBL-enhanced classifiers consistently outperform the basic KNN. These findings underscore the potential of OBL as a lightweight yet powerful data transformation strategy for enhancing classification performance, especially in complex or sparse learning environments.
Authors: Yilin Zhai, Hongyuan Shi, Chao Zhan, Qing Wang, Zaijin You, Nan Wang
Abstract: Accurate wave height prediction is critical for maritime safety and coastal resilience, yet conventional physics-based models and traditional machine learning methods face challenges in computational efficiency and nonlinear dynamics modeling. This study introduces Chronos, the first implementation of a large language model (LLM)-powered temporal architecture (Chronos) optimized for wave forecasting. Through advanced temporal pattern recognition applied to historical wave data from three strategically chosen marine zones in the Northwest Pacific basin, our framework achieves multimodal improvements: (1) 14.3% reduction in training time with 2.5x faster inference speed compared to PatchTST baselines, achieving 0.575 mean absolute scaled error (MASE) units; (2) superior short-term forecasting (1-24h) across comprehensive metrics; (3) sustained predictive leadership in extended-range forecasts (1-120h); and (4) demonstrated zero-shot capability maintaining median performance (rank 4/12) against specialized operational models. This LLM-enhanced temporal modeling paradigm establishes a new standard in wave prediction, offering both computationally efficient solutions and a transferable framework for complex geophysical systems modeling.
Authors: Jun Wu, Jiangtao Wen, Yuxing Han
Abstract: The rapid advancement of large-language models (LLMs) has driven extensive research into parameter compression after training has been completed, yet compression during the training phase remains largely unexplored. In this work, we introduce Rate-Constrained Training (BackSlash), a novel training-time compression approach based on rate-distortion optimization (RDO). BackSlash enables a flexible trade-off between model accuracy and complexity, significantly reducing parameter redundancy while preserving performance. Experiments in various architectures and tasks demonstrate that BackSlash can reduce memory usage by 60% - 90% without accuracy loss and provides significant compression gain compared to compression after training. Moreover, BackSlash proves to be highly versatile: it enhances generalization with small Lagrange multipliers, improves model robustness to pruning (maintaining accuracy even at 80% pruning rates), and enables network simplification for accelerated inference on edge devices.
Authors: Xinyu Zhou, Simin Fan, Martin Jaggi, Jie Fu
Abstract: Grokking is proposed and widely studied as an intricate phenomenon in which generalization is achieved after a long-lasting period of overfitting. In this work, we propose NeuralGrok, a novel gradient-based approach that learns an optimal gradient transformation to accelerate the generalization of transformers in arithmetic tasks. Specifically, NeuralGrok trains an auxiliary module (e.g., an MLP block) in conjunction with the base model. This module dynamically modulates the influence of individual gradient components based on their contribution to generalization, guided by a bilevel optimization algorithm. Our extensive experiments demonstrate that NeuralGrok significantly accelerates generalization, particularly in challenging arithmetic tasks. We also show that NeuralGrok promotes a more stable training paradigm, constantly reducing the model's complexity, while traditional regularization methods, such as weight decay, can introduce substantial instability and impede generalization. We further investigate the intrinsic model complexity leveraging a novel Absolute Gradient Entropy (AGE) metric, which explains that NeuralGrok effectively facilitates generalization by reducing the model complexity. We offer valuable insights on the grokking phenomenon of Transformer models, which encourages a deeper understanding of the fundamental principles governing generalization ability.
Authors: Miaoyun Zhao, Qiang Zhang, Chenrong Li
Abstract: Spurious correlations that lead models to correct predictions for the wrong reasons pose a critical challenge for robust real-world generalization. Existing research attributes this issue to group imbalance and addresses it by maximizing group-balanced or worst-group accuracy, which heavily relies on expensive bias annotations. A compromise approach involves predicting bias information using extensively pretrained foundation models, which requires large-scale data and becomes impractical for resource-limited rare domains. To address these challenges, we offer a novel perspective by reframing the spurious correlations as imbalances or mismatches in class-conditional distributions, and propose a simple yet effective robust learning method that eliminates the need for both bias annotations and predictions. With the goal of reducing the mutual information between spurious factors and label information, our method leverages a sample reweighting strategy to achieve class-conditional distribution balancing, which automatically highlights minority groups and classes, effectively dismantling spurious correlations and producing a debiased data distribution for classification. Extensive experiments and analysis demonstrate that our approach consistently delivers state-of-the-art performance, rivaling methods that rely on bias supervision.
Authors: Shengtao Zhang, Haokai Zhang, Shiqi Lou, Zicheng Wang, Zinan Zeng, Yilin Wang, Minnan Luo
Abstract: Dynamic node classification is critical for modeling evolving systems like financial transactions and academic collaborations. In such systems, dynamically capturing node information changes is critical for dynamic node classification, which usually requires all labels at every timestamp. However, it is difficult to collect all dynamic labels in real-world scenarios due to high annotation costs and label uncertainty (e.g., ambiguous or delayed labels in fraud detection). In contrast, final timestamp labels are easier to obtain as they rely on complete temporal patterns and are usually maintained as a unique label for each user in many open platforms, without tracking the history data. To bridge this gap, we propose PTCL(Pseudo-label Temporal Curriculum Learning), a pioneering method addressing label-limited dynamic node classification where only final labels are available. PTCL introduces: (1) a temporal decoupling architecture separating the backbone (learning time-aware representations) and decoder (strictly aligned with final labels), which generate pseudo-labels, and (2) a Temporal Curriculum Learning strategy that prioritizes pseudo-labels closer to the final timestamp by assigning them higher weights using an exponentially decaying function. We contribute a new academic dataset (CoOAG), capturing long-range research interest in dynamic graph. Experiments across real-world scenarios demonstrate PTCL's consistent superiority over other methods adapted to this task. Beyond methodology, we propose a unified framework FLiD (Framework for Label-Limited Dynamic Node Classification), consisting of a complete preparation workflow, training pipeline, and evaluation standards, and supporting various models and datasets. The code can be found at https://github.com/3205914485/FLiD.
Authors: Max H. Farrell, Tengyuan Liang, Sanjog Misra
Abstract: This paper integrates deep neural networks (DNNs) into structural economic models to increase flexibility and capture rich heterogeneity while preserving interpretability. Economic structure and machine learning are complements in empirical modeling, not substitutes: DNNs provide the capacity to learn complex, non-linear heterogeneity patterns, while the structural model ensures the estimates remain interpretable and suitable for decision making and policy analysis. We start with a standard parametric structural model and then enrich its parameters into fully flexible functions of observables, which are estimated using a particular DNN architecture whose structure reflects the economic model. We illustrate our framework by studying demand estimation in consumer choice. We show that by enriching a standard demand model we can capture rich heterogeneity, and further, exploit this heterogeneity to create a personalized pricing strategy. This type of optimization is not possible without economic structure, but cannot be heterogeneous without machine learning. Finally, we provide theoretical justification of each step in our proposed methodology. We first establish non-asymptotic bounds and convergence rates of our structural deep learning approach. Next, a novel and quite general influence function calculation allows for feasible inference via double machine learning in a wide variety of contexts. These results may be of interest in many other contexts, as they generalize prior work.
Authors: Emile Anand, Charles Steinhardt, Martin Hansen
Abstract: Civilizations have tried to make drinking water safe to consume for thousands of years. The process of determining water contaminants has evolved with the complexity of the contaminants due to pesticides and heavy metals. The routine procedure to determine water safety is to use targeted analysis which searches for specific substances from some known list; however, we do not explicitly know which substances should be on this list. Before experimentally determining which substances are contaminants, how do we answer the sampling problem of identifying all the substances in the water? Here, we present an approach that builds on the work of Jaanus Liigand et al., which used non-targeted analysis that conducts a broader search on the sample to develop a random-forest regression model, to predict the names of all the substances in a sample, as well as their respective concentrations[1]. This work utilizes techniques from dimensionality reduction and linear decompositions to present a more accurate model using data from the European Massbank Metabolome Library to produce a global list of chemicals that researchers can then identify and test for when purifying water.
Authors: Robin Str\"asser, Manuel Schaller, Karl Worthmann, Julian Berberich, Frank Allg\"ower
Abstract: The Koopman operator serves as the theoretical backbone for machine learning of dynamical control systems, where the operator is heuristically approximated by extended dynamic mode decomposition (EDMD). In this paper, we propose SafEDMD, a novel stability- and certificate-oriented EDMD-based controller design framework. Our approach leverages a reliable surrogate model generated in a data-driven fashion in order to provide closed-loop guarantees. In particular, we establish a controller design based on semi-definite programming with guaranteed stabilization of the underlying nonlinear system. As central ingredient, we derive proportional error bounds that vanish at the origin and are tailored to control tasks. We illustrate the developed method by means of several benchmark examples and highlight the advantages over state-of-the-art methods.
Authors: Nicolas Nguyen, Imad Aouali, Andr\'as Gy\"orgy, Claire Vernade
Abstract: We study the problem of Bayesian fixed-budget best-arm identification (BAI) in structured bandits. We propose an algorithm that uses fixed allocations based on the prior information and the structure of the environment. We provide theoretical bounds on its performance across diverse models, including the first prior-dependent upper bounds for linear and hierarchical BAI. Our key contribution is introducing new proof methods that result in tighter bounds for multi-armed BAI compared to existing methods. We extensively compare our approach to other fixed-budget BAI methods, demonstrating its consistent and robust performance in various settings. Our work improves our understanding of Bayesian fixed-budget BAI in structured bandits and highlights the effectiveness of our approach in practical scenarios.
Authors: Kehan Shi, Martin Burger
Abstract: This paper studies the $p$-biharmonic equation on graphs, which arises in point cloud processing and can be interpreted as a natural extension of the graph $p$-Laplacian from the perspective of hypergraph. The asymptotic behavior of the solution is investigated when the random geometric graph is considered and the number of data points goes to infinity. We show that the continuum limit is an appropriately weighted $p$-biharmonic equation with homogeneous Neumann boundary conditions. The result relies on the uniform $L^p$ estimates for solutions and gradients of nonlocal and graph Poisson equations. The $L^\infty$ estimates of solutions are also obtained as a byproduct.
Authors: Antonin Schrab, Ilmun Kim
Abstract: We propose a general method for constructing robust permutation tests under data corruption. The proposed tests effectively control the non-asymptotic type I error under data corruption, and we prove their consistency in power under minimal conditions. This contributes to the practical deployment of hypothesis tests for real-world applications with potential adversarial attacks. For the two-sample and independence settings, we show that our kernel robust tests are minimax optimal, in the sense that they are guaranteed to be non-asymptotically powerful against alternatives uniformly separated from the null in the kernel MMD and HSIC metrics at some optimal rate (tight with matching lower bound). We point out that existing differentially private tests can be adapted to be robust to data corruption, and we demonstrate in experiments that our proposed tests achieve much higher power than these private tests. Finally, we provide publicly available implementations and empirically illustrate the practicality of our robust tests.
Authors: Ognjen Kundacina, Vladimir Vincan, Dragisa Miskovic
Abstract: This paper introduces a novel two-stage active learning (AL) pipeline for automatic speech recognition (ASR), combining unsupervised and supervised AL methods. The first stage utilizes unsupervised AL by using x-vectors clustering for diverse sample selection from unlabeled speech data, thus establishing a robust initial dataset for the subsequent supervised AL. The second stage incorporates a supervised AL strategy, with a batch AL method specifically developed for ASR, aimed at selecting diverse and informative batches of samples. Here, sample diversity is also achieved using x-vectors clustering, while the most informative samples are identified using a Bayesian AL method tailored for ASR with an adaptation of Monte Carlo dropout to approximate Bayesian inference. This approach enables precise uncertainty estimation, thereby enhancing ASR model training with significantly reduced data requirements. Our method has shown superior performance compared to competing methods on homogeneous, heterogeneous, and OOD test sets, demonstrating that strategic sample selection and innovative Bayesian modeling can substantially optimize both labeling effort and data utilization in deep learning-based ASR applications.
Authors: Kota Kondo, Claudius T. Tewari, Andrea Tagliabue, Jesus Tordesillas, Parker C. Lusk, Mason B. Peterson, Jonathan P. How
Abstract: In decentralized multiagent trajectory planners, agents need to communicate and exchange their positions to generate collision-free trajectories. However, due to localization errors/uncertainties, trajectory deconfliction can fail even if trajectories are perfectly shared between agents. To address this issue, we first present PARM and PARM*, perception-aware, decentralized, asynchronous multiagent trajectory planners that enable a team of agents to navigate uncertain environments while deconflicting trajectories and avoiding obstacles using perception information. PARM* differs from PARM as it is less conservative, using more computation to find closer-to-optimal solutions. While these methods achieve state-of-the-art performance, they suffer from high computational costs as they need to solve large optimization problems onboard, making it difficult for agents to replan at high rates. To overcome this challenge, we present our second key contribution, PRIMER, a learning-based planner trained with imitation learning (IL) using PARM* as the expert demonstrator. PRIMER leverages the low computational requirements at deployment of neural networks and achieves a computation speed up to 5500 times faster than optimization-based approaches.
Authors: Orson Mengara
Abstract: With the growing use of voice-activated systems and speech recognition technologies, the danger of backdoor attacks on audio data has grown significantly. This research looks at a specific type of attack, known as a Stochastic investment-based backdoor attack (MarketBack), in which adversaries strategically manipulate the stylistic properties of audio to fool speech recognition systems. The security and integrity of machine learning models are seriously threatened by backdoor attacks, in order to maintain the reliability of audio applications and systems, the identification of such attacks becomes crucial in the context of audio data. Experimental results demonstrated that MarketBack is feasible to achieve an average attack success rate close to 100% in seven victim models when poisoning less than 1% of the training data.
Authors: Jungeum Kim, Sean O'Hagan, Veronika Rockova
Abstract: This work is concerned with conformal prediction in contemporary applications (including generative AI) where a black-box model has been trained on data that are not accessible to the user. Mirroring split-conformal inference, we design a wrapper around a black-box algorithm which calibrates conformity scores. This calibration is local and proceeds in two stages by first adaptively partitioning the predictor space into groups and then calibrating sectionally group by group. Adaptive partitioning (self-grouping) is achieved by fitting a robust regression tree to the conformity scores on the calibration set. This new tree variant is designed in such a way that adding a single new observation does not change the tree fit with overwhelmingly large probability. This add-one-in robustness property allows us to conclude a finite sample group-conditional coverage guarantee, a refinement of the marginal guarantee. In addition, unlike traditional split-conformal inference, adaptive splitting and within-group calibration yields adaptive bands which can stretch and shrink locally. We demonstrate benefits of local tightening on several simulated as well as real examples using non-parametric regression. Finally, we consider two contemporary classification applications for obtaining uncertainty quantification around GPT-4o predictions. We conformalize skin disease diagnoses based on self-reported symptoms as well as predicted states of U.S. legislators based on summaries of their ideology. We demonstrate substantial local tightening of the uncertainty sets while attaining similar marginal coverage.
Authors: Zaile Li, Weiwei Fan, L. Jeff Hong
Abstract: Screening tasks that aim to identify a small subset of top alternatives from a large pool are common in business decision-making processes. These tasks often require substantial human effort to evaluate each alternative's performance, making them time-consuming and costly. Motivated by recent advances in large language models (LLMs), particularly their ability to generate outputs that align well with human evaluations, we consider an LLM-as-human-evaluator approach for conducting screening virtually, thereby reducing the cost burden. To achieve scalability and cost-effectiveness in virtual screening, we identify that the stochastic nature of LLM outputs and their cost structure necessitate efficient budget allocation across all alternatives. To address this, we propose using a top-$m$ greedy evaluation mechanism, a simple yet effective approach that keeps evaluating the current top-$m$ alternatives, and design the explore-first top-$m$ greedy (EFG-$m$) algorithm. We prove that EFG-$m$ is both sample-optimal and consistent in large-scale virtual screening. Surprisingly, we also uncover a bonus ranking effect, where the algorithm naturally induces an indifference-based ranking within the selected subset. To further enhance practicality, we design a suite of algorithm variants to improve screening performance and computational efficiency. Numerical experiments validate our results and demonstrate the effectiveness of our algorithms. Lastly, we conduct a case study on LLM-based virtual screening. The study shows that while LLMs alone may not provide meaningful screening and ranking results when directly queried, integrating them with our sample-optimal algorithms unlocks their potential for cost-effective, large-scale virtual screening.
Authors: Yile Chen, Weiming Huang, Kaiqi Zhao, Yue Jiang, Gao Cong
Abstract: The proliferation of various data sources in urban and territorial environments has significantly facilitated the development of geospatial artificial intelligence (GeoAI) across a wide range of geospatial applications. However, geospatial data, which is inherently linked to geospatial objects, often exhibits data heterogeneity that necessitates specialized fusion and representation strategies while simultaneously being inherently sparse in labels for downstream tasks. Consequently, there is a growing demand for techniques that can effectively leverage geospatial data without heavy reliance on task-specific labels and model designs. This need aligns with the principles of self-supervised learning (SSL), which has garnered increasing attention for its ability to learn effective and generalizable representations directly from data without extensive labeled supervision. This paper presents a comprehensive and up-to-date survey of SSL techniques specifically applied to or developed for geospatial objects in three primary vector geometric types: Point, Polyline, and Polygon. We systematically categorize various SSL techniques into predictive and contrastive methods, and analyze their adaptation to different data types for representation learning across various downstream tasks. Furthermore, we examine the emerging trends in SSL for geospatial objects, particularly the gradual advancements towards geospatial foundation models. Finally, we discuss key challenges in current research and outline promising directions for future investigation. By offering a structured analysis of existing studies, this paper aims to inspire continued progress in integrating SSL with geospatial objects, and the development of geospatial foundation models in a longer term.
Authors: Yuejiang Liu, Jubayer Ibn Hamid, Annie Xie, Yoonho Lee, Maximilian Du, Chelsea Finn
Abstract: Predicting and executing a sequence of actions without intermediate replanning, known as action chunking, is increasingly used in robot learning from human demonstrations. Yet, its effects on the learned policy remain inconsistent: some studies find it crucial for achieving strong results, while others observe decreased performance. In this paper, we first dissect how action chunking impacts the divergence between a learner and a demonstrator. We find that action chunking allows the learner to better capture the temporal dependencies in demonstrations but at the cost of reduced reactivity to unexpected states. To address this tradeoff, we propose Bidirectional Decoding (BID), a test-time inference algorithm that bridges action chunking with closed-loop adaptation. At each timestep, BID samples multiple candidate predictions and searches for the optimal one based on two criteria: (i) backward coherence, which favors samples that align with previous decisions; (ii) forward contrast, which seeks samples of high likelihood for future plans. By coupling decisions within and across action chunks, BID promotes both long-term consistency and short-term reactivity. Experimental results show that our method boosts the performance of two state-of-the-art generative policies across seven simulation benchmarks and two real-world tasks. Code and videos are available at https://bid-robot.github.io.
Authors: Parth Nobel, Daniel LeJeune, Emmanuel J. Cand\`es
Abstract: Estimating out-of-sample risk for models trained on large high-dimensional datasets is an expensive but essential part of the machine learning process, enabling practitioners to optimally tune hyperparameters. Cross-validation (CV) serves as the de facto standard for risk estimation but poorly trades off high bias ($K$-fold CV) for computational cost (leave-one-out CV). We propose a randomized approximate leave-one-out (RandALO) risk estimator that is not only a consistent estimator of risk in high dimensions but also less computationally expensive than $K$-fold CV. We support our claims with extensive simulations on synthetic and real data and provide a user-friendly Python package implementing RandALO available on PyPI as randalo and at https://github.com/cvxgrp/randalo.
Authors: Ningyuan Xi, Xiaoyu Wang, Yetao Wu, Teng Chen, Qingqing Gu, Yue Zhao, Jinxian Qu, Zhonglin Jiang, Yong Chen, Luo Ji
Abstract: Large Language Model can reasonably understand and generate human expressions but may lack of thorough thinking and reasoning mechanisms. Recently there have been several studies which enhance the thinking ability of language models but most of them are not data-driven or training-based. In this paper, we are motivated by the cognitive mechanism in the natural world, and design a novel model architecture called TaS which allows it to first consider the thoughts and then express the response based upon the query. We design several pipelines to annotate or generate the thought contents from prompt-response samples, then add language heads in a middle layer which behaves as the thinking layer. We train the language model by the thoughts-augmented data and successfully let the thinking layer automatically generate reasonable thoughts and finally output more reasonable responses. Both qualitative examples and quantitative results validate the effectiveness and performance of TaS. Our code is available at https://anonymous.4open.science/r/TadE.
Authors: Tifanny Portela, Andrei Cramariuc, Mayank Mittal, Marco Hutter
Abstract: Combining manipulation with the mobility of legged robots is essential for a wide range of robotic applications. However, integrating an arm with a mobile base significantly increases the system's complexity, making precise end-effector control challenging. Existing model-based approaches are often constrained by their modeling assumptions, leading to limited robustness. Meanwhile, recent Reinforcement Learning (RL) implementations restrict the arm's workspace to be in front of the robot or track only the position to obtain decent tracking accuracy. In this work, we address these limitations by introducing a whole-body RL formulation for end-effector pose tracking in a large workspace on rough, unstructured terrains. Our proposed method involves a terrain-aware sampling strategy for the robot's initial configuration and end-effector pose commands, as well as a game-based curriculum to extend the robot's operating range. We validate our approach on the ANYmal quadrupedal robot with a six DoF robotic arm. Through our experiments, we show that the learned controller achieves precise command tracking over a large workspace and adapts across varying terrains such as stairs and slopes. On deployment, it achieves a pose-tracking error of 2.64 cm and 3.64 degrees, outperforming existing competitive baselines.
Authors: Kianusch Vahid Yousefnia, Christoph Metzl, Tobias B\"olle
Abstract: Thunderstorms have significant social and economic impacts due to heavy precipitation, hail, lightning, and strong winds, necessitating reliable forecasts. Thunderstorm forecasts based on numerical weather prediction (NWP) often rely on single-level surrogate predictors, like convective available potential energy and convective inhibition, derived from vertical profiles of three-dimensional atmospheric variables. In this study, we develop SALAMA 1D, a deep neural network which directly infers the probability of thunderstorm occurrence from vertical profiles of ten atmospheric variables, bypassing single-level predictors. By training the model on convection-permitting NWP forecasts, we allow SALAMA 1D to flexibly identify convective patterns, with the goal of enhancing forecast accuracy. The model's architecture is physically motivated: sparse connections encourage interactions at similar height levels while keeping model size and inference times computationally efficient, whereas a shuffling mechanism prevents the model from learning non-physical patterns tied to the vertical grid. SALAMA 1D is trained over Central Europe with lightning observations as the ground truth. Comparative analysis against a baseline machine learning model that uses single-level predictors shows SALAMA 1D's superior skill across various metrics and lead times of up to at least 11 hours. Moreover, expanding the archive of forecasts from which training examples are sampled improves skill, even when training set size remains constant. Finally, a sensitivity analysis using saliency maps indicates that our model relies on physically interpretable patterns consistent with established theoretical understanding, such as ice particle content near the tropopause, cloud cover, conditional instability, and low-level moisture.
Authors: Yifei Ming, Senthil Purushwalkam, Shrey Pandit, Zixuan Ke, Xuan-Phi Nguyen, Caiming Xiong, Shafiq Joty
Abstract: Ensuring faithfulness to context in large language models (LLMs) and retrieval-augmented generation (RAG) systems is crucial for reliable deployment in real-world applications, as incorrect or unsupported information can erode user trust. Despite advancements on standard benchmarks, faithfulness hallucination-where models generate responses misaligned with the provided context-remains a significant challenge. In this work, we introduce FaithEval, a novel and comprehensive benchmark tailored to evaluate the faithfulness of LLMs in contextual scenarios across three diverse tasks: unanswerable, inconsistent, and counterfactual contexts. These tasks simulate real-world challenges where retrieval mechanisms may surface incomplete, contradictory, or fabricated information. FaithEval comprises 4.9K high-quality problems in total, validated through a rigorous four-stage context construction and validation framework, employing both LLM-based auto-evaluation and human validation. Our extensive study across a wide range of open-source and proprietary models reveals that even state-of-the-art models often struggle to remain faithful to the given context, and that larger models do not necessarily exhibit improved faithfulness.Project is available at: https://github.com/SalesforceAIResearch/FaithEval.
Authors: Vitalis Vosylius, Edward Johns
Abstract: Following the impressive capabilities of in-context learning with large transformers, In-Context Imitation Learning (ICIL) is a promising opportunity for robotics. We introduce Instant Policy, which learns new tasks instantly (without further training) from just one or two demonstrations, achieving ICIL through two key components. First, we introduce inductive biases through a graph representation and model ICIL as a graph generation problem with a learned diffusion process, enabling structured reasoning over demonstrations, observations, and actions. Second, we show that such a model can be trained using pseudo-demonstrations - arbitrary trajectories generated in simulation - as a virtually infinite pool of training data. Simulated and real experiments show that Instant Policy enables rapid learning of various everyday robot tasks. We also show how it can serve as a foundation for cross-embodiment and zero-shot transfer to language-defined tasks. Code and videos are available at https://www.robot-learning.uk/instant-policy.
Authors: Petar Bevanda, Nicolas Hoischen, Tobias Wittmann, Jan Br\"udigam, Sandra Hirche, Boris Houska
Abstract: This paper presents a novel operator-theoretic approach for optimal control of nonlinear stochastic systems within reproducing kernel Hilbert spaces. Our learning framework leverages data samples of system dynamics and stage cost functions, with only control penalties and constraints provided. The proposed method directly learns the infinitesimal generator of a controlled stochastic diffusion in an infinite-dimensional hypothesis space. We demonstrate that our approach seamlessly integrates with modern convex operator-theoretic Hamilton-Jacobi-Bellman recursions, enabling a data-driven solution to the optimal control problems. Furthermore, our learning framework includes nonparametric estimators for uncontrolled infinitesimal generators as a special case. Numerical experiments, ranging from synthetic differential equations to simulated robotic systems, showcase the advantages of our approach compared to both modern data-driven and classical nonlinear programming methods for optimal control.
Authors: Ziyang Cheng, Xiangyu Tian, Ruomin Sui, Tiemin Li, Yao Jiang
Abstract: Accurate grasp force control is one of the key skills for ensuring successful and damage-free robotic grasping of objects. Although existing methods have conducted in-depth research on slip detection and grasping force planning, they often overlook the issue of adaptive tracking of the actual force to the target force when handling objects with different material properties. The optimal parameters of a force tracking controller are significantly influenced by the object's stiffness, and many adaptive force tracking algorithms rely on stiffness estimation. However, real-world objects often exhibit viscous, plastic, or other more complex nonlinear time-varying behaviors, and existing studies provide insufficient support for these materials in terms of stiffness definition and estimation. To address this, this paper introduces the concept of generalized stiffness, extending the definition of stiffness to nonlinear time-varying grasp system models, and proposes an online generalized stiffness estimator based on Long Short-Term Memory (LSTM) networks. Based on generalized stiffness, this paper proposes an adaptive parameter adjustment strategy using a PI controller as an example, enabling dynamic force tracking for objects with varying characteristics. Experimental results demonstrate that the proposed method achieves high precision and short probing time, while showing better adaptability to non-ideal objects compared to existing methods. The method effectively solves the problem of grasp force tracking in unknown, nonlinear, and time-varying grasp systems, demonstrating the generalization capability of our neural network and enhancing the robotic grasping ability in unstructured environments.
Authors: Nikolaos Bousias, Stefanos Pertigkiozoglou, Kostas Daniilidis, George Pappas
Abstract: Multi-agent reinforcement learning has emerged as a powerful framework for enabling agents to learn complex, coordinated behaviors but faces persistent challenges regarding its generalization, scalability and sample efficiency. Recent advancements have sought to alleviate those issues by embedding intrinsic symmetries of the systems in the policy. Yet, most dynamical systems exhibit little to no symmetries to exploit. This paper presents a novel framework for embedding extrinsic symmetries in multi-agent system dynamics that enables the use of symmetry-enhanced methods to address systems with insufficient intrinsic symmetries, expanding the scope of equivariant learning to a wide variety of MARL problems. Central to our framework is the Group Equivariant Graphormer, a group-modular architecture specifically designed for distributed swarming tasks. Extensive experiments on a swarm of symmetry-breaking quadrotors validate the effectiveness of our approach, showcasing its potential for improved generalization and zero-shot scalability. Our method achieves significant reductions in collision rates and enhances task success rates across a diverse range of scenarios and varying swarm sizes.
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: George Drayson, Emine Yilmaz, Vasileios Lampos
Abstract: As Large Language Models (LLMs) become increasingly prevalent, their generated outputs are proliferating across the web, risking a future where machine-generated content dilutes human-authored text. Since online data is the primary resource for LLM pre-training, subsequent models could be trained on an unknown portion of synthetic samples. This will lead to model collapse, a degenerative process whereby LLMs reinforce their own errors, and ultimately yield a declining performance. In this study, we investigate the impact of decoding strategy on model collapse, analysing the characteristics of text at each model generation, the similarity to human references, and the resulting model performance. Using the decoding strategies that lead to the most significant degradation, we evaluate model collapse in more realistic scenarios where the origin of the data (human or synthetic) is unknown. We train a machine-generated text detector and propose an importance sampling approach to alleviate model collapse. Our method is validated on two LLM variants (GPT-2 and SmolLM2) on the open-ended text generation task. We demonstrate that it can not only prevent model collapse but also improve performance when sufficient human-authored samples are present. We release our code at https://github.com/GeorgeDrayson/model_collapse.
Authors: Jason Jingzhou Liu, Yulong Li, Kenneth Shaw, Tony Tao, Ruslan Salakhutdinov, Deepak Pathak
Abstract: Many contact-rich tasks humans perform, such as box pickup or rolling dough, rely on force feedback for reliable execution. However, this force information, which is readily available in most robot arms, is not commonly used in teleoperation and policy learning. Consequently, robot behavior is often limited to quasi-static kinematic tasks that do not require intricate force-feedback. In this paper, we first present a low-cost, intuitive, bilateral teleoperation setup that relays external forces of the follower arm back to the teacher arm, facilitating data collection for complex, contact-rich tasks. We then introduce FACTR, a policy learning method that employs a curriculum which corrupts the visual input with decreasing intensity throughout training. The curriculum prevents our transformer-based policy from over-fitting to the visual input and guides the policy to properly attend to the force modality. We demonstrate that by fully utilizing the force information, our method significantly improves generalization to unseen objects by 43\% compared to baseline approaches without a curriculum. Video results, codebases, and instructions at https://jasonjzliu.com/factr/
Authors: Chen Xu, Tony Khuong Nguyen, Emma Dixon, Christopher Rodriguez, Patrick Miller, Robert Lee, Paarth Shah, Rares Ambrus, Haruki Nishimura, Masha Itkina
Abstract: Recent years have witnessed impressive robotic manipulation systems driven by advances in imitation learning and generative modeling, such as diffusion- and flow-based approaches. As robot policy performance increases, so does the complexity and time horizon of achievable tasks, inducing unexpected and diverse failure modes that are difficult to predict a priori. To enable trustworthy policy deployment in safety-critical human environments, reliable runtime failure detection becomes important during policy inference. However, most existing failure detection approaches rely on prior knowledge of failure modes and require failure data during training, which imposes a significant challenge in practicality and scalability. In response to these limitations, we present FAIL-Detect, a modular two-stage approach for failure detection in imitation learning-based robotic manipulation. To accurately identify failures from successful training data alone, we frame the problem as sequential out-of-distribution (OOD) detection. We first distill policy inputs and outputs into scalar signals that correlate with policy failures and capture epistemic uncertainty. FAIL-Detect then employs conformal prediction (CP) as a versatile framework for uncertainty quantification with statistical guarantees. Empirically, we thoroughly investigate both learned and post-hoc scalar signal candidates on diverse robotic manipulation tasks. Our experiments show learned signals to be mostly consistently effective, particularly when using our novel flow-based density estimator. Furthermore, our method detects failures more accurately and faster than state-of-the-art (SOTA) failure detection baselines. These results highlight the potential of FAIL-Detect to enhance the safety and reliability of imitation learning-based robotic systems as they progress toward real-world deployment.
Authors: Jiuding Sun, Jing Huang, Sidharth Baskaran, Karel D'Oosterlinck, Christopher Potts, Michael Sklar, Atticus Geiger
Abstract: Mechanistic interpretability has made great strides in identifying neural network features (e.g., directions in hidden activation space) that mediate concepts(e.g., the birth year of a person) and enable predictable manipulation. Distributed alignment search (DAS) leverages supervision from counterfactual data to learn concept features within hidden states, but DAS assumes we can afford to conduct a brute force search over potential feature locations. To address this, we present HyperDAS, a transformer-based hypernetwork architecture that (1) automatically locates the token-positions of the residual stream that a concept is realized in and (2) constructs features of those residual stream vectors for the concept. In experiments with Llama3-8B, HyperDAS achieves state-of-the-art performance on the RAVEL benchmark for disentangling concepts in hidden states. In addition, we review the design decisions we made to mitigate the concern that HyperDAS (like all powerful interpretabilty methods) might inject new information into the target model rather than faithfully interpreting it.
Authors: Zijun Jia
Abstract: Road rage, driven by emotional outbursts, endangers road and public safety. Speech Emotion Recognition (SER) can detect early negative emotions to reduce accidents, but traditional methods (e.g., HMMs, LSTMs) using 1D speech signals face overfitting and miscalibration issues. This paper proposes a risk management framework ensuring statistically rigorous correctness coverage for test data. We separate a calibration set, design a binary loss function to check if ground-truth labels are in prediction sets, calibrated by data-driven threshold $\lambda$. A joint loss function on the calibration set adjusts $\lambda$ according to user-specified risk level $\alpha$, bounding the test loss expectation by $\alpha$. Evaluations on 6 models across 2 datasets show our framework strictly maintains average correctness coverage $\geq 1-\alpha$ and controls marginal error rates under various calibration-test splits (e.g., 0.1). Additionally, a small-batch online calibration framework based on local exchangeability is proposed for complex scenarios with data domain offset or non-IID batches. By constructing a non-negative test martingale, it ensures prediction set coverage in dynamic environments, validated via cross-dataset experiments.
Authors: Manolo Canales Cuba, Vin\'icius do Carmo Mel\'icio, Jo\~ao Paulo Gois
Abstract: Achieving high-fidelity and temporally smooth 3D human motion generation remains a challenge, particularly within resource-constrained environments. We introduce FlowMotion, a novel method leveraging Conditional Flow Matching (CFM). FlowMotion incorporates a training objective within CFM that focuses on more accurately predicting target motion in 3D human motion generation, resulting in enhanced generation fidelity and temporal smoothness while maintaining the fast synthesis times characteristic of flow-matching-based methods. FlowMotion achieves state-of-the-art jitter performance, achieving the best jitter in the KIT dataset and the second-best jitter in the HumanML3D dataset, and a competitive FID value in both datasets. This combination provides robust and natural motion sequences, offering a promising equilibrium between generation quality and temporal naturalness.
Authors: Haowei Lin, Xiangyu Wang, Ruilin Yan, Baizhou Huang, Haotian Ye, Jianhua Zhu, Zihao Wang, James Zou, Jianzhu Ma, Yitao Liang
Abstract: With powerful large language models (LLMs) demonstrating superhuman reasoning capabilities, a critical question arises: Do LLMs genuinely reason, or do they merely recall answers from their extensive, web-scraped training datasets? Publicly released benchmarks inevitably become contaminated once incorporated into subsequent LLM training sets, undermining their reliability as faithful assessments. To address this, we introduce KUMO, a generative evaluation framework designed specifically for assessing reasoning in LLMs. KUMO synergistically combines LLMs with symbolic engines to dynamically produce diverse, multi-turn reasoning tasks that are partially observable and adjustable in difficulty. Through an automated pipeline, KUMO continuously generates novel tasks across open-ended domains, compelling models to demonstrate genuine generalization rather than memorization. We evaluated 23 state-of-the-art LLMs on 5,000 tasks across 100 domains created by KUMO, benchmarking their reasoning abilities against university students. Our findings reveal that many LLMs have outperformed university-level performance on easy reasoning tasks, and reasoning-scaled LLMs reach university-level performance on complex reasoning challenges. Moreover, LLM performance on KUMO tasks correlates strongly with results on newly released real-world reasoning benchmarks, underscoring KUMO's value as a robust, enduring assessment tool for genuine LLM reasoning capabilities.
Authors: David Landry, Claire Monteleoni, Anastase Charantonis
Abstract: We propose a machine-learning-based methodology for in-situ weather forecast postprocessing that is both spatially coherent and multivariate. Compared to previous work, our Flow MAtching Postprocessing (FMAP) better represents the correlation structures of the observations distribution, while also improving marginal performance at the stations. FMAP generates forecasts that are not bound to what is already modeled by the underlying gridded prediction and can infer new correlation structures from data. The resulting model can generate an arbitrary number of forecasts from a limited number of numerical simulations, allowing for low-cost forecasting systems. A single training is sufficient to perform postprocessing at multiple lead times, in contrast with other methods which use multiple trained networks at generation time. This work details our methodology, including a spatial attention transformer backbone trained within a flow matching generative modeling framework. FMAP shows promising performance in experiments on the EUPPBench dataset, forecasting surface temperature and wind gust values at station locations in western Europe up to five-day lead times.
Authors: Kidist Amde Mekonnen, Yubao Tang, Maarten de Rijke
Abstract: Generative information retrieval (GenIR) is a promising neural retrieval paradigm that formulates document retrieval as a document identifier (docid) generation task, allowing for end-to-end optimization toward a unified global retrieval objective. However, existing GenIR models suffer from token-level misalignment, where models trained to predict the next token often fail to capture document-level relevance effectively. While reinforcement learning-based methods, such as reinforcement learning from relevance feedback (RLRF), aim to address this misalignment through reward modeling, they introduce significant complexity, requiring the optimization of an auxiliary reward function followed by reinforcement fine-tuning, which is computationally expensive and often unstable. To address these challenges, we propose direct document relevance optimization (DDRO), which aligns token-level docid generation with document-level relevance estimation through direct optimization via pairwise ranking, eliminating the need for explicit reward modeling and reinforcement learning. Experimental results on benchmark datasets, including MS MARCO document and Natural Questions, show that DDRO outperforms reinforcement learning-based methods, achieving a 7.4% improvement in MRR@10 for MS MARCO and a 19.9% improvement for Natural Questions. These findings highlight DDRO's potential to enhance retrieval effectiveness with a simplified optimization approach. By framing alignment as a direct optimization problem, DDRO simplifies the ranking optimization pipeline of GenIR models while offering a viable alternative to reinforcement learning-based methods.
Authors: Szymon Mazurek, Jakub Caputa, Jan K. Argasi\'nski, Maciej Wielgosz
Abstract: Three-factor learning rules in Spiking Neural Networks (SNNs) have emerged as a crucial extension to traditional Hebbian learning and Spike-Timing-Dependent Plasticity (STDP), incorporating neuromodulatory signals to improve adaptation and learning efficiency. These mechanisms enhance biological plausibility and facilitate improved credit assignment in artificial neural systems. This paper takes a view on this topic from a machine learning perspective, providing an overview of recent advances in three-factor learning, discusses theoretical foundations, algorithmic implementations, and their relevance to reinforcement learning and neuromorphic computing. In addition, we explore interdisciplinary approaches, scalability challenges, and potential applications in robotics, cognitive modeling, and AI systems. Finally, we highlight key research gaps and propose future directions for bridging the gap between neuroscience and artificial intelligence.
Authors: Hamidreza Eivazi, Jendrik-Alexander Tr\"oger, Stefan Wittek, Stefan Hartmann, Andreas Rausch
Abstract: Multiscale problems are ubiquitous in physics. Numerical simulations of such problems by solving partial differential equations (PDEs) at high resolution are computationally too expensive for many-query scenarios, e.g., uncertainty quantification, remeshing applications, topology optimization, and so forth. This limitation has motivated the application of data-driven surrogate models, where the microscale computations are $\textit{substituted}$ with a surrogate, usually acting as a black-box mapping between macroscale quantities. These models offer significant speedups but struggle with incorporating microscale physical constraints, such as the balance of linear momentum and constitutive models. In this contribution, we propose Equilibrium Neural Operator (EquiNO) as a $\textit{complementary}$ physics-informed PDE surrogate for predicting microscale physics and compare it with variational physics-informed neural and operator networks. Our framework, applicable to the so-called multiscale FE$^{\,2}\,$ computations, introduces the FE-OL approach by integrating the finite element (FE) method with operator learning (OL). We apply the proposed FE-OL approach to quasi-static problems of solid mechanics. The results demonstrate that FE-OL can yield accurate solutions even when confronted with a restricted dataset during model development. Our results show that EquiNO achieves speedup factors exceeding 8000-fold compared to traditional methods and offers an optimal balance between data-driven and physics-based strategies.
Authors: Minjie Deng, Yan Wei, Hao Zhai, An Wu, Yuncan Ouyang, Qianyao Peng
Abstract: In image fusion tasks, the absence of real fused images as priors presents a fundamental challenge. Most deep learning-based fusion methods rely on large-scale paired datasets to extract global weighting features from raw images, thereby generating fused outputs that approximate real fused images. In contrast to previous studies, this paper explores few-shot training of neural networks under the condition of having prior knowledge. We propose a novel fusion framework named GBFF, and a Granular Ball Significant Extraction algorithm specifically designed for the few-shot prior setting. All pixel pairs involved in the fusion process are initially modeled as a Coarse-Grained Granular Ball. At the local level, Fine-Grained Granular Balls are used to slide through the brightness space to extract Non-Salient Pixel Pairs, and perform splitting operations to obtain Salient Pixel Pairs. Pixel-wise weights are then computed to generate a pseudo-supervised image. At the global level, pixel pairs with significant contributions to the fusion process are categorized into the Positive Region, while those whose contributions cannot be accurately determined are assigned to the Boundary Region. The Granular Ball performs modality-aware adaptation based on the proportion of the positive region, thereby adjusting the neural network's loss function and enabling it to complement the information of the boundary region. Extensive experiments demonstrate the effectiveness of both the proposed algorithm and the underlying theory. Compared with state-of-the-art (SOTA) methods, our approach shows strong competitiveness in terms of both fusion time and image expressiveness. Our code is publicly available at:
Authors: Shuming Ma, Hongyu Wang, Shaohan Huang, Xingxing Zhang, Ying Hu, Ting Song, Yan Xia, Furu Wei
Abstract: We introduce BitNet b1.58 2B4T, the first open-source, native 1-bit Large Language Model (LLM) at the 2-billion parameter scale. Trained on a corpus of 4 trillion tokens, the model has been rigorously evaluated across benchmarks covering language understanding, mathematical reasoning, coding proficiency, and conversational ability. Our results demonstrate that BitNet b1.58 2B4T achieves performance on par with leading open-weight, full-precision LLMs of similar size, while offering significant advantages in computational efficiency, including substantially reduced memory footprint, energy consumption, and decoding latency. To facilitate further research and adoption, the model weights are released via Hugging Face along with open-source inference implementations for both GPU and CPU architectures.
Authors: Ole-Christian Galbo Engstr{\o}m, Michela Albano-Gaglio, Erik Schou Dreier, Yamine Bouzembrak, Maria Font-i-Furnols, Puneet Mishra, Kim Steenstrup Pedersen
Abstract: Current approaches to chemical map generation from hyperspectral images are based on models such as partial least squares (PLS) regression, generating pixel-wise predictions that do not consider spatial context and suffer from a high degree of noise. This study proposes an end-to-end deep learning approach using a modified version of U-Net and a custom loss function to directly obtain chemical maps from hyperspectral images, skipping all intermediate steps required for traditional pixel-wise analysis. We compare the U-Net with the traditional PLS regression on a real dataset of pork belly samples with associated mean fat reference values. The U-Net obtains a test set root mean squared error of between 9% and 13% lower than that of PLS regression on the task of mean fat prediction. At the same time, U-Net generates fine detail chemical maps where 99.91% of the variance is spatially correlated. Conversely, only 2.53% of the variance in the PLS-generated chemical maps is spatially correlated, indicating that each pixel-wise prediction is largely independent of neighboring pixels. Additionally, while the PLS-generated chemical maps contain predictions far beyond the physically possible range of 0-100%, U-Net learns to stay inside this range. Thus, the findings of this study indicate that U-Net is superior to PLS for chemical map generation.
Authors: Yihan Lin, Zhirong Bella Yu, Simon Lee
Abstract: Synthetic Electronic Health Records (EHRs) offer a valuable opportunity to create privacy preserving and harmonized structured data, supporting numerous applications in healthcare. Key benefits of synthetic data include precise control over the data schema, improved fairness and representation of patient populations, and the ability to share datasets without concerns about compromising real individuals privacy. Consequently, the AI community has increasingly turned to Large Language Models (LLMs) to generate synthetic data across various domains. However, a significant challenge in healthcare is ensuring that synthetic health records reliably generalize across different hospitals, a long standing issue in the field. In this work, we evaluate the current state of commercial LLMs for generating synthetic data and investigate multiple aspects of the generation process to identify areas where these models excel and where they fall short. Our main finding from this work is that while LLMs can reliably generate synthetic health records for smaller subsets of features, they struggle to preserve realistic distributions and correlations as the dimensionality of the data increases, ultimately limiting their ability to generalize across diverse hospital settings.
Authors: Zexi Fan, Yan Sun, Shihao Yang, Yiping Lu
Abstract: High-dimensional partial differential equations (PDEs) pose significant computational challenges across fields ranging from quantum chemistry to economics and finance. Although scientific machine learning (SciML) techniques offer approximate solutions, they often suffer from bias and neglect crucial physical insights. Inspired by inference-time scaling strategies in language models, we propose Simulation-Calibrated Scientific Machine Learning (SCaSML), a physics-informed framework that dynamically refines and debiases the SCiML predictions during inference by enforcing the physical laws. SCaSML leverages derived new physical laws that quantifies systematic errors and employs Monte Carlo solvers based on the Feynman-Kac and Elworthy-Bismut-Li formulas to dynamically correct the prediction. Both numerical and theoretical analysis confirms enhanced convergence rates via compute-optimal inference methods. Our numerical experiments demonstrate that SCaSML reduces errors by 20-50% compared to the base surrogate model, establishing it as the first algorithm to refine approximated solutions to high-dimensional PDE during inference. Code of SCaSML is available at https://github.com/Francis-Fan-create/SCaSML.
Authors: Ye Qiao, Zhiheng Chen, Yifan Zhang, Yian Wang, Sitao Huang
Abstract: Deploying large language models (LLMs) on edge platforms is challenged by their high computational and memory demands. Although recent low-bit quantization methods (e.g., BitNet, DeepSeek) compress weights to as little as 1.58 bits with minimal accuracy loss, edge deployment is still constrained by limited on-chip resources, power budgets, and the often-neglected latency of the prefill phase. We present TeLLMe, the first ternary LLM accelerator for low-power FPGAs (e.g., AMD KV260) that fully supports both prefill and autoregressive decoding using 1.58-bit weights and 8-bit activations. Our contributions include: (1) a table-lookup matrix engine for ternary matmul that merges grouped activations with online precomputation to minimize resource use; (2) a fused, bandwidth-efficient attention module featuring a reversed reordering scheme to accelerate prefill; and (3) a tightly integrated normalization and quantization--dequantization unit optimized for ultra-low-bit inference. Under a 7W power budget, TeLLMe delivers up to 9 tokens/s throughput over 1,024-token contexts and prefill latencies of 0.55--1.15 s for 64--128 token prompts, marking a significant energy-efficiency advance and establishing a new edge FPGA benchmark for generative AI.
Authors: Ye Qiao, Zhiheng Chen, Yian Wang, Yifan Zhang, Yunzhe Deng, Sitao Huang
Abstract: Transformer-based models have demonstrated superior performance in various fields, including natural language processing and computer vision. However, their enormous model size and high demands in computation, memory, and communication limit their deployment to edge platforms for local, secure inference. Binary transformers offer a compact, low-complexity solution for edge deployment with reduced bandwidth needs and acceptable accuracy. However, existing binary transformers perform inefficiently on current hardware due to the lack of binary specific optimizations. To address this, we introduce COBRA, an algorithm-architecture co-optimized binary Transformer accelerator for edge computing. COBRA features a real 1-bit binary multiplication unit, enabling matrix operations with -1, 0, and +1 values, surpassing ternary methods. With further hardware-friendly optimizations in the attention block, COBRA achieves up to 3,894.7 GOPS throughput and 448.7 GOPS/Watt energy efficiency on edge FPGAs, delivering a 311x energy efficiency improvement over GPUs and a 3.5x throughput improvement over the state-of-the-art binary accelerator, with only negligible inference accuracy degradation.
Authors: Yuanchang Ye, Weiyan Wen
Abstract: This study addresses the critical challenge of hallucination mitigation in Large Vision-Language Models (LVLMs) for Visual Question Answering (VQA) tasks through a Split Conformal Prediction (SCP) framework. While LVLMs excel in multi-modal reasoning, their outputs often exhibit hallucinated content with high confidence, posing risks in safety-critical applications. We propose a model-agnostic uncertainty quantification method that integrates dynamic threshold calibration and cross-modal consistency verification. By partitioning data into calibration and test sets, the framework computes nonconformity scores to construct prediction sets with statistical guarantees under user-defined risk levels ($\alpha$). Key innovations include: (1) rigorous control of \textbf{marginal coverage} to ensure empirical error rates remain strictly below $\alpha$; (2) dynamic adjustment of prediction set sizes inversely with $\alpha$, filtering low-confidence outputs; (3) elimination of prior distribution assumptions and retraining requirements. Evaluations on benchmarks (ScienceQA, MMMU) with eight LVLMs demonstrate that SCP enforces theoretical guarantees across all $\alpha$ values. The framework achieves stable performance across varying calibration-to-test split ratios, underscoring its robustness for real-world deployment in healthcare, autonomous systems, and other safety-sensitive domains. This work bridges the gap between theoretical reliability and practical applicability in multi-modal AI systems, offering a scalable solution for hallucination detection and uncertainty-aware decision-making.