Authors: Tong Nie, Jian Sun, Wei Ma
Abstract: Networked urban systems facilitate the flow of people, resources, and services, and are essential for economic and social interactions. These systems often involve complex processes with unknown governing rules, observed by sensor-based time series. To aid decision-making in industrial and engineering contexts, data-driven predictive models are used to forecast spatiotemporal dynamics of urban systems. Current models such as graph neural networks have shown promise but face a trade-off between efficacy and efficiency due to computational demands. Hence, their applications in large-scale networks still require further efforts. This paper addresses this trade-off challenge by drawing inspiration from physical laws to inform essential model designs that align with fundamental principles and avoid architectural redundancy. By understanding both micro- and macro-processes, we present a principled interpretable neural diffusion scheme based on Transformer-like structures whose attention layers are induced by low-dimensional embeddings. The proposed scalable spatiotemporal Transformer (ScaleSTF), with linear complexity, is validated on large-scale urban systems including traffic flow, solar power, and smart meters, showing state-of-the-art performance and remarkable scalability. Our results constitute a fresh perspective on the dynamics prediction in large-scale urban networks.
Authors: Kaustav Chatterjee, Joshua Q. Li, Fatemeh Ansari, Masud Rana Munna, Kundan Parajulee, Jared Schwennesen
Abstract: Hump crossings, or high-profile Highway Railway Grade Crossings (HRGCs), pose safety risks to highway vehicles due to potential hang-ups. These crossings typically result from post-construction railway track maintenance activities or non-compliance with design guidelines for HRGC vertical alignments. Conventional methods for measuring HRGC profiles are costly, time-consuming, traffic-disruptive, and present safety challenges. To address these issues, this research employed advanced, cost-effective techniques and innovative modeling approaches for HRGC profile measurement. A novel hybrid deep learning framework combining Long Short-Term Memory (LSTM) and Transformer architectures was developed by utilizing instrumentation and ground truth data. Instrumentation data were gathered using a highway testing vehicle equipped with Inertial Measurement Unit (IMU) and Global Positioning System (GPS) sensors, while ground truth data were obtained via an industrial-standard walking profiler. Field data was collected at the Red Rock Railroad Corridor in Oklahoma. Three advanced deep learning models Transformer-LSTM sequential (model 1), LSTM-Transformer sequential (model 2), and LSTM-Transformer parallel (model 3) were evaluated to identify the most efficient architecture. Models 2 and 3 outperformed the others and were deployed to generate 2D/3D HRGC profiles. The deep learning models demonstrated significant potential to enhance highway and railroad safety by enabling rapid and accurate assessment of HRGC hang-up susceptibility.
Authors: Abhinav Das, Stephan Schl\"uter
Abstract: This work integrates Bayesian regime detection with conditional neural processes for 24-hour electricity price prediction in the German market. Our methodology integrates regime detection using a disentangled sticky hierarchical Dirichlet process hidden Markov model (DS-HDP-HMM) applied to daily electricity prices. Each identified regime is subsequently modeled by an independent conditional neural process (CNP), trained to learn localized mappings from input contexts to 24-dimensional hourly price trajectories, with final predictions computed as regime-weighted mixtures of these CNP outputs. We rigorously evaluate R-NP against deep neural networks (DNN) and Lasso estimated auto-regressive (LEAR) models by integrating their forecasts into diverse battery storage optimization frameworks, including price arbitrage, risk management, grid services, and cost minimization. This operational utility assessment revealed complex performance trade-offs: LEAR often yielded superior absolute profits or lower costs, while DNN showed exceptional optimality in specific cost-minimization contexts. Recognizing that raw prediction accuracy doesn't always translate to optimal operational outcomes, we employed TOPSIS as a comprehensive multi-criteria evaluation layer. Our TOPSIS analysis identified LEAR as the top-ranked model for 2021, but crucially, our proposed R-NP model emerged as the most balanced and preferred solution for 2021, 2022 and 2023.
Authors: Yebo Wu, Jingguang Li, Zhijiang Guo, Li Li
Abstract: Federated fine-tuning enables Large Language Models (LLMs) to adapt to downstream tasks while preserving data privacy, but its resource-intensive nature limits deployment on edge devices. In this paper, we introduce Developmental Federated Tuning (DevFT), a resource-efficient approach inspired by cognitive development that progressively builds a powerful LLM from a compact foundation. DevFT decomposes the fine-tuning process into developmental stages, each optimizing submodels with increasing parameter capacity. Knowledge from earlier stages transfers to subsequent submodels, providing optimized initialization parameters that prevent convergence to local minima and accelerate training. This paradigm mirrors human learning, gradually constructing comprehensive knowledge structure while refining existing skills. To efficiently build stage-specific submodels, DevFT introduces deconfliction-guided layer grouping and differential-based layer fusion to distill essential information and construct representative layers. Evaluations across multiple benchmarks demonstrate that DevFT significantly outperforms state-of-the-art methods, achieving up to 4.59$\times$ faster convergence, 10.67$\times$ reduction in communication overhead, and 9.07% average performance improvement, while maintaining compatibility with existing approaches.
Authors: Nhut Truong, Uri Hasson
Abstract: Topographic neural networks are computational models that can simulate the spatial and functional organization of the brain. Topographic constraints in neural networks can be implemented in multiple ways, with potentially different impacts on the representations learned by the network. The impact of such different implementations has not been systematically examined. To this end, here we compare topographic convolutional neural networks trained with two spatial constraints: Weight Similarity (WS), which pushes neighboring units to develop similar incoming weights, and Activation Similarity (AS), which enforces similarity in unit activations. We evaluate the resulting models on classification accuracy, robustness to weight perturbations and input degradation, and the spatial organization of learned representations. Compared to both AS and standard CNNs, WS provided three main advantages: i) improved robustness to noise, also showing higher accuracy under weight corruption; ii) greater input sensitivity, reflected in higher activation variance; and iii) stronger functional localization, with units showing similar activations positioned at closer distances. In addition, WS produced differences in orientation tuning, symmetry sensitivity, and eccentricity profiles of units, indicating an influence of this spatial constraint on the representational geometry of the network. Our findings suggest that during end-to-end training, WS constraints produce more robust representations than AS or non-topographic CNNs. These findings also suggest that weight-based spatial constraints can shape feature learning and functional organization in biophysical inspired models.
Authors: Ruo Yu Tao, Kaicheng Guo, Cameron Allen, George Konidaris
Abstract: Mitigating partial observability is a necessary but challenging task for general reinforcement learning algorithms. To improve an algorithm's ability to mitigate partial observability, researchers need comprehensive benchmarks to gauge progress. Most algorithms tackling partial observability are only evaluated on benchmarks with simple forms of state aliasing, such as feature masking and Gaussian noise. Such benchmarks do not represent the many forms of partial observability seen in real domains, like visual occlusion or unknown opponent intent. We argue that a partially observable benchmark should have two key properties. The first is coverage in its forms of partial observability, to ensure an algorithm's generalizability. The second is a large gap between the performance of a agents with more or less state information, all other factors roughly equal. This gap implies that an environment is memory improvable: where performance gains in a domain are from an algorithm's ability to cope with partial observability as opposed to other factors. We introduce best-practice guidelines for empirically benchmarking reinforcement learning under partial observability, as well as the open-source library POBAX: Partially Observable Benchmarks in JAX. We characterize the types of partial observability present in various environments and select representative environments for our benchmark. These environments include localization and mapping, visual control, games, and more. Additionally, we show that these tasks are all memory improvable and require hard-to-learn memory functions, providing a concrete signal for partial observability research. This framework includes recommended hyperparameters as well as algorithm implementations for fast, out-of-the-box evaluation, as well as highly performant environments implemented in JAX for GPU-scalable experimentation.
Authors: Yuan-Cheng Yu, Yen-Chieh Ouyang, Chun-An Lin
Abstract: Time-series anomaly detection plays a central role across a wide range of application domains. With the increasing proliferation of the Internet of Things (IoT) and smart manufacturing, time-series data has dramatically increased in both scale and dimensionality. This growth has exposed the limitations of traditional statistical methods in handling the high heterogeneity and complexity of such data. Inspired by the recent success of large language models (LLMs) in multimodal tasks across language and vision domains, we propose a novel unsupervised anomaly detection framework: A Tri-Branch Patch-wise Large Language Model Framework for Time-Series Anomaly Detection (TriP-LLM). TriP-LLM integrates local and global temporal features through a tri-branch design-Patching, Selection, and Global-to encode the input time series into patch-wise tokens, which are then processed by a frozen, pretrained LLM. A lightweight patch-wise decoder reconstructs the input, from which anomaly scores are derived. We evaluate TriP-LLM on several public benchmark datasets using PATE, a recently proposed threshold-free evaluation metric, and conduct all comparisons within a unified open-source framework to ensure fairness. Experimental results show that TriP-LLM consistently outperforms recent state-of-the-art methods across all datasets, demonstrating strong detection capabilities. Furthermore, through extensive ablation studies, we verify the substantial contribution of the LLM to the overall architecture. Compared to LLM-based approaches using Channel Independence (CI) patch processing, TriP-LLM achieves significantly lower memory consumption, making it more suitable for GPU memory-constrained environments. All code and model checkpoints are publicly available on https://github.com/YYZStart/TriP-LLM.git
Authors: Imen Mahmoud, Andrei Velichko
Abstract: This study proposes a novel methodological framework integrating a LightGBM regression model and genetic algorithm (GA) optimization to systematically evaluate the contribution of COVID-19-related indicators to Bitcoin return prediction. The primary objective was not merely to forecast Bitcoin returns but rather to determine whether including pandemic-related health data significantly enhances prediction accuracy. A comprehensive dataset comprising daily Bitcoin returns and COVID-19 metrics (vaccination rates, hospitalizations, testing statistics) was constructed. Predictive models, trained with and without COVID-19 features, were optimized using GA over 31 independent runs, allowing robust statistical assessment. Performance metrics (R2, RMSE, MAE) were statistically compared through distribution overlaps and Mann-Whitney U tests. Permutation Feature Importance (PFI) analysis quantified individual feature contributions. Results indicate that COVID-19 indicators significantly improved model performance, particularly in capturing extreme market fluctuations (R2 increased by 40%, RMSE decreased by 2%, both highly significant statistically). Among COVID-19 features, vaccination metrics, especially the 75th percentile of fully vaccinated individuals, emerged as dominant predictors. The proposed methodology extends existing financial analytics tools by incorporating public health signals, providing investors and policymakers with refined indicators to navigate market uncertainty during systemic crises.
Authors: Ashkan Shakarami, Yousef Yeganeh, Azade Farshad, Lorenzo Nicole, Stefano Ghidoni, Nassir Navab
Abstract: This paper introduces Stress-Aware Learning, a resilient neural training paradigm in which deep neural networks dynamically adjust their optimization behavior - whether under stable training regimes or in settings with uncertain dynamics - based on the concept of Temporary (Elastic) and Permanent (Plastic) Deformation, inspired by structural fatigue in materials science. To instantiate this concept, we propose Plastic Deformation Optimizer, a stress-aware mechanism that injects adaptive noise into model parameters whenever an internal stress signal - reflecting stagnation in training loss and accuracy - indicates persistent optimization difficulty. This enables the model to escape sharp minima and converge toward flatter, more generalizable regions of the loss landscape. Experiments across six architectures, four optimizers, and seven vision benchmarks demonstrate improved robustness and generalization with minimal computational overhead. The code and 3D visuals will be available on GitHub: https://github.com/Stress-Aware-Learning/SAL.
Authors: Md. Ehsanul Haque, S. M. Jahidul Islam, Shakil Mia, Rumana Sharmin, Ashikuzzaman, Md Samir Morshed, Md. Tahmidul Huque
Abstract: Liver diseases are a serious health concern in the world, which requires precise and timely diagnosis to enhance the survival chances of patients. The current literature implemented numerous machine learning and deep learning models to classify liver diseases, but most of them had some issues like high misclassification error, poor interpretability, prohibitive computational expense, and lack of good preprocessing strategies. In order to address these drawbacks, we introduced StackLiverNet in this study; an interpretable stacked ensemble model tailored to the liver disease detection task. The framework uses advanced data preprocessing and feature selection technique to increase model robustness and predictive ability. Random undersampling is performed to deal with class imbalance and make the training balanced. StackLiverNet is an ensemble of several hyperparameter-optimized base classifiers, whose complementary advantages are used through a LightGBM meta-model. The provided model demonstrates excellent performance, with the testing accuracy of 99.89%, Cohen Kappa of 0.9974, and AUC of 0.9993, having only 5 misclassifications, and efficient training and inference speeds that are amenable to clinical practice (training time 4.2783 seconds, inference time 0.1106 seconds). Besides, Local Interpretable Model-Agnostic Explanations (LIME) are applied to generate transparent explanations of individual predictions, revealing high concentrations of Alkaline Phosphatase and moderate SGOT as important observations of liver disease. Also, SHAP was used to rank features by their global contribution to predictions, while the Morris method confirmed the most influential features through sensitivity analysis.
Authors: Saleh Nikooroo, Thomas Engel
Abstract: Despite their impressive performance, contemporary neural networks often lack structural safeguards that promote stable learning and interpretable behavior. In this work, we introduce a reformulation of layer-level transformations that departs from the standard unconstrained affine paradigm. Each transformation is decomposed into a structured linear operator and a residual corrective component, enabling more disciplined signal propagation and improved training dynamics. Our formulation encourages internal consistency and supports stable information flow across depth, while remaining fully compatible with standard learning objectives and backpropagation. Through a series of synthetic and real-world experiments, we demonstrate that models constructed with these structured transformations exhibit improved gradient conditioning, reduced sensitivity to perturbations, and layer-wise robustness. We further show that these benefits persist across architectural scales and training regimes. This study serves as a foundation for a more principled class of neural architectures that prioritize stability and transparency-offering new tools for reasoning about learning behavior without sacrificing expressive power.
Authors: Christopher Harvey, Sumaiya Shomaji, Zijun Yao, Amit Noheria
Abstract: The electrocardiogram (ECG) is an inexpensive and widely available tool for cardiac assessment. Despite its standardized format and small file size, the high complexity and inter-individual variability of ECG signals (typically a 60,000-size vector with 12 leads at 500 Hz) make it challenging to use in deep learning models, especially when only small training datasets are available. This study addresses these challenges by exploring feature generation methods from representative beat ECGs, focusing on Principal Component Analysis (PCA) and Autoencoders to reduce data complexity. We introduce three novel Variational Autoencoder (VAE) variants-Stochastic Autoencoder (SAE), Annealed beta-VAE (A beta-VAE), and Cyclical beta VAE (C beta-VAE)-and compare their effectiveness in maintaining signal fidelity and enhancing downstream prediction tasks using a Light Gradient Boost Machine (LGBM). The A beta-VAE achieved superior signal reconstruction, reducing the mean absolute error (MAE) to 15.7+/-3.2 muV, which is at the level of signal noise. Moreover, the SAE encodings, when combined with traditional ECG summary features, improved the prediction of reduced Left Ventricular Ejection Fraction (LVEF), achieving an holdout test set area under the receiver operating characteristic curve (AUROC) of 0.901 with a LGBM classifier. This performance nearly matches the 0.909 AUROC of state-of-the-art CNN model but requires significantly less computational resources. Further, the ECG feature extraction-LGBM pipeline avoids overfitting and retains predictive performance when trained with less data. Our findings demonstrate that these VAE encodings are not only effective in simplifying ECG data but also provide a practical solution for applying deep learning in contexts with limited-scale labeled training data.
Authors: Mohit Gupta, Debjit Bhowmick, Rhys Newbury, Meead Saberi, Shirui Pan, Ben Beck
Abstract: Accurate link-level bicycling volume estimation is essential for sustainable urban transportation planning. However, many cities face significant challenges of high data sparsity due to limited bicycling count sensor coverage. To address this issue, we propose INSPIRE-GNN, a novel Reinforcement Learning (RL)-boosted hybrid Graph Neural Network (GNN) framework designed to optimize sensor placement and improve link-level bicycling volume estimation in data-sparse environments. INSPIRE-GNN integrates Graph Convolutional Networks (GCN) and Graph Attention Networks (GAT) with a Deep Q-Network (DQN)-based RL agent, enabling a data-driven strategic selection of sensor locations to maximize estimation performance. Applied to Melbourne's bicycling network, comprising 15,933 road segments with sensor coverage on only 141 road segments (99% sparsity) - INSPIRE-GNN demonstrates significant improvements in volume estimation by strategically selecting additional sensor locations in deployments of 50, 100, 200 and 500 sensors. Our framework outperforms traditional heuristic methods for sensor placement such as betweenness centrality, closeness centrality, observed bicycling activity and random placement, across key metrics such as Mean Squared Error (MSE), Root Mean Squared Error (RMSE) and Mean Absolute Error (MAE). Furthermore, our experiments benchmark INSPIRE-GNN against standard machine learning and deep learning models in the bicycle volume estimation performance, underscoring its effectiveness. Our proposed framework provides transport planners actionable insights to effectively expand sensor networks, optimize sensor placement and maximize volume estimation accuracy and reliability of bicycling data for informed transportation planning decisions.
Authors: Ziqian Zhong, Aditi Raghunathan
Abstract: The releases of powerful open-weight large language models (LLMs) are often not accompanied by access to their full training data. Existing interpretability methods, particularly those based on activations, often require or assume distributionally similar data. This is a significant limitation when detecting and defending against novel potential threats like backdoors, which are by definition out-of-distribution. In this work, we introduce a new method for understanding, monitoring and controlling fine-tuned LLMs that interprets weights, rather than activations, thereby side stepping the need for data that is distributionally similar to the unknown training data. We demonstrate that the top singular vectors of the weight difference between a fine-tuned model and its base model correspond to newly acquired behaviors. By monitoring the cosine similarity of activations along these directions, we can detect salient behaviors introduced during fine-tuning with high precision. For backdoored models that bypasses safety mechanisms when a secret trigger is present, our method stops up to 100% of attacks with a false positive rate below 1.2%. For models that have undergone unlearning, we detect inference on erased topics with accuracy up to 95.42% and can even steer the model to recover "unlearned" information. Besides monitoring, our method also shows potential for pre-deployment model auditing: by analyzing commercial instruction-tuned models (OLMo, Llama, Qwen), we are able to uncover model-specific fine-tuning focus including marketing strategies and Midjourney prompt generation. Our implementation can be found at https://github.com/fjzzq2002/WeightWatch.
Authors: Fupei Guo, Hao Zheng, Xiang Zhang, Li Chen, Yue Wang, Songyang Zhang
Abstract: The rapid development of artificial intelligence has driven smart health with next-generation wireless communication technologies, stimulating exciting applications in remote diagnosis and intervention. To enable a timely and effective response for remote healthcare, efficient transmission of medical data through noisy channels with limited bandwidth emerges as a critical challenge. In this work, we propose a novel diffusion-based semantic communication framework, namely DiSC-Med, for the medical image transmission, where medical-enhanced compression and denoising blocks are developed for bandwidth efficiency and robustness, respectively. Unlike conventional pixel-wise communication framework, our proposed DiSC-Med is able to capture the key semantic information and achieve superior reconstruction performance with ultra-high bandwidth efficiency against noisy channels. Extensive experiments on real-world medical datasets validate the effectiveness of our framework, demonstrating its potential for robust and efficient telehealth applications.
Authors: Yongchao Huang
Abstract: Standard regression techniques, while powerful, are often constrained by predefined, differentiable loss functions such as mean squared error. These functions may not fully capture the desired behavior of a system, especially when dealing with asymmetric costs or complex, non-differentiable objectives. In this paper, we explore an alternative paradigm: framing regression as a Reinforcement Learning (RL) problem. We demonstrate this by treating a model's prediction as an action and defining a custom reward signal based on the prediction error, and we can leverage powerful RL algorithms to perform function approximation. Through a progressive case study of learning a noisy sine wave, we illustrate the development of an Actor-Critic agent, iteratively enhancing it with Prioritized Experience Replay, increased network capacity, and positional encoding to enable a capable RL agent for this regression task. Our results show that the RL framework not only successfully solves the regression problem but also offers enhanced flexibility in defining objectives and guiding the learning process.
Authors: Adam Block, Cyril Zhang
Abstract: Stochasticity in language model fine-tuning, often caused by the small batch sizes typically used in this regime, can destabilize training by introducing large oscillations in generation quality. A popular approach to mitigating this instability is to take an Exponential moving average (EMA) of weights throughout training. While EMA reduces stochasticity, thereby smoothing training, the introduction of bias from old iterates often creates a lag in optimization relative to vanilla training. In this work, we propose the Bias-Corrected Exponential Moving Average (BEMA), a simple and practical augmentation of EMA that retains variance-reduction benefits while eliminating bias. BEMA is motivated by a simple theoretical model wherein we demonstrate provable acceleration of BEMA over both a standard EMA and vanilla training. Through an extensive suite of experiments on Language Models, we show that BEMA leads to significantly improved convergence rates and final performance over both EMA and vanilla training in a variety of standard LM benchmarks, making BEMA a practical and theoretically motivated intervention for more stable and efficient fine-tuning.
Authors: Mehdi Ben Ayed, Fei Feng, Jay Adams, Vishwakarma Singh, Kritarth Anand, Jiajing Xu
Abstract: Existing web-scale recommendation systems commonly use supervised learning methods that prioritize immediate user feedback. Although reinforcement learning (RL) offers a solution to optimize longer-term goals, such as in-session engagement, applying it at web scale is challenging due to the extremely large action space and engineering complexity. In this paper, we introduce RecoMind, a simulator-based RL framework designed for the effective optimization of session-based goals at web-scale. RecoMind leverages existing recommendation models to establish a simulation environment and to bootstrap the RL policy to optimize immediate user interactions from the outset. This method integrates well with existing industry pipelines, simplifying the training and deployment of RL policies. Additionally, RecoMind introduces a custom exploration strategy to efficiently explore web-scale action spaces with hundreds of millions of items. We evaluated RecoMind through extensive offline simulations and online A/B testing on a video streaming platform. Both methods showed that the RL policy trained using RecoMind significantly outperforms traditional supervised learning recommendation approaches in in-session user satisfaction. In online A/B tests, the RL policy increased videos watched for more than 10 seconds by 15.81\% and improved session depth by 4.71\% for sessions with at least 10 interactions. As a result, RecoMind presents a systematic and scalable approach for embedding RL into web-scale recommendation systems, showing great promise for optimizing session-based user satisfaction.
Authors: Ecem Bozkurt, Antonio Ortega
Abstract: Foundation models (FMs) pretrained on large datasets have become fundamental for various downstream machine learning tasks, in particular in scenarios where obtaining perfectly labeled data is prohibitively expensive. In this paper, we assume an FM has to be fine-tuned with noisy data and present a two-stage framework to ensure robust classification in the presence of label noise without model retraining. Recent work has shown that simple k-nearest neighbor (kNN) approaches using an embedding derived from an FM can achieve good performance even in the presence of severe label noise. Our work is motivated by the fact that these methods make use of local geometry. In this paper, following a similar two-stage procedure, reliability estimation followed by reliability-weighted inference, we show that improved performance can be achieved by introducing geometry information. For a given instance, our proposed inference uses a local neighborhood of training data, obtained using the non-negative kernel (NNK) neighborhood construction. We propose several methods for reliability estimation that can rely less on distance and local neighborhood as the label noise increases. Our evaluation on CIFAR-10 and DermaMNIST shows that our methods improve robustness across various noise conditions, surpassing standard K-NN approaches and recent adaptive-neighborhood baselines.
Authors: Paul Albert, Frederic Z. Zhang, Hemanth Saratchandran, Anton van den Hengel, Ehsan Abbasnejad
Abstract: Parameter-efficient fine-tuning (PEFT) has become a standard approach for adapting large pre-trained models. Amongst PEFT methods, low-rank adaptation (LoRA) has achieved notable success. However, recent studies have highlighted its limitations compared against full-rank alternatives, particularly when applied to multimodal and large language models. In this work, we present a quantitative comparison amongst full-rank and low-rank PEFT methods using a synthetic matrix approximation benchmark with controlled spectral properties. Our results confirm that LoRA struggles to approximate matrices with relatively flat spectrums or high frequency components -- signs of high effective ranks. To this end, we introduce KRAdapter, a novel PEFT algorithm that leverages the Khatri-Rao product to produce weight updates, which, by construction, tends to produce matrix product with a high effective rank. We demonstrate performance gains with KRAdapter on vision-language models up to 1B parameters and on large language models up to 8B parameters, particularly on unseen common-sense reasoning tasks. In addition, KRAdapter maintains the memory and compute efficiency of LoRA, making it a practical and robust alternative to fine-tune billion-scale parameter models.
Authors: Jerry Huang, Peng Lu, Qiuhao Zeng
Abstract: Recent advances in natural language processing (NLP) have opened up greater opportunities to enable fine-tuned large language models (LLMs) to behave as more powerful interactive agents through improved instruction-following ability. However, understanding how this impacts confidence calibration for reliable model output has not been researched in full. In this work, we examine various open-sourced LLMs, identifying significant calibration degradation after instruction tuning in each. Seeking a practical solution, we look towards label smoothing, which has been shown as an effective method to regularize for overconfident predictions but has yet to be widely adopted in the supervised fine-tuning (SFT) of LLMs. We first provide insight as to why label smoothing is sufficient to maintain calibration throughout the SFT process. However, settings remain where the effectiveness of smoothing is severely diminished, in particular the case of large vocabulary LLMs (LV-LLMs). We posit the cause to stem from the ability to become over-confident, which has a direct relationship with the hidden size and vocabulary size, and justify this theoretically and experimentally. Finally, we address an outstanding issue regarding the memory footprint of the cross-entropy loss computation in the label smoothed loss setting, designing a customized kernel to dramatically reduce memory consumption without sacrificing speed or performance in comparison to existing solutions for non-smoothed losses.
Authors: Robin Schmucker, Nimish Pachapurkar, Shanmuga Bala, Miral Shah, Tom Mitchell
Abstract: We present an online tutoring system that learns to provide effective feedback to students after they answer questions incorrectly. Using data from one million students, the system learns which assistance action (e.g., one of multiple hints) to provide for each question to optimize student learning. Employing the multi-armed bandit (MAB) framework and offline policy evaluation, we assess 43,000 assistance actions, and identify trade-offs between assistance policies optimized for different student outcomes (e.g., response correctness, session completion). We design an algorithm that for each question decides on a suitable policy training objective to enhance students' immediate second attempt success and overall practice session performance. We evaluate the resulting MAB policies in 166,000 practice sessions, verifying significant improvements in student outcomes. While MAB policies optimize feedback for the overall student population, we further investigate whether contextual bandit (CB) policies can enhance outcomes by personalizing feedback based on individual student features (e.g., ability estimates, response times). Using causal inference, we examine (i) how effects of assistance actions vary across students and (ii) whether CB policies, which leverage such effect heterogeneity, outperform MAB policies. While our analysis reveals that some actions for some questions exhibit effect heterogeneity, effect sizes may often be too small for CB policies to provide significant improvements beyond what well-optimized MAB policies that deliver the same action to all students already achieve. We discuss insights gained from deploying data-driven systems at scale and implications for future refinements. Today, the teaching policies optimized by our system support thousands of students daily.
Authors: Mohsen Zaker Esteghamati
Abstract: This study presents a methodology to treat performance-based seismic design as an inverse engineering problem, where design parameters are directly derived to achieve specific performance objectives. By implementing explainable machine learning models, this methodology directly maps design variables and performance metrics, tackling computational inefficiencies of performance-based design. The resultant machine learning model is integrated as an evaluation function into a genetic optimization algorithm to solve the inverse problem. The developed methodology is then applied to two different inventories of steel and concrete moment frames in Los Angeles and Charleston to obtain sectional properties of frame members that minimize expected annualized seismic loss in terms of repair costs. The results show high accuracy of the surrogate models (e.g., R2> 90%) across a diverse set of building types, geometries, seismic design, and site hazard, where the optimization algorithm could identify the optimum values of members' properties for a fixed set of geometric variables, consistent with engineering principles.
Authors: Tianyin Liao, Ziwei Zhang, Yufei Sun, Chunyu Hu, Jianxin Li
Abstract: Graph Transformers (GTs) have demonstrated great effectiveness across various graph analytical tasks. However, the existing GTs focus on training and testing graph data originated from the same distribution, but fail to generalize under distribution shifts. Graph invariant learning, aiming to capture generalizable graph structural patterns with labels under distribution shifts, is potentially a promising solution, but how to design attention mechanisms and positional and structural encodings (PSEs) based on graph invariant learning principles remains challenging. To solve these challenges, we introduce Graph Out-Of-Distribution generalized Transformer (GOODFormer), aiming to learn generalized graph representations by capturing invariant relationships between predictive graph structures and labels through jointly optimizing three modules. Specifically, we first develop a GT-based entropy-guided invariant subgraph disentangler to separate invariant and variant subgraphs while preserving the sharpness of the attention function. Next, we design an evolving subgraph positional and structural encoder to effectively and efficiently capture the encoding information of dynamically changing subgraphs during training. Finally, we propose an invariant learning module utilizing subgraph node representations and encodings to derive generalizable graph representations that can to unseen graphs. We also provide theoretical justifications for our method. Extensive experiments on benchmark datasets demonstrate the superiority of our method over state-of-the-art baselines under distribution shifts.
Authors: Yongquan Qu, Matthieu Blanke, Sara Shamekh, Pierre Gentine
Abstract: Earth system modeling presents a fundamental challenge in scientific computing: capturing complex, multiscale nonlinear dynamics in computationally efficient models while minimizing forecast errors caused by necessary simplifications. Even the most powerful AI- or physics-based forecast system suffer from gradual error accumulation. Data assimilation (DA) aims to mitigate these errors by optimally blending (noisy) observations with prior model forecasts, but conventional variational methods often assume Gaussian error statistics that fail to capture the true, non-Gaussian behavior of chaotic dynamical systems. We propose PnP-DA, a Plug-and-Play algorithm that alternates (1) a lightweight, gradient-based analysis update (using a Mahalanobis-distance misfit on new observations) with (2) a single forward pass through a pretrained generative prior conditioned on the background forecast via a conditional Wasserstein coupling. This strategy relaxes restrictive statistical assumptions and leverages rich historical data without requiring an explicit regularization functional, and it also avoids the need to backpropagate gradients through the complex neural network that encodes the prior during assimilation cycles. Experiments on standard chaotic testbeds demonstrate that this strategy consistently reduces forecast errors across a range of observation sparsities and noise levels, outperforming classical variational methods.
Authors: George Wang, Garrett Baker, Andrew Gordon, Daniel Murfet
Abstract: Understanding how language models develop their internal computational structure is a central problem in the science of deep learning. While susceptibilities, drawn from statistical physics, offer a promising analytical tool, their full potential for visualizing network organization remains untapped. In this work, we introduce an embryological approach, applying UMAP to the susceptibility matrix to visualize the model's structural development over training. Our visualizations reveal the emergence of a clear ``body plan,'' charting the formation of known features like the induction circuit and discovering previously unknown structures, such as a ``spacing fin'' dedicated to counting space tokens. This work demonstrates that susceptibility analysis can move beyond validation to uncover novel mechanisms, providing a powerful, holistic lens for studying the developmental principles of complex neural networks.
Authors: Qilin Liao, Shuo Yang, Bo Zhao, Ping Luo, Hengshuang Zhao
Abstract: Harnessing the power of diffusion models to synthesize auxiliary training data based on latent space features has proven effective in enhancing out-of-distribution (OOD) detection performance. However, extracting effective features outside the in-distribution (ID) boundary in latent space remains challenging due to the difficulty of identifying decision boundaries between classes. This paper proposes a novel framework called Boundary-based Out-Of-Distribution data generation (BOOD), which synthesizes high-quality OOD features and generates human-compatible outlier images using diffusion models. BOOD first learns a text-conditioned latent feature space from the ID dataset, selects ID features closest to the decision boundary, and perturbs them to cross the decision boundary to form OOD features. These synthetic OOD features are then decoded into images in pixel space by a diffusion model. Compared to previous works, BOOD provides a more training efficient strategy for synthesizing informative OOD features, facilitating clearer distinctions between ID and OOD data. Extensive experimental results on common benchmarks demonstrate that BOOD surpasses the state-of-the-art method significantly, achieving a 29.64% decrease in average FPR95 (40.31% vs. 10.67%) and a 7.27% improvement in average AUROC (90.15% vs. 97.42%) on the CIFAR-100 dataset.
Authors: Yoonhyuk Choi, Jiho Choi, Chong-Kwon Kim
Abstract: Over-smoothing in Graph Neural Networks (GNNs) causes collapse in distinct node features, particularly on heterophilic graphs where adjacent nodes often have dissimilar labels. Although sheaf neural networks partially mitigate this problem, they typically rely on static or heavily parameterized sheaf structures that hinder generalization and scalability. Existing sheaf-based models either predefine restriction maps or introduce excessive complexity, yet fail to provide rigorous stability guarantees. In this paper, we introduce a novel scheme called SGPC (Sheaf GNNs with PAC-Bayes Calibration), a unified architecture that combines cellular-sheaf message passing with several mechanisms, including optimal transport-based lifting, variance-reduced diffusion, and PAC-Bayes spectral regularization for robust semi-supervised node classification. We establish performance bounds theoretically and demonstrate that the resulting bound-aware objective can be achieved via end-to-end training in linear computational complexity. Experiments on nine homophilic and heterophilic benchmarks show that SGPC outperforms state-of-the-art spectral and sheaf-based GNNs while providing certified confidence intervals on unseen nodes.
Authors: Chanyoung Yoon, Sangbong Yoo, Soobin Yim, Chansoo Kim, Yun Jang
Abstract: Designing residential interiors strongly impacts occupant satisfaction but remains challenging due to unstructured spatial layouts, high computational demands, and reliance on expert knowledge. Existing methods based on optimization or deep learning are either computationally expensive or constrained by data scarcity. Reinforcement learning (RL) approaches often limit furniture placement to discrete positions and fail to incorporate design principles adequately. We propose OID-PPO, a novel RL framework for Optimal Interior Design using Proximal Policy Optimization, which integrates expert-defined functional and visual guidelines into a structured reward function. OID-PPO utilizes a diagonal Gaussian policy for continuous and flexible furniture placement, effectively exploring latent environmental dynamics under partial observability. Experiments conducted across diverse room shapes and furniture configurations demonstrate that OID-PPO significantly outperforms state-of-the-art methods in terms of layout quality and computational efficiency. Ablation studies further demonstrate the impact of structured guideline integration and reveal the distinct contributions of individual design constraints.
Authors: Lijun Zhang, Wenhao Yang, Guanghui Wang, Wei Jiang, Zhi-Hua Zhou
Abstract: To deal with changing environments, a new performance measure -- adaptive regret, defined as the maximum static regret over any interval, was proposed in online learning. Under the setting of online convex optimization, several algorithms have been successfully developed to minimize the adaptive regret. However, existing algorithms lack universality in the sense that they can only handle one type of convex functions and need apriori knowledge of parameters, which hinders their application in real-world scenarios. To address this limitation, this paper investigates universal algorithms with dual adaptivity, which automatically adapt to the property of functions (convex, exponentially concave, or strongly convex), as well as the nature of environments (stationary or changing). Specifically, we propose a meta-expert framework for dual adaptive algorithms, where multiple experts are created dynamically and aggregated by a meta-algorithm. The meta-algorithm is required to yield a second-order bound, which can accommodate unknown function types. We further incorporate the technique of sleeping experts to capture the changing environments. For the construction of experts, we introduce two strategies (increasing the number of experts or enhancing the capabilities of experts) to achieve universality. Theoretical analysis shows that our algorithms are able to minimize the adaptive regret for multiple types of convex functions simultaneously, and also allow the type of functions to switch between rounds. Moreover, we extend our meta-expert framework to online composite optimization, and develop a universal algorithm for minimizing the adaptive regret of composite functions.
Authors: Antonis Klironomos, Baifan Zhou, Zhipeng Tan, Zhuoxun Zheng, Mohamed H. Gad-Elrab, Heiko Paulheim, Evgeny Kharlamov
Abstract: Nowadays machine learning (ML) practitioners have access to numerous ML libraries available online. Such libraries can be used to create ML pipelines that consist of a series of steps where each step may invoke up to several ML libraries that are used for various data-driven analytical tasks. Development of high-quality ML pipelines is non-trivial; it requires training, ML expertise, and careful development of each step. At the same time, domain experts in science and engineering may not possess such ML expertise and training while they are in pressing need of ML-based analytics. In this paper, we present our ExeKGLib, a Python library enhanced with a graphical interface layer that allows users with minimal ML knowledge to build ML pipelines. This is achieved by relying on knowledge graphs that encode ML knowledge in simple terms accessible to non-ML experts. ExeKGLib also allows improving the transparency and reusability of the built ML workflows and ensures that they are executable. We show the usability and usefulness of ExeKGLib by presenting real use cases.
Authors: Zizhuo Zhang, Jianing Zhu, Xinmu Ge, Zihua Zhao, Zhanke Zhou, Xuan Li, Xiao Feng, Jiangchao Yao, Bo Han
Abstract: Although reinforcement learning with verifiable rewards (RLVR) shows promise in improving the reasoning ability of large language models (LLMs), the scaling up dilemma remains due to the reliance on human annotated labels especially for complex tasks. Recent alternatives that explore various self-reward signals exhibit the eliciting potential of LLM reasoning, but suffer from the non-negligible collapse issue. Inspired by the success of self-supervised learning, we propose \textit{Co-Reward}, a novel RL framework that leverages contrastive agreement across semantically analogical questions as a reward basis. Specifically, we construct a similar question for each training sample (without labels) and synthesize their individual surrogate labels through a simple rollout voting, and then the reward is constructed by cross-referring the labels of each question pair to enforce the internal reasoning consistency across analogical inputs. Intuitively, such a self-supervised reward-shaping mechanism increases the difficulty of learning collapse into a trivial solution, and promotes stable reasoning elicitation and improvement through expanding the input sample variants. Empirically, Co-Reward achieves superior performance compared to other self-reward baselines on multiple reasoning benchmarks and LLM series, and reaches or even surpasses ground-truth (GT) labeled reward, with improvements of up to $+6.8\%$ on MATH500 over GT reward on Llama-3.2-3B-Instruct. Our code is publicly available at https://github.com/tmlr-group/Co-Reward.
Authors: Yue Yang, Yuxiang Lin, Ying Zhang, Zihan Su, Chang Chuan Goh, Tangtangfang Fang, Anthony Graham Bellotti, Boon Giin Lee
Abstract: Prediction of post-loan default is an important task in credit risk management, and can be addressed by detection of financial anomalies using machine learning. This study introduces a ResE-BiLSTM model, using a sliding window technique, and is evaluated on 44 independent cohorts from the extensive Freddie Mac US mortgage dataset, to improve prediction performance. The ResE-BiLSTM is compared with five baseline models: Long Short-Term Memory (LSTM), BiLSTM, Gated Recurrent Units (GRU), Convolutional Neural Networks (CNN), and Recurrent Neural Networks (RNN), across multiple metrics, including Accuracy, Precision, Recall, F1, and AUC. An ablation study was conducted to evaluate the contribution of individual components in the ResE-BiLSTM architecture. Additionally, SHAP analysis was employed to interpret the underlying features the model relied upon for its predictions. Experimental results demonstrate that ResE-BiLSTM achieves superior predictive performance compared to baseline models, underscoring its practical value and applicability in real-world scenarios.
Authors: Leonidas Akritidis, Panayiotis Bozanis
Abstract: The tabular form constitutes the standard way of representing data in relational database systems and spreadsheets. But, similarly to other forms, tabular data suffers from class imbalance, a problem that causes serious performance degradation in a wide variety of machine learning tasks. One of the most effective solutions dictates the usage of Generative Adversarial Networks (GANs) in order to synthesize artificial data instances for the under-represented classes. Despite their good performance, none of the proposed GAN models takes into account the vector subspaces of the input samples in the real data space, leading to data generation in arbitrary locations. Moreover, the class labels are treated in the same manner as the other categorical variables during training, so conditional sampling by class is rendered less effective. To overcome these problems, this study presents ctdGAN, a conditional GAN for alleviating class imbalance in tabular datasets. Initially, ctdGAN executes a space partitioning step to assign cluster labels to the input samples. Subsequently, it utilizes these labels to synthesize samples via a novel probabilistic sampling strategy and a new loss function that penalizes both cluster and class mis-predictions. In this way, ctdGAN is trained to generate samples in subspaces that resemble those of the original data distribution. We also introduce several other improvements, including a simple, yet effective cluster-wise scaling technique that captures multiple feature modes without affecting data dimensionality. The exhaustive evaluation of ctdGAN with 14 imbalanced datasets demonstrated its superiority in generating high fidelity samples and improving classification accuracy.
Authors: Yiming Xu, Jiarun Chen, Zhen Peng, Zihan Chen, Qika Lin, Lan Ma, Bin Shi, Bo Dong
Abstract: The natural combination of intricate topological structures and rich textual information in text-attributed graphs (TAGs) opens up a novel perspective for graph anomaly detection (GAD). However, existing GAD methods primarily focus on designing complex optimization objectives within the graph domain, overlooking the complementary value of the textual modality, whose features are often encoded by shallow embedding techniques, such as bag-of-words or skip-gram, so that semantic context related to anomalies may be missed. To unleash the enormous potential of textual modality, large language models (LLMs) have emerged as promising alternatives due to their strong semantic understanding and reasoning capabilities. Nevertheless, their application to TAG anomaly detection remains nascent, and they struggle to encode high-order structural information inherent in graphs due to input length constraints. For high-quality anomaly detection in TAGs, we propose CoLL, a novel framework that combines LLMs and graph neural networks (GNNs) to leverage their complementary strengths. CoLL employs multi-LLM collaboration for evidence-augmented generation to capture anomaly-relevant contexts while delivering human-readable rationales for detected anomalies. Moreover, CoLL integrates a GNN equipped with a gating mechanism to adaptively fuse textual features with evidence while preserving high-order topological information. Extensive experiments demonstrate the superiority of CoLL, achieving an average improvement of 13.37% in AP. This study opens a new avenue for incorporating LLMs in advancing GAD.
Authors: Yiming Xu, Xu Hua, Zhen Peng, Bin Shi, Jiarun Chen, Xingbo Fu, Song Wang, Bo Dong
Abstract: The widespread application of graph data in various high-risk scenarios has increased attention to graph anomaly detection (GAD). Faced with real-world graphs that often carry node descriptions in the form of raw text sequences, termed text-attributed graphs (TAGs), existing graph anomaly detection pipelines typically involve shallow embedding techniques to encode such textual information into features, and then rely on complex self-supervised tasks within the graph domain to detect anomalies. However, this text encoding process is separated from the anomaly detection training objective in the graph domain, making it difficult to ensure that the extracted textual features focus on GAD-relevant information, seriously constraining the detection capability. How to seamlessly integrate raw text and graph topology to unleash the vast potential of cross-modal data in TAGs for anomaly detection poses a challenging issue. This paper presents a novel end-to-end paradigm for text-attributed graph anomaly detection, named CMUCL. We simultaneously model data from both text and graph structures, and jointly train text and graph encoders by leveraging cross-modal and uni-modal multi-scale consistency to uncover potential anomaly-related information. Accordingly, we design an anomaly score estimator based on inconsistency mining to derive node-specific anomaly scores. Considering the lack of benchmark datasets tailored for anomaly detection on TAGs, we release 8 datasets to facilitate future research. Extensive evaluations show that CMUCL significantly advances in text-attributed graph anomaly detection, delivering an 11.13% increase in average accuracy (AP) over the suboptimal.
Authors: Sifan Yang, Yuanyu Wan, Lijun Zhang
Abstract: We investigate the online nonsubmodular optimization with delayed feedback in the bandit setting, where the loss function is $\alpha$-weakly DR-submodular and $\beta$-weakly DR-supermodular. Previous work has established an $(\alpha,\beta)$-regret bound of $\mathcal{O}(nd^{1/3}T^{2/3})$, where $n$ is the dimensionality and $d$ is the maximum delay. However, its regret bound relies on the maximum delay and is thus sensitive to irregular delays. Additionally, it couples the effects of delays and bandit feedback as its bound is the product of the delay term and the $\mathcal{O}(nT^{2/3})$ regret bound in the bandit setting without delayed feedback. In this paper, we develop two algorithms to address these limitations, respectively. Firstly, we propose a novel method, namely DBGD-NF, which employs the one-point gradient estimator and utilizes all the available estimated gradients in each round to update the decision. It achieves a better $\mathcal{O}(n\bar{d}^{1/3}T^{2/3})$ regret bound, which is relevant to the average delay $\bar{d} = \frac{1}{T}\sum_{t=1}^T d_t\leq d$. Secondly, we extend DBGD-NF by employing a blocking update mechanism to decouple the joint effect of the delays and bandit feedback, which enjoys an $\mathcal{O}(n(T^{2/3} + \sqrt{dT}))$ regret bound. When $d = \mathcal{O}(T^{1/3})$, our regret bound matches the $\mathcal{O}(nT^{2/3})$ bound in the bandit setting without delayed feedback. Compared to our first $\mathcal{O}(n\bar{d}^{1/3}T^{2/3})$ bound, it is more advantageous when the maximum delay $d = o(\bar{d}^{2/3}T^{1/3})$. Finally, we conduct experiments on structured sparse learning to demonstrate the superiority of our methods.
Authors: Judy X Yang
Abstract: Hyperspectral imaging offers detailed spectral information for mineral mapping; however, weak mineral signatures are often masked by noisy and redundant bands, limiting detection performance. To address this, we propose a two-stage integrated framework for enhanced mineral detection in the Cuprite mining district. In the first stage, we compute the signal-to-noise ratio (SNR) for each spectral band and apply a phase-locked thresholding technique to discard low-SNR bands, effectively removing redundancy and suppressing background noise. Savitzky-Golay filtering is then employed for spectral smoothing, serving a dual role first to stabilize trends during band selection, and second to preserve fine-grained spectral features during preprocessing. In the second stage, the refined HSI data is reintroduced into the model, where KMeans clustering is used to extract 12 endmember spectra (W1 custom), followed by non negative least squares (NNLS) for abundance unmixing. The resulting endmembers are quantitatively compared with laboratory spectra (W1 raw) using cosine similarity and RMSE metrics. Experimental results confirm that our proposed pipeline improves unmixing accuracy and enhances the detection of weak mineral zones. This two-pass strategy demonstrates a practical and reproducible solution for spectral dimensionality reduction and unmixing in geological HSI applications.
Authors: Pietro Barbiero, Mateo Espinosa Zarlenga, Alberto Termine, Mateja Jamnik, Giuseppe Marra
Abstract: We argue that existing definitions of interpretability are not actionable in that they fail to inform users about general, sound, and robust interpretable model design. This makes current interpretability research fundamentally ill-posed. To address this issue, we propose a definition of interpretability that is general, simple, and subsumes existing informal notions within the interpretable AI community. We show that our definition is actionable, as it directly reveals the foundational properties, underlying assumptions, principles, data structures, and architectural features necessary for designing interpretable models. Building on this, we propose a general blueprint for designing interpretable models and introduce the first open-sourced library with native support for interpretable data structures and processes.
Authors: Marlen Neubert, Patrick Reiser, Frauke Gr\"ater, Pascal Friederich
Abstract: Hydrogen atom transfer (HAT) reactions are essential in many biological processes, such as radical migration in damaged proteins, but their mechanistic pathways remain incompletely understood. Simulating HAT is challenging due to the need for quantum chemical accuracy at biologically relevant scales; thus, neither classical force fields nor DFT-based molecular dynamics are applicable. Machine-learned potentials offer an alternative, able to learn potential energy surfaces (PESs) with near-quantum accuracy. However, training these models to generalize across diverse HAT configurations, especially at radical positions in proteins, requires tailored data generation and careful model selection. Here, we systematically generate HAT configurations in peptides to build large datasets using semiempirical methods and DFT. We benchmark three graph neural network architectures (SchNet, Allegro, and MACE) on their ability to learn HAT PESs and indirectly predict reaction barriers from energy predictions. MACE consistently outperforms the others in energy, force, and barrier prediction, achieving a mean absolute error of 1.13 kcal/mol on out-of-distribution DFT barrier predictions. This accuracy enables integration of ML potentials into large-scale collagen simulations to compute reaction rates from predicted barriers, advancing mechanistic understanding of HAT and radical migration in peptides. We analyze scaling laws, model transferability, and cost-performance trade-offs, and outline strategies for improvement by combining ML potentials with transition state search algorithms and active learning. Our approach is generalizable to other biomolecular systems, enabling quantum-accurate simulations of chemical reactivity in complex environments.
Authors: Thorben Werner, Lars Schmidt-Thieme, Vijaya Krishna Yalavarthi
Abstract: Even though Active Learning (AL) is widely studied, it is rarely applied in contexts outside its own scientific literature. We posit that the reason for this is AL's high computational cost coupled with the comparatively small lifts it is typically able to generate in scenarios with few labeled points. In this work we study the impact of different methods to combat this low data scenario, namely data augmentation (DA), semi-supervised learning (SSL) and AL. We find that AL is by far the least efficient method of solving the low data problem, generating a lift of only 1-4\% over random sampling, while DA and SSL methods can generate up to 60\% lift in combination with random sampling. However, when AL is combined with strong DA and SSL techniques, it surprisingly is still able to provide improvements. Based on these results, we frame AL not as a method to combat missing labels, but as the final building block to squeeze the last bits of performance out of data after appropriate DA and SSL methods as been applied.
Authors: Mukesh Kumar Sahu, Pinki Roy
Abstract: Accurately predicting the criticalness of ICU patients (such as in-ICU mortality risk) is vital for early intervention in critical care. However, conventional models often treat each patient in isolation and struggle to exploit the relational structure in Electronic Health Records (EHR). We propose a Similarity-Based Self-Construct Graph Model (SBSCGM) that dynamically builds a patient similarity graph from multi-modal EHR data, and a HybridGraphMedGNN architecture that operates on this graph to predict patient mortality and a continuous criticalness score. SBSCGM uses a hybrid similarity measure (combining feature-based and structural similarities) to connect patients with analogous clinical profiles in real-time. The HybridGraphMedGNN integrates Graph Convolutional Network (GCN), GraphSAGE, and Graph Attention Network (GAT) layers to learn robust patient representations, leveraging both local and global graph patterns. In experiments on 6,000 ICU stays from the MIMIC-III dataset, our model achieves state-of-the-art performance (AUC-ROC $0.94$) outperforming baseline classifiers and single-type GNN models. We also demonstrate improved precision/recall and show that the attention mechanism provides interpretable insights into model predictions. Our framework offers a scalable and interpretable solution for critical care risk prediction, with potential to support clinicians in real-world ICU deployment.
Authors: Paul Tresson, Pierre Le Coz, Hadrien Tulet, Anthony Malkassian, Maxime R\'ejou M\'echain
Abstract: Remote sensing has entered a new era with the rapid development of artificial intelligence approaches. However, the implementation of deep learning has largely remained restricted to specialists and has been impractical because it often requires (i) large reference datasets for model training and validation; (ii) substantial computing resources; and (iii) strong coding skills. Here, we introduce IAMAP, a user-friendly QGIS plugin that addresses these three challenges in an easy yet flexible way. IAMAP builds on recent advancements in self-supervised learning strategies, which now provide robust feature extractors, often referred to as foundation models. These generalist models can often be reliably used in few-shot or zero-shot scenarios (i.e., with little to no fine-tuning). IAMAP's interface allows users to streamline several key steps in remote sensing image analysis: (i) extracting image features using a wide range of deep learning architectures; (ii) reducing dimensionality with built-in algorithms; (iii) performing clustering on features or their reduced representations; (iv) generating feature similarity maps; and (v) calibrating and validating supervised machine learning models for prediction. By enabling non-AI specialists to leverage the high-quality features provided by recent deep learning approaches without requiring GPU capacity or extensive reference datasets, IAMAP contributes to the democratization of computationally efficient and energy-conscious deep learning methods.
Authors: Xiong Xiong, Zhuo Zhang, Rongchun Hu, Chen Gao, Zichen Deng
Abstract: Solving high-frequency oscillatory partial differential equations (PDEs) is a critical challenge in scientific computing, with applications in fluid mechanics, quantum mechanics, and electromagnetic wave propagation. Traditional physics-informed neural networks (PINNs) suffer from spectral bias, limiting their ability to capture high-frequency solution components. We introduce Separated-Variable Spectral Neural Networks (SV-SNN), a novel framework that addresses these limitations by integrating separation of variables with adaptive spectral methods. Our approach features three key innovations: (1) decomposition of multivariate functions into univariate function products, enabling independent spatial and temporal networks; (2) adaptive Fourier spectral features with learnable frequency parameters for high-frequency capture; and (3) theoretical framework based on singular value decomposition to quantify spectral bias. Comprehensive evaluation on benchmark problems including Heat equation, Helmholtz equation, Poisson equations and Navier-Stokes equations demonstrates that SV-SNN achieves 1-3 orders of magnitude improvement in accuracy while reducing parameter count by over 90\% and training time by 60\%. These results establish SV-SNN as an effective solution to the spectral bias problem in neural PDE solving. The implementation will be made publicly available upon acceptance at https://github.com/xgxgnpu/SV-SNN.
Authors: Changning Wu, Gao Wu, Rongyao Cai, Yong Liu, Kexin Zhang
Abstract: Multi-scale decomposition architectures have emerged as predominant methodologies in time series forecasting. However, real-world time series exhibit noise interference across different scales, while heterogeneous information distribution among frequency components at varying scales leads to suboptimal multi-scale representation. Inspired by Kolmogorov-Arnold Networks (KAN) and Parseval's theorem, we propose a KAN based adaptive Frequency Selection learning architecture (KFS) to address these challenges. This framework tackles prediction challenges stemming from cross-scale noise interference and complex pattern modeling through its FreK module, which performs energy-distribution-based dominant frequency selection in the spectral domain. Simultaneously, KAN enables sophisticated pattern representation while timestamp embedding alignment synchronizes temporal representations across scales. The feature mixing module then fuses scale-specific patterns with aligned temporal features. Extensive experiments across multiple real-world time series datasets demonstrate that KT achieves state-of-the-art performance as a simple yet effective architecture.
Authors: Alessandro Palmas
Abstract: The growing threat of low-cost kamikaze drone swarms poses a critical challenge to modern defense systems demanding rapid and strategic decision-making to prioritize interceptions across multiple effectors and high-value target zones. In this work, we present a case study demonstrating the practical advantages of reinforcement learning in addressing this challenge. We introduce a high-fidelity simulation environment that captures realistic operational constraints, within which a decision-level reinforcement learning agent learns to coordinate multiple effectors for optimal interception prioritization. Operating in a discrete action space, the agent selects which drone to engage per effector based on observed state features such as positions, classes, and effector status. We evaluate the learned policy against a handcrafted rule-based baseline across hundreds of simulated attack scenarios. The reinforcement learning based policy consistently achieves lower average damage and higher defensive efficiency in protecting critical zones. This case study highlights the potential of reinforcement learning as a strategic layer within defense architectures, enhancing resilience without displacing existing control systems. All code and simulation assets are publicly released for full reproducibility, and a video demonstration illustrates the policy's qualitative behavior.
Authors: Albert Matveev, Sanmitra Ghosh, Aamal Hussain, James-Michael Leahy, Michalis Michaelides
Abstract: Operator learning is a powerful paradigm for solving partial differential equations, with Fourier Neural Operators serving as a widely adopted foundation. However, FNOs face significant scalability challenges due to overparameterization and offer no native uncertainty quantification -- a key requirement for reliable scientific and engineering applications. Instead, neural operators rely on post hoc UQ methods that ignore geometric inductive biases. In this work, we introduce DINOZAUR: a diffusion-based neural operator parametrization with uncertainty quantification. Inspired by the structure of the heat kernel, DINOZAUR replaces the dense tensor multiplier in FNOs with a dimensionality-independent diffusion multiplier that has a single learnable time parameter per channel, drastically reducing parameter count and memory footprint without compromising predictive performance. By defining priors over those time parameters, we cast DINOZAUR as a Bayesian neural operator to yield spatially correlated outputs and calibrated uncertainty estimates. Our method achieves competitive or superior performance across several PDE benchmarks while providing efficient uncertainty quantification.
Authors: Sihang Zeng, Lucas Jing Liu, Jun Wen, Meliha Yetisgen, Ruth Etzioni, Gang Luo
Abstract: Trustworthy survival prediction is essential for clinical decision making. Longitudinal electronic health records (EHRs) provide a uniquely powerful opportunity for the prediction. However, it is challenging to accurately model the continuous clinical progression of patients underlying the irregularly sampled clinical features and to transparently link the progression to survival outcomes. To address these challenges, we develop TrajSurv, a model that learns continuous latent trajectories from longitudinal EHR data for trustworthy survival prediction. TrajSurv employs a neural controlled differential equation (NCDE) to extract continuous-time latent states from the irregularly sampled data, forming continuous latent trajectories. To ensure the latent trajectories reflect the clinical progression, TrajSurv aligns the latent state space with patient state space through a time-aware contrastive learning approach. To transparently link clinical progression to the survival outcome, TrajSurv uses latent trajectories in a two-step divide-and-conquer interpretation process. First, it explains how the changes in clinical features translate into the latent trajectory's evolution using a learned vector field. Second, it clusters these latent trajectories to identify key clinical progression patterns associated with different survival outcomes. Evaluations on two real-world medical datasets, MIMIC-III and eICU, show TrajSurv's competitive accuracy and superior transparency over existing deep learning methods.
Authors: Jialun Zheng, Jie Liu, Jiannong Cao, Xiao Wang, Hanchen Yang, Yankai Chen, Philip S. Yu
Abstract: Dynamic graph anomaly detection (DGAD) is essential for identifying anomalies in evolving graphs across domains such as finance, traffic, and social networks. Recently, generalist graph anomaly detection (GAD) models have shown promising results. They are pretrained on multiple source datasets and generalize across domains. While effective on static graphs, they struggle to capture evolving anomalies in dynamic graphs. Moreover, the continuous emergence of new domains and the lack of labeled data further challenge generalist DGAD. Effective cross-domain DGAD requires both domain-specific and domain-agnostic anomalous patterns. Importantly, these patterns evolve temporally within and across domains. Building on these insights, we propose a DGAD model with Dynamic Prototypes (DP) to capture evolving domain-specific and domain-agnostic patterns. Firstly, DP-DGAD extracts dynamic prototypes, i.e., evolving representations of normal and anomalous patterns, from temporal ego-graphs and stores them in a memory buffer. The buffer is selectively updated to retain general, domain-agnostic patterns while incorporating new domain-specific ones. Then, an anomaly scorer compares incoming data with dynamic prototypes to flag both general and domain-specific anomalies. Finally, DP-DGAD employs confidence-based pseudo-labeling for effective self-supervised adaptation in target domains. Extensive experiments demonstrate state-of-the-art performance across ten real-world datasets from different domains.
Authors: Young-ho Cho, Hao Zhu, Duehee Lee, Ross Baldick
Abstract: For conducting resource adequacy studies, we synthesize multiple long-term wind power scenarios of distributed wind farms simultaneously by using the spatio-temporal features: spatial and temporal correlation, waveforms, marginal and ramp rates distributions of waveform, power spectral densities, and statistical characteristics. Generating the spatial correlation in scenarios requires the design of common factors for neighboring wind farms and antithetical factors for distant wind farms. The generalized dynamic factor model (GDFM) can extract the common factors through cross spectral density analysis, but it cannot closely imitate waveforms. The GAN can synthesize plausible samples representing the temporal correlation by verifying samples through a fake sample discriminator. To combine the advantages of GDFM and GAN, we use the GAN to provide a filter that extracts dynamic factors with temporal information from the observation data, and we then apply this filter in the GDFM to represent both spatial and frequency correlations of plausible waveforms. Numerical tests on the combination of GDFM and GAN have demonstrated performance improvements over competing alternatives in synthesizing wind power scenarios from Australia, better realizing plausible statistical characteristics of actual wind power compared to alternatives such as the GDFM with a filter synthesized from distributions of actual dynamic filters and the GAN with direct synthesis without dynamic factors.
Authors: Sergio Rubio-Mart\'in, Mar\'ia Teresa Garc\'ia-Ord\'as, Antonio Serrano-Garc\'ia, Clara Margarita Franch-Pato, Arturo Crespo-\'Alvaro, Jos\'e Alberto Ben\'itez-Andrades
Abstract: The classification of clinical notes into specific diagnostic categories is critical in healthcare, especially for mental health conditions like Anxiety and Adjustment Disorder. In this study, we compare the performance of various Artificial Intelligence models, including both traditional Machine Learning approaches (Random Forest, Support Vector Machine, K-nearest neighbors, Decision Tree, and eXtreme Gradient Boost) and Deep Learning models (DistilBERT and SciBERT), to classify clinical notes into these two diagnoses. Additionally, we implemented three oversampling strategies: No Oversampling, Random Oversampling, and Synthetic Minority Oversampling Technique (SMOTE), to assess their impact on model performance. Hyperparameter tuning was also applied to optimize model accuracy. Our results indicate that oversampling techniques had minimal impact on model performance overall. The only exception was SMOTE, which showed a positive effect specifically with BERT-based models. However, hyperparameter optimization significantly improved accuracy across the models, enhancing their ability to generalize and perform on the dataset. The Decision Tree and eXtreme Gradient Boost models achieved the highest accuracy among machine learning approaches, both reaching 96%, while the DistilBERT and SciBERT models also attained 96% accuracy in the deep learning category. These findings underscore the importance of hyperparameter tuning in maximizing model performance. This study contributes to the ongoing research on AI-assisted diagnostic tools in mental health by providing insights into the efficacy of different model architectures and data balancing methods.
Authors: Haozhe Tian, Pietro Ferraro, Robert Shorten, Mahdi Jalili, Homayoun Hamedmoghadam
Abstract: The application of message-passing Graph Neural Networks has been a breakthrough for important network science problems. However, the competitive performance often relies on using handcrafted structural features as inputs, which increases computational cost and introduces bias into the otherwise purely data-driven network representations. Here, we eliminate the need for handcrafted features by introducing an attention mechanism and utilizing message-iteration profiles, in addition to an effective algorithmic approach to generate a structurally diverse training set of small synthetic networks. Thereby, we build an expressive message-passing framework and use it to efficiently solve the NP-hard problem of Network Dismantling, virtually equivalent to vital node identification, with significant real-world applications. Trained solely on diversified synthetic networks, our proposed model -- MIND: Message Iteration Network Dismantler -- generalizes to large, unseen real networks with millions of nodes, outperforming state-of-the-art network dismantling methods. Increased efficiency and generalizability of the proposed model can be leveraged beyond dismantling in a range of complex network problems.
Authors: Yannik Schnitzer, Alessandro Abate, David Parker
Abstract: Robust Markov decision processes (r-MDPs) extend MDPs by explicitly modelling epistemic uncertainty about transition dynamics. Learning r-MDPs from interactions with an unknown environment enables the synthesis of robust policies with provable (PAC) guarantees on performance, but this can require a large number of sample interactions. We propose novel methods for solving and learning r-MDPs based on factored state-space representations that leverage the independence between model uncertainty across system components. Although policy synthesis for factored r-MDPs leads to hard, non-convex optimisation problems, we show how to reformulate these into tractable linear programs. Building on these, we also propose methods to learn factored model representations directly. Our experimental results show that exploiting factored structure can yield dimensional gains in sample efficiency, producing more effective robust policies with tighter performance guarantees than state-of-the-art methods.
Authors: Dien Nguyen, Diego Perez-Liebana, Simon Lucas
Abstract: We introduce JSON Bag-of-Tokens model (JSON-Bag) as a method to generically represent game trajectories by tokenizing their JSON descriptions and apply Jensen-Shannon distance (JSD) as distance metric for them. Using a prototype-based nearest-neighbor search (P-NNS), we evaluate the validity of JSON-Bag with JSD on six tabletop games -- \textit{7 Wonders}, \textit{Dominion}, \textit{Sea Salt and Paper}, \textit{Can't Stop}, \textit{Connect4}, \textit{Dots and boxes} -- each over three game trajectory classification tasks: classifying the playing agents, game parameters, or game seeds that were used to generate the trajectories. Our approach outperforms a baseline using hand-crafted features in the majority of tasks. Evaluating on N-shot classification suggests using JSON-Bag prototype to represent game trajectory classes is also sample efficient. Additionally, we demonstrate JSON-Bag ability for automatic feature extraction by treating tokens as individual features to be used in Random Forest to solve the tasks above, which significantly improves accuracy on underperforming tasks. Finally, we show that, across all six games, the JSD between JSON-Bag prototypes of agent classes highly correlates with the distances between agents' policies.
Authors: Yingxu Wang, Mengzhu Wang, Zhichao Huang, Suyu Liu
Abstract: Graph Domain Adaptation (GDA) facilitates knowledge transfer from labeled source graphs to unlabeled target graphs by learning domain-invariant representations, which is essential in applications such as molecular property prediction and social network analysis. However, most existing GDA methods rely on the assumption of clean source labels, which rarely holds in real-world scenarios where annotation noise is pervasive. This label noise severely impairs feature alignment and degrades adaptation performance under domain shifts. To address this challenge, we propose Nested Graph Pseudo-Label Refinement (NeGPR), a novel framework tailored for graph-level domain adaptation with noisy labels. NeGPR first pretrains dual branches, i.e., semantic and topology branches, by enforcing neighborhood consistency in the feature space, thereby reducing the influence of noisy supervision. To bridge domain gaps, NeGPR employs a nested refinement mechanism in which one branch selects high-confidence target samples to guide the adaptation of the other, enabling progressive cross-domain learning. Furthermore, since pseudo-labels may still contain noise and the pre-trained branches are already overfitted to the noisy labels in the source domain, NeGPR incorporates a noise-aware regularization strategy. This regularization is theoretically proven to mitigate the adverse effects of pseudo-label noise, even under the presence of source overfitting, thus enhancing the robustness of the adaptation process. Extensive experiments on benchmark datasets demonstrate that NeGPR consistently outperforms state-of-the-art methods under severe label noise, achieving gains of up to 12.7% in accuracy.
Authors: Ivona Krchova, Mariana Vargas Vieyra, Mario Scriminaci, Andrey Sidorenko
Abstract: Machine learning development critically depends on access to high-quality data. However, increasing restrictions due to privacy, proprietary interests, and ethical concerns have created significant barriers to data accessibility. Synthetic data offers a viable solution by enabling safe, broad data usage without compromising sensitive information. This paper presents the MOSTLY AI Synthetic Data Software Development Kit (SDK), an open-source toolkit designed specifically for synthesizing high-quality tabular data. The SDK integrates robust features such as differential privacy guarantees, fairness-aware data generation, and automated quality assurance into a flexible and accessible Python interface. Leveraging the TabularARGN autoregressive framework, the SDK supports diverse data types and complex multi-table and sequential datasets, delivering competitive performance with notable improvements in speed and usability. Currently deployed both as a cloud service and locally installable software, the SDK has seen rapid adoption, highlighting its practicality in addressing real-world data bottlenecks and promoting widespread data democratization.
Authors: Liuyun Xu, Seymour M. J. Spence
Abstract: Existing variance reduction techniques used in stochastic simulations for rare event analysis still require a substantial number of model evaluations to estimate small failure probabilities. In the context of complex, nonlinear finite element modeling environments, this can become computationally challenging-particularly for systems subjected to stochastic excitation. To address this challenge, a multi-fidelity stratified sampling scheme with adaptive machine learning metamodels is introduced for efficiently propagating uncertainties and estimating small failure probabilities. In this approach, a high-fidelity dataset generated through stratified sampling is used to train a deep learning-based metamodel, which then serves as a cost-effective and highly correlated low-fidelity model. An adaptive training scheme is proposed to balance the trade-off between approximation quality and computational demand associated with the development of the low-fidelity model. By integrating the low-fidelity outputs with additional high-fidelity results, an unbiased estimate of the strata-wise failure probabilities is obtained using a multi-fidelity Monte Carlo framework. The overall probability of failure is then computed using the total probability theorem. Application to a full-scale high-rise steel building subjected to stochastic wind excitation demonstrates that the proposed scheme can accurately estimate exceedance probability curves for nonlinear responses of interest, while achieving significant computational savings compared to single-fidelity variance reduction approaches.
Authors: Yaxin Ma, Benjamin Colburn, Jose C. Principe
Abstract: Bayesian neural networks and deep ensemble methods have been proposed for uncertainty quantification; however, they are computationally intensive and require large storage. By utilizing a single deterministic model, we can solve the above issue. We propose an effective method based on feature space density to quantify uncertainty for distributional shifts and out-of-distribution (OOD) detection. Specifically, we leverage the information potential field derived from kernel density estimation to approximate the feature space density of the training set. By comparing this density with the feature space representation of test samples, we can effectively determine whether a distributional shift has occurred. Experiments were conducted on a 2D synthetic dataset (Two Moons and Three Spirals) as well as an OOD detection task (CIFAR-10 vs. SVHN). The results demonstrate that our method outperforms baseline models.
Authors: Timur Sattarov, Marco Schreyer, Damian Borth
Abstract: Anomaly detection in tabular data remains challenging due to complex feature interactions and the scarcity of anomalous examples. Denoising autoencoders rely on fixed-magnitude noise, limiting adaptability to diverse data distributions. Diffusion models introduce scheduled noise and iterative denoising, but lack explicit reconstruction mappings. We propose the Diffusion-Scheduled Denoising Autoencoder (DDAE), a framework that integrates diffusion-based noise scheduling and contrastive learning into the encoding process to improve anomaly detection. We evaluated DDAE on 57 datasets from ADBench. Our method outperforms in semi-supervised settings and achieves competitive results in unsupervised settings, improving PR-AUC by up to 65% (9%) and ROC-AUC by 16% (6%) over state-of-the-art autoencoder (diffusion) model baselines. We observed that higher noise levels benefit unsupervised training, while lower noise with linear scheduling is optimal in semi-supervised settings. These findings underscore the importance of principled noise strategies in tabular anomaly detection.
Authors: Antonio Tudisco, Andrea Marchesin, Maurizio Zamboni, Mariagrazia Graziano, Giovanna Turvani
Abstract: Recent advancements in Quantum Computing and Machine Learning have increased attention to Quantum Machine Learning (QML), which aims to develop machine learning models by exploiting the quantum computing paradigm. One of the widely used models in this area is the Variational Quantum Circuit (VQC), a hybrid model where the quantum circuit handles data inference while classical optimization adjusts the parameters of the circuit. The quantum circuit consists of an encoding layer, which loads data into the circuit, and a template circuit, known as the ansatz, responsible for processing the data. This work involves performing an analysis by considering both Amplitude- and Angle-encoding models, and examining how the type of rotational gate applied affects the classification performance of the model. This comparison is carried out by training the different models on two datasets, Wine and Diabetes, and evaluating their performance. The study demonstrates that, under identical model topologies, the difference in accuracy between the best and worst models ranges from 10% to 30%, with differences reaching up to 41%. Moreover, the results highlight how the choice of rotational gates used in encoding can significantly impact the model's classification performance. The findings confirm that the embedding represents a hyperparameter for VQC models.
Authors: Bushra Akter, Md Biplob Hosen, Sabbir Ahmed, Mehrin Anannya, Md. Farhad Hossain
Abstract: Academic performance depends on a multivariable nexus of socio-academic and financial factors. This study investigates these influences to develop effective strategies for optimizing students' CGPA. To achieve this, we reviewed various literature to identify key influencing factors and constructed an initial hypothetical causal graph based on the findings. Additionally, an online survey was conducted, where 1,050 students participated, providing comprehensive data for analysis. Rigorous data preprocessing techniques, including cleaning and visualization, ensured data quality before analysis. Causal analysis validated the relationships among variables, offering deeper insights into their direct and indirect effects on CGPA. Regression models were implemented for CGPA prediction, while classification models categorized students based on performance levels. Ridge Regression demonstrated strong predictive accuracy, achieving a Mean Absolute Error of 0.12 and a Mean Squared Error of 0.023. Random Forest outperformed in classification, attaining an F1-score near perfection and an accuracy of 98.68%. Explainable AI techniques such as SHAP, LIME, and Interpret enhanced model interpretability, highlighting critical factors such as study hours, scholarships, parental education, and prior academic performance. The study culminated in the development of a web-based application that provides students with personalized insights, allowing them to predict academic performance, identify areas for improvement, and make informed decisions to enhance their outcomes.
Authors: Ping Chen, Zhuohong Deng, Ping Li, Shuibing He, Hongzi Zhu, Yi Zheng, Zhefeng Wang, Baoxing Huai, Minyi Guo
Abstract: Training large language models often employs recomputation to alleviate memory pressure, which can introduce up to 30% overhead in real-world scenarios. In this paper, we propose Adacc, a novel memory management framework that combines adaptive compression and activation checkpointing to reduce the GPU memory footprint. It comprises three modules: (1) We design layer-specific compression algorithms that account for outliers in LLM tensors, instead of directly quantizing floats from FP16 to INT4, to ensure model accuracy. (2) We propose an optimal scheduling policy that employs MILP to determine the best memory optimization for each tensor. (3) To accommodate changes in training tensors, we introduce an adaptive policy evolution mechanism that adjusts the policy during training to enhance throughput. Experimental results show that Adacc can accelerate the LLM training by 1.01x to 1.37x compared to state-of-the-art frameworks, while maintaining comparable model accuracy to the Baseline.
Authors: Muhammad Farid Adilazuarda, Musa Izzanardi Wijanarko, Lucky Susanto, Khumaisa Nur'aini, Derry Wijaya, Alham Fikri Aji
Abstract: Indonesia is rich in languages and scripts. However, most NLP progress has been made using romanized text. In this paper, we present NusaAksara, a novel public benchmark for Indonesian languages that includes their original scripts. Our benchmark covers both text and image modalities and encompasses diverse tasks such as image segmentation, OCR, transliteration, translation, and language identification. Our data is constructed by human experts through rigorous steps. NusaAksara covers 8 scripts across 7 languages, including low-resource languages not commonly seen in NLP benchmarks. Although unsupported by Unicode, the Lampung script is included in this dataset. We benchmark our data across several models, from LLMs and VLMs such as GPT-4o, Llama 3.2, and Aya 23 to task-specific systems such as PP-OCR and LangID, and show that most NLP technologies cannot handle Indonesia's local scripts, with many achieving near-zero performance.
Authors: Ahmet Melih Ince, Ayse Elif Canbilen, Halim Yanikomeroglu
Abstract: Sixth-generation (6G) networks are designed to meet the hyper-reliable and low-latency communication (HRLLC) requirements of safety-critical applications such as autonomous driving. Integrating non-terrestrial networks (NTN) into the 6G infrastructure brings redundancy to the network, ensuring continuity of communications even under extreme conditions. In particular, high-altitude platform stations (HAPS) stand out for their wide coverage and low latency advantages, supporting communication reliability and enhancing information freshness, especially in rural areas and regions with infrastructure constraints. In this paper, we present reinforcement learning-based approaches using deep deterministic policy gradient (DDPG) to dynamically optimize the age-of-information (AoI) in HAPS-enabled vehicle-to-everything (V2X) networks. The proposed method improves information freshness and overall network reliability by enabling independent learning without centralized coordination. The findings reveal the potential of HAPS-supported solutions, combined with DDPG-based learning, for efficient AoI-aware resource allocation in platoon-based autonomous vehicle systems.
Authors: Sebasti\'an Andr\'es Cajas Ord\'o\~nez, Luis Fernando Torres Torres, Mario Bifulco, Carlos Andr\'es Dur\'an, Cristian Bosch, Ricardo Sim\'on Carbajo
Abstract: Quantum Support Vector Machines face scalability challenges due to high-dimensional quantum states and hardware limitations. We propose an embedding-aware quantum-classical pipeline combining class-balanced k-means distillation with pretrained Vision Transformer embeddings. Our key finding: ViT embeddings uniquely enable quantum advantage, achieving up to 8.02% accuracy improvements over classical SVMs on Fashion-MNIST and 4.42% on MNIST, while CNN features show performance degradation. Using 16-qubit tensor network simulation via cuTensorNet, we provide the first systematic evidence that quantum kernel advantage depends critically on embedding choice, revealing fundamental synergy between transformer attention and quantum feature spaces. This provides a practical pathway for scalable quantum machine learning that leverages modern neural architectures.
Authors: Azadeh Alavi, Fatemeh Kouchmeshki, Abdolrahman Alavi, Yongli Ren, Jiayang Niu
Abstract: Modern recommenders describe each item with hundreds of sparse semantic tags, yet most quantum pipelines still map one qubit per tag, demanding well beyond one hundred qubits, far out of reach for current noisy-intermediate-scale quantum (NISQ) devices and prone to deep, error-amplifying circuits. We close this gap with a three-stage hybrid machine learning algorithm that compresses tag profiles, optimizes feature selection under a fixed qubit budget via QAOA, and scores recommendations with a Quantum semi-Random Forest (QsRF) built on just five qubits, while performing similarly to the state-of-the-art methods. Leveraging SVD sketching and k-means, we learn a 1000-atom dictionary ($>$97 \% variance), then solve a 2020 QUBO via depth-3 QAOA to select 5 atoms. A 100-tree QsRF trained on these codes matches full-feature baselines on ICM-150/500.
Authors: Azadeh Alavi, Sanduni Jayasinghe, Mojtaba Mahmoodian, Sam Mazaheri, John Thangarajah, Sujeeva Setunge
Abstract: Large-scale civil structures, such as bridges, pipelines, and offshore platforms, are vital to modern infrastructure, where unexpected failures can cause significant economic and safety repercussions. Although finite element (FE) modeling is widely used for real-time structural health monitoring (SHM), its high computational cost and the complexity of inverse FE analysis, where low dimensional sensor data must map onto high-dimensional displacement or stress fields pose ongoing challenges. Here, we propose a hybrid quantum classical multilayer perceptron (QMLP) framework to tackle these issues and facilitate swift updates to digital twins across a range of structural applications. Our approach embeds sensor data using symmetric positive definite (SPD) matrices and polynomial features, yielding a representation well suited to quantum processing. A parameterized quantum circuit (PQC) transforms these features, and the resultant quantum outputs feed into a classical neural network for final inference. By fusing quantum capabilities with classical modeling, the QMLP handles large scale inverse FE mapping while preserving computational viability. Through extensive experiments on a bridge, we demonstrate that the QMLP achieves a mean squared error (MSE) of 0.0000000000316, outperforming purely classical baselines with a large margin. These findings confirm the potential of quantum-enhanced methods for real time SHM, establishing a pathway toward more efficient, scalable digital twins that can robustly monitor and diagnose structural integrity in near real time.
Authors: Athanasios Tziouvaras, Carolina Fortuna, George Floros, Kostas Kolomvatsos, Panagiotis Sarigiannidis, Marko Grobelnik, Bla\v{z} Bertalani\v{c}
Abstract: AI-native 6G networks promise unprecedented automation and performance by embedding machine-learning models throughout the radio access and core segments of the network. However, the non-stationary nature of wireless environments due to infrastructure changes, user mobility, and emerging traffic patterns, induces concept drifts that can quickly degrade these model accuracies. Existing methods in general are very domain specific, or struggle with certain type of concept drift. In this paper, we introduce two unsupervised, model-agnostic, batch concept drift detectors. Both methods compute an expected-utility score to decide when concept drift occurred and if model retraining is warranted, without requiring ground-truth labels after deployment. We validate our framework on two real-world wireless use cases in outdoor fingerprinting for localization and for link-anomaly detection, and demonstrate that both methods are outperforming classical detectors such as ADWIN, DDM, CUSUM by 20-40 percentage points. Additionally, they achieve an F1-score of 0.94 and 1.00 in correctly triggering retraining alarm, thus reducing the false alarm rate by up to 20 percentage points compared to the best classical detectors.
Authors: Ammar Daskin
Abstract: Schmidt decomposition of a vector can be understood as writing the singular value decomposition (SVD) in vector form. A vector can be written as a linear combination of tensor product of two dimensional vectors by recursively applying Schmidt decompositions via SVD to all subsystems. Given a vector expressed as a linear combination of tensor products, using only the $k$ principal terms yields a $k$-rank approximation of the vector. Therefore, writing a vector in this reduced form allows to retain most important parts of the vector while removing small noises from it, analogous to SVD-based denoising. In this paper, we show that quantum circuits designed based on a value $k$ (determined from the tensor network decomposition of the mean vector of the training sample) can approximate the reduced-form representations of entire datasets. We then employ this circuit ansatz with a classical neural network head to construct a hybrid machine learning model. Since the output of the quantum circuit for an $2^n$ dimensional vector is an $n$ dimensional probability vector, this provides an exponential compression of the input and potentially can reduce the number of learnable parameters for training large-scale models. We use datasets provided in the Python scikit-learn module for the experiments. The results confirm the quantum circuit is able to compress data successfully to provide effective $k$-rank approximations to the classical processing component.
Authors: Yihong Dong, Xue Jiang, Jiaru Qian, Tian Wang, Kechi Zhang, Zhi Jin, Ge Li
Abstract: Code generation agents powered by large language models (LLMs) are revolutionizing the software development paradigm. Distinct from previous code generation techniques, code generation agents are characterized by three core features. 1) Autonomy: the ability to independently manage the entire workflow, from task decomposition to coding and debugging. 2) Expanded task scope: capabilities that extend beyond generating code snippets to encompass the full software development lifecycle (SDLC). 3) Enhancement of engineering practicality: a shift in research emphasis from algorithmic innovation toward practical engineering challenges, such as system reliability, process management, and tool integration. This domain has recently witnessed rapid development and an explosion in research, demonstrating significant application potential. This paper presents a systematic survey of the field of LLM-based code generation agents. We trace the technology's developmental trajectory from its inception and systematically categorize its core techniques, including both single-agent and multi-agent architectures. Furthermore, this survey details the applications of LLM-based agents across the full SDLC, summarizes mainstream evaluation benchmarks and metrics, and catalogs representative tools. Finally, by analyzing the primary challenges, we identify and propose several foundational, long-term research directions for the future work of the field.
Authors: Chandler Smith, HanQin Cai, Abiy Tasissa
Abstract: The problem of recovering a configuration of points from partial pairwise distances, referred to as the Euclidean Distance Geometry (EDG) problem, arises in a broad range of applications, including sensor network localization, molecular conformation, and manifold learning. In this paper, we propose a Riemannian optimization framework for solving the EDG problem by formulating it as a low-rank matrix completion task over the space of positive semi-definite Gram matrices. The available distance measurements are encoded as expansion coefficients in a non-orthogonal basis, and optimization over the Gram matrix implicitly enforces geometric consistency through the triangle inequality, a structure inherited from classical multidimensional scaling. Under a Bernoulli sampling model for observed distances, we prove that Riemannian gradient descent on the manifold of rank-$r$ matrices locally converges linearly with high probability when the sampling probability satisfies $p \geq \mathcal{O}(\nu^2 r^2 \log(n)/n)$, where $\nu$ is an EDG-specific incoherence parameter. Furthermore, we provide an initialization candidate using a one-step hard thresholding procedure that yields convergence, provided the sampling probability satisfies $p \geq \mathcal{O}(\nu r^{3/2} \log^{3/4}(n)/n^{1/4})$. A key technical contribution of this work is the analysis of a symmetric linear operator arising from a dual basis expansion in the non-orthogonal basis, which requires a novel application of the Hanson--Wright inequality to establish an optimal restricted isometry property in the presence of coupled terms. Empirical evaluations on synthetic data demonstrate that our algorithm achieves competitive performance relative to state-of-the-art methods. Moreover, we propose a novel notion of matrix incoherence tailored to the EDG setting and provide robustness guarantees for our method.
Authors: Alena Kopani\v{c}\'akov\'a, Youngkyu Lee, George Em Karniadakis
Abstract: We propose a new deflation strategy to accelerate the convergence of the preconditioned conjugate gradient(PCG) method for solving parametric large-scale linear systems of equations. Unlike traditional deflation techniques that rely on eigenvector approximations or recycled Krylov subspaces, we generate the deflation subspaces using operator learning, specifically the Deep Operator Network~(DeepONet). To this aim, we introduce two complementary approaches for assembling the deflation operators. The first approach approximates near-null space vectors of the discrete PDE operator using the basis functions learned by the DeepONet. The second approach directly leverages solutions predicted by the DeepONet. To further enhance convergence, we also propose several strategies for prescribing the sparsity pattern of the deflation operator. A comprehensive set of numerical experiments encompassing steady-state, time-dependent, scalar, and vector-valued problems posed on both structured and unstructured geometries is presented and demonstrates the effectiveness of the proposed DeepONet-based deflated PCG method, as well as its generalization across a wide range of model parameters and problem resolutions.
Authors: Ernest Bonnah, Luan Viet Nguyen, Khaza Anuarul Hoque
Abstract: Hyperproperties for Time Window Temporal Logic (HyperTWTL) is a domain-specific formal specification language known for its effectiveness in compactly representing security, opacity, and concurrency properties for robotics applications. This paper focuses on HyperTWTL-constrained secure reinforcement learning (SecRL). Although temporal logic-constrained safe reinforcement learning (SRL) is an evolving research problem with several existing literature, there is a significant research gap in exploring security-aware reinforcement learning (RL) using hyperproperties. Given the dynamics of an agent as a Markov Decision Process (MDP) and opacity/security constraints formalized as HyperTWTL, we propose an approach for learning security-aware optimal policies using dynamic Boltzmann softmax RL while satisfying the HyperTWTL constraints. The effectiveness and scalability of our proposed approach are demonstrated using a pick-up and delivery robotic mission case study. We also compare our results with two other baseline RL algorithms, showing that our proposed method outperforms them.
Authors: Katharine M. Clark, Paul D. McNicholas
Abstract: Functional data present unique challenges for clustering due to their infinite-dimensional nature and potential sensitivity to outliers. An extension of the OCLUST algorithm to the functional setting is proposed to address these issues. The approach leverages the OCLUST framework, creating a robust method to cluster curves and trim outliers. The methodology is evaluated on both simulated and real-world functional datasets, demonstrating strong performance in clustering and outlier identification.
Authors: Basna Mohammed Salih Hasan, Ramadhan J. Mstafa
Abstract: Gender classification has emerged as a crucial aspect in various fields, including security, human-machine interaction, surveillance, and advertising. Nonetheless, the accuracy of this classification can be influenced by factors such as cosmetics and disguise. Consequently, our study is dedicated to addressing this concern by concentrating on gender classification using color images of the periocular region. The periocular region refers to the area surrounding the eye, including the eyelids, eyebrows, and the region between them. It contains valuable visual cues that can be used to extract key features for gender classification. This paper introduces a sophisticated Convolutional Neural Network (CNN) model that utilizes color image databases to evaluate the effectiveness of the periocular region for gender classification. To validate the model's performance, we conducted tests on two eye datasets, namely CVBL and (Female and Male). The recommended architecture achieved an outstanding accuracy of 99% on the previously unused CVBL dataset while attaining a commendable accuracy of 96% with a small number of learnable parameters (7,235,089) on the (Female and Male) dataset. To ascertain the effectiveness of our proposed model for gender classification using the periocular region, we evaluated its performance through an extensive range of metrics and compared it with other state-of-the-art approaches. The results unequivocally demonstrate the efficacy of our model, thereby suggesting its potential for practical application in domains such as security and surveillance.
Authors: Babak Esmaeili, Hamidreza Modares, Stefano Di Cairano
Abstract: This paper proposes a data-driven motion-planning framework for nonlinear systems that constructs a sequence of overlapping invariant polytopes. Around each randomly sampled waypoint, the algorithm identifies a convex admissible region and solves data-driven linear-matrix-inequality problems to learn several ellipsoidal invariant sets together with their local state-feedback gains. The convex hull of these ellipsoids, still invariant under a piece-wise-affine controller obtained by interpolating the gains, is then approximated by a polytope. Safe transitions between nodes are ensured by verifying the intersection of consecutive convex-hull polytopes and introducing an intermediate node for a smooth transition. Control gains are interpolated in real time via simplex-based interpolation, keeping the state inside the invariant polytopes throughout the motion. Unlike traditional approaches that rely on system dynamics models, our method requires only data to compute safe regions and design state-feedback controllers. The approach is validated through simulations, demonstrating the effectiveness of the proposed method in achieving safe, dynamically feasible paths for complex nonlinear systems.
Authors: Tomasz Szczepa\'nski, Szymon P{\l}otka, Michal K. Grzeszczyk, Arleta Adamowicz, Piotr Fudalej, Przemys{\l}aw Korzeniowski, Tomasz Trzci\'nski, Arkadiusz Sitek
Abstract: Tooth segmentation in Cone-Beam Computed Tomography (CBCT) remains challenging, especially for fine structures like root apices, which is critical for assessing root resorption in orthodontics. We introduce GEPAR3D, a novel approach that unifies instance detection and multi-class segmentation into a single step tailored to improve root segmentation. Our method integrates a Statistical Shape Model of dentition as a geometric prior, capturing anatomical context and morphological consistency without enforcing restrictive adjacency constraints. We leverage a deep watershed method, modeling each tooth as a continuous 3D energy basin encoding voxel distances to boundaries. This instance-aware representation ensures accurate segmentation of narrow, complex root apices. Trained on publicly available CBCT scans from a single center, our method is evaluated on external test sets from two in-house and two public medical centers. GEPAR3D achieves the highest overall segmentation performance, averaging a Dice Similarity Coefficient (DSC) of 95.0% (+2.8% over the second-best method) and increasing recall to 95.2% (+9.5%) across all test sets. Qualitative analyses demonstrated substantial improvements in root segmentation quality, indicating significant potential for more accurate root resorption assessment and enhanced clinical decision-making in orthodontics. We provide the implementation and dataset at https://github.com/tomek1911/GEPAR3D.
Authors: Jobst Heitzig, Ram Potham
Abstract: Power is a key concept in AI safety: power-seeking as an instrumental goal, sudden or gradual disempowerment of humans, power balance in human-AI interaction and international AI governance. At the same time, power as the ability to pursue diverse goals is essential for wellbeing. This paper explores the idea of promoting both safety and wellbeing by forcing AI agents explicitly to empower humans and to manage the power balance between humans and AI agents in a desirable way. Using a principled, partially axiomatic approach, we design a parametrizable and decomposable objective function that represents an inequality- and risk-averse long-term aggregate of human power. It takes into account humans' bounded rationality and social norms, and, crucially, considers a wide variety of possible human goals. We derive algorithms for computing that metric by backward induction or approximating it via a form of multi-agent reinforcement learning from a given world model. We exemplify the consequences of (softly) maximizing this metric in a variety of paradigmatic situations and describe what instrumental sub-goals it will likely imply. Our cautious assessment is that softly maximizing suitable aggregate metrics of human power might constitute a beneficial objective for agentic AI systems that is safer than direct utility-based objectives.
Authors: Eric Mjolsness, Cory B. Scott
Abstract: Graphs, and sequences of growing graphs, can be used to specify the architecture of mathematical models in many fields including machine learning and computational science. Here we define structured graph "lineages" (ordered by level number) that grow in a hierarchical fashion, so that: (1) the number of graph vertices and edges increases exponentially in level number; (2) bipartite graphs connect successive levels within a graph lineage and, as in multigrid methods, can constrain matrices relating successive levels; (3) using prolongation maps within a graph lineage, process-derived distance measures between graphs at successive levels can be defined; (4) a category of "graded graphs" can be defined, and using it low-cost "skeletal" variants of standard algebraic graph operations and type constructors (cross product, box product, disjoint sum, and function types) can be derived for graded graphs and hence hierarchical graph lineages; (5) these skeletal binary operators have similar but not identical algebraic and category-theoretic properties to their standard counterparts; (6) graph lineages and their skeletal product constructors can approach continuum limit objects. Additional space-efficient unary operators on graded graphs are also derived: thickening, which creates a graph lineage of multiscale graphs, and escalation to a graph lineage of search frontiers (useful as a generalization of adaptive grids and in defining "skeletal" functions). The result is an algebraic type theory for graded graphs and (hierarchical) graph lineages. The approach is expected to be well suited to defining hierarchical model architectures - "hierarchitectures" - and local sampling, search, or optimization algorithms on them. We demonstrate such application to deep neural networks (including visual and feature scale spaces) and to multigrid numerical methods.
Authors: Shayan Jalilian, Abdul Bais
Abstract: The Segment Anything Model (SAM) has demonstrated impressive generalization in prompt-based segmentation. Yet, the potential of semantic text prompts remains underexplored compared to traditional spatial prompts like points and boxes. This paper introduces SAM-PTx, a parameter-efficient approach for adapting SAM using frozen CLIP-derived text embeddings as class-level semantic guidance. Specifically, we propose a lightweight adapter design called Parallel-Text that injects text embeddings into SAM's image encoder, enabling semantics-guided segmentation while keeping most of the original architecture frozen. Our adapter modifies only the MLP-parallel branch of each transformer block, preserving the attention pathway for spatial reasoning. Through supervised experiments and ablations on the COD10K dataset as well as low-data subsets of COCO and ADE20K, we show that incorporating fixed text embeddings as input improves segmentation performance over purely spatial prompt baselines. To our knowledge, this is the first work to use text prompts for segmentation on the COD10K dataset. These results suggest that integrating semantic conditioning into SAM's architecture offers a practical and scalable path for efficient adaptation with minimal computational complexity.
Authors: Xiaofeng Wu, Alan Ritter, Wei Xu
Abstract: Tables have gained significant attention in large language models (LLMs) and multimodal large language models (MLLMs) due to their complex and flexible structure. Unlike linear text inputs, tables are two-dimensional, encompassing formats that range from well-structured database tables to complex, multi-layered spreadsheets, each with different purposes. This diversity in format and purpose has led to the development of specialized methods and tasks, instead of universal approaches, making navigation of table understanding tasks challenging. To address these challenges, this paper introduces key concepts through a taxonomy of tabular input representations and an introduction of table understanding tasks. We highlight several critical gaps in the field that indicate the need for further research: (1) the predominance of retrieval-focused tasks that require minimal reasoning beyond mathematical and logical operations; (2) significant challenges faced by models when processing complex table structures, large-scale tables, length context, or multi-table scenarios; and (3) the limited generalization of models across different tabular representations and formats.
Authors: Aymane Abdali, Bartosz Boguslawski, Lucas Drumetz, Vincent Gripon
Abstract: In the domain of Few-Shot Image Classification, operating with as little as one example per class, the presence of image ambiguities stemming from multiple objects or complex backgrounds can significantly deteriorate performance. Our research demonstrates that incorporating additional information about the local positioning of an object within its image markedly enhances classification across established benchmarks. More importantly, we show that a significant fraction of the improvement can be achieved through the use of the Segment Anything Model, requiring only a pixel of the object of interest to be pointed out, or by employing fully unsupervised foreground object extraction methods.
Authors: Yihong Dong, Xue Jiang, Yongding Tao, Huanyu Liu, Kechi Zhang, Lili Mou, Rongyu Cao, Yingwei Ma, Jue Chen, Binhua Li, Zhi Jin, Fei Huang, Yongbin Li, Ge Li
Abstract: Reinforcement Learning with Verifiable Reward (RLVR) has significantly advanced the complex reasoning abilities of Large Language Models (LLMs). However, it struggles to break through the inherent capability boundaries of the base LLM, due to its inherently on-policy strategy with LLM's immense action space and sparse reward. Further, RLVR can lead to the capability boundary collapse, narrowing the LLM's problem-solving scope. To address this problem, we propose RL-PLUS, a novel approach that synergizes internal exploitation (i.e., Thinking) with external data (i.e., Learning) to achieve stronger reasoning capabilities and surpass the boundaries of base models. RL-PLUS integrates two core components: Multiple Importance Sampling to address for distributional mismatch from external data, and an Exploration-Based Advantage Function to guide the model towards high-value, unexplored reasoning paths. We provide both theoretical analysis and extensive experiments to demonstrate the superiority and generalizability of our approach. The results show that RL-PLUS achieves state-of-the-art performance compared with existing RLVR methods on six math reasoning benchmarks and exhibits superior performance on six out-of-distribution reasoning tasks. It also achieves consistent and significant gains across diverse model families, with average relative improvements ranging from 21.1\% to 69.2\%. Moreover, Pass@k curves across multiple benchmarks indicate that RL-PLUS effectively resolves the capability boundary collapse problem.
Authors: Sergei Gleyzer, Hanh Nguyen, Dinesh P. Ramakrishnan, Eric A. F. Reinhardt
Abstract: The Kolmogorov-Arnold representation theorem states that any continuous multivariable function can be exactly represented as a finite superposition of continuous single variable functions. Subsequent simplifications of this representation involve expressing these functions as parameterized sums of a smaller number of unique monotonic functions. These developments led to the proof of the universal approximation capabilities of multilayer perceptron networks with sigmoidal activations, forming the alternative theoretical direction of most modern neural networks. Kolmogorov-Arnold Networks (KANs) have been recently proposed as an alternative to multilayer perceptrons. KANs feature learnable nonlinear activations applied directly to input values, modeled as weighted sums of basis spline functions. This approach replaces the linear transformations and sigmoidal post-activations used in traditional perceptrons. Subsequent works have explored alternatives to spline-based activations. In this work, we propose a novel KAN variant by replacing both the inner and outer functions in the Kolmogorov-Arnold representation with weighted sinusoidal functions of learnable frequencies. Inspired by simplifications introduced by Lorentz and Sprecher, we fix the phases of the sinusoidal activations to linearly spaced constant values and provide a proof of its theoretical validity. We also conduct numerical experiments to evaluate its performance on a range of multivariable functions, comparing it with fixed-frequency Fourier transform methods and multilayer perceptrons (MLPs). We show that it outperforms the fixed-frequency Fourier transform and achieves comparable performance to MLPs.
Authors: Victor D. Martinez, Vidya Manian, Sudhir Malik
Abstract: This article presents, for the first time, the application of diffusion models for generating jet images corresponding to proton-proton collision events at the Large Hadron Collider (LHC). The kinematic variables of quark, gluon, W-boson, Z-boson, and top quark jets from the JetNet simulation dataset are mapped to two-dimensional image representations. Diffusion models are trained on these images to learn the spatial distribution of jet constituents. We compare the performance of score-based diffusion models and consistency models in accurately generating class-conditional jet images. Unlike approaches based on latent distributions, our method operates directly in image space. The fidelity of the generated images is evaluated using several metrics, including the Fr\'echet Inception Distance (FID), which demonstrates that consistency models achieve higher fidelity and generation stability compared to score-based diffusion models. These advancements offer significant improvements in computational efficiency and generation accuracy, providing valuable tools for High Energy Physics (HEP) research.
Authors: Molly Noel, Gabriel Mancino-Ball, Yangyang Xu
Abstract: Graph convolutional networks (GCNs) are a powerful tool for graph representation learning. Due to the recursive neighborhood aggregations employed by GCNs, efficient training methods suffer from a lack of theoretical guarantees or are missing important practical elements from modern deep learning algorithms, such as adaptivity and momentum. In this paper, we present several neighbor-sampling (NS) based Adam-type stochastic methods for solving a nonconvex GCN training problem. We utilize the control variate technique proposed by [1] to reduce the stochastic error caused by neighbor sampling. Under standard assumptions for Adam-type methods, we show that our methods enjoy the optimal convergence rate. In addition, we conduct extensive numerical experiments on node classification tasks with several benchmark datasets. The results demonstrate superior performance of our methods over classic NS-based SGD that also uses the control-variate technique, especially for large-scale graph datasets. Our code is available at https://github.com/RPI-OPT/CV-ADAM-GNN .
Authors: Shruthi Chari, Oshani Seneviratne, Prithwish Chakraborty, Pablo Meyer, Deborah L. McGuinness
Abstract: Explanations are crucial for building trustworthy AI systems, but a gap often exists between the explanations provided by models and those needed by users. To address this gap, we introduce MetaExplainer, a neuro-symbolic framework designed to generate user-centered explanations. Our approach employs a three-stage process: first, we decompose user questions into machine-readable formats using state-of-the-art large language models (LLM); second, we delegate the task of generating system recommendations to model explainer methods; and finally, we synthesize natural language explanations that summarize the explainer outputs. Throughout this process, we utilize an Explanation Ontology to guide the language models and explainer methods. By leveraging LLMs and a structured approach to explanation generation, MetaExplainer aims to enhance the interpretability and trustworthiness of AI systems across various applications, providing users with tailored, question-driven explanations that better meet their needs. Comprehensive evaluations of MetaExplainer demonstrate a step towards evaluating and utilizing current state-of-the-art explanation frameworks. Our results show high performance across all stages, with a 59.06% F1-score in question reframing, 70% faithfulness in model explanations, and 67% context-utilization in natural language synthesis. User studies corroborate these findings, highlighting the creativity and comprehensiveness of generated explanations. Tested on the Diabetes (PIMA Indian) tabular dataset, MetaExplainer supports diverse explanation types, including Contrastive, Counterfactual, Rationale, Case-Based, and Data explanations. The framework's versatility and traceability from using ontology to guide LLMs suggest broad applicability beyond the tested scenarios, positioning MetaExplainer as a promising tool for enhancing AI explainability across various domains.
Authors: Ammar Ahmed, Sheng Di, Franck Cappello, Zirui Liu, Jingoo Han, Ali Anwar
Abstract: Large language models (LLMs) excel across diverse natural language processing tasks but face resource demands and limited context windows. Although techniques like pruning, quantization, and token dropping can mitigate these issues, their efficacy in long-context scenarios and system evaluation remains underexplored. This paper systematically benchmarks these optimizations, characterizing memory usage, latency, and throughput, and studies how these methods impact the quality of text generation. We first analyze individual optimization methods for two LLM architectures supporting long context and then systematically evaluate combinations of these techniques to assess how this deeper analysis impacts performance metrics. We subsequently study the scalability of individual optimization methods on a larger variant with 70 billion-parameter model. Our novel insights reveal that naive combination inference optimization algorithms can adversely affect larger models due to compounded approximation errors, as compared to their smaller counterparts. Experiments show that relying solely on F1 obscures these effects by hiding precision-recall trade-offs in question answering tasks. By integrating system-level profiling with task-specific insights, this study helps LLM practitioners and researchers explore and balance efficiency, accuracy, and scalability across tasks and hardware configurations.
Authors: Sunghyun Park, Seokeon Choi, Hyoungwoo Park, Sungrack Yun
Abstract: Personalizing text-to-image diffusion models is crucial for adapting the pre-trained models to specific target concepts, enabling diverse image generation. However, fine-tuning with few images introduces an inherent trade-off between aligning with the target distribution (e.g., subject fidelity) and preserving the broad knowledge of the original model (e.g., text editability). Existing sampling guidance methods, such as classifier-free guidance (CFG) and autoguidance (AG), fail to effectively guide the output toward well-balanced space: CFG restricts the adaptation to the target distribution, while AG compromises text alignment. To address these limitations, we propose personalization guidance, a simple yet effective method leveraging an unlearned weak model conditioned on a null text prompt. Moreover, our method dynamically controls the extent of unlearning in a weak model through weight interpolation between pre-trained and fine-tuned models during inference. Unlike existing guidance methods, which depend solely on guidance scales, our method explicitly steers the outputs toward a balanced latent space without additional computational overhead. Experimental results demonstrate that our proposed guidance can improve text alignment and target distribution fidelity, integrating seamlessly with various fine-tuning strategies.
Authors: Alessandro Gaudenzi, Lorenzo Nodari, Lance Kaplan, Alessandra Russo, Murat Sensoy, Federico Cerutti
Abstract: Advanced Persistent Threats (APTs) represent a significant challenge in cybersecurity due to their prolonged, multi-stage nature and the sophistication of their operators. Traditional detection systems typically focus on identifying malicious activity in binary terms (benign or malicious) without accounting for the progression of an attack. However, effective response strategies depend on accurate inference of the attack's current stage, as countermeasures must be tailored to whether an adversary is in the early reconnaissance phase or actively conducting exploitation or exfiltration. This work addresses the problem of attack stage inference under uncertainty, with a focus on robustness to out-of-distribution (OOD) inputs. We propose a classification approach based on Evidential Deep Learning (EDL), which models predictive uncertainty by outputting parameters of a Dirichlet distribution over possible stages. This allows the system not only to predict the most likely stage of an attack but also to indicate when it is uncertain or the input lies outside the training distribution. Preliminary experiments in a simulated environment demonstrate that the proposed model can accurately infer the stage of an attack with calibrated confidence while effectively detecting OOD inputs, which may indicate changes in the attackers' tactics. These results support the feasibility of deploying uncertainty-aware models for staged threat detection in dynamic and adversarial environments.
Authors: Jiyu Chen, Poh Seng Lim, Shuang Peng, Daxiong Luo, JungHau Foo, Yap Deep, Timothy Lee Jun Jie, Kelvin Teh Kae Wen, Fan Yang, Danyu Feng, Hao-Yun Chen, Peng-Wen Chen, Fangyuan Li, Xiaoxin Chen, Wong Wai Mun
Abstract: Deploying Transformer-based large language models (LLMs) on resource-constrained edge devices for long-sequence tasks remains challenging due to the quadratic time complexity of self-attention and growing Key-Value (KV) cache demands. While existing KV cache optimizations improve memory efficiency, they often fail to reduce time to first token (TTFT) and may degrade performance through token pruning. Alternative sequence modeling architectures address some of these limitations, but typically require full retraining and lack infrastructure support. EdgeInfinite offers an efficient solution by fine-tuning only a small subset of parameters, maintaining quality while reducing both computational and memory costs, including improved TTFT. However, its instruction-following ability is limited, and it lacks mobile-specific optimizations. To address these issues, we propose EdgeInfinite-Instruct, which introduces a Segmented Supervised Fine-Tuning (S-SFT) strategy tailored to long-sequence tasks such as summarization and question answering. We further optimized EdgeInfinite-Instruct for efficient deployment on edge NPUs by employing fine-grained post-training quantization (PTQ) to reduce computational demands while maintaining accuracy, and by implementing a fixed-shape computation graph that balances memory usage and on-device efficiency through scenario-specific customization of input token and cache sizes. Experiments on long-context benchmarks and real-world mobile tasks show that our approach improves domain-specific performance while maintaining efficiency on NPU-accelerated edge devices.
Authors: Kamal Basha S, Athira Nambiar
Abstract: Weld defect detection is crucial for ensuring the safety and reliability of piping systems in the oil and gas industry, especially in challenging marine and offshore environments. Traditional non-destructive testing (NDT) methods often fail to detect subtle or internal defects, leading to potential failures and costly downtime. Furthermore, existing neural network-based approaches for defect classification frequently rely on arbitrarily selected pretrained architectures and lack interpretability, raising safety concerns for deployment. To address these challenges, this paper introduces ``Adapt-WeldNet", an adaptive framework for welding defect detection that systematically evaluates various pre-trained architectures, transfer learning strategies, and adaptive optimizers to identify the best-performing model and hyperparameters, optimizing defect detection and providing actionable insights. Additionally, a novel Defect Detection Interpretability Analysis (DDIA) framework is proposed to enhance system transparency. DDIA employs Explainable AI (XAI) techniques, such as Grad-CAM and LIME, alongside domain-specific evaluations validated by certified ASNT NDE Level II professionals. Incorporating a Human-in-the-Loop (HITL) approach and aligning with the principles of Trustworthy AI, DDIA ensures the reliability, fairness, and accountability of the defect detection system, fostering confidence in automated decisions through expert validation. By improving both performance and interpretability, this work enhances trust, safety, and reliability in welding defect detection systems, supporting critical operations in offshore and marine environments.
Authors: Won June Cho, Hongjun Yoon, Daeky Jeong, Hyeongyeol Lim, Yosep Chong
Abstract: Spatial transcriptomics reveals gene expression patterns within tissue context, enabling precision oncology applications such as treatment response prediction, but its high cost and technical complexity limit clinical adoption. Predicting spatial gene expression (biomarkers) from routine histopathology images offers a practical alternative, yet current vision foundation models (VFMs) in pathology based on Vision Transformer (ViT) backbones perform below clinical standards. Given that VFMs are already trained on millions of diverse whole slide images, we hypothesize that architectural innovations beyond ViTs may better capture the low-frequency, subtle morphological patterns correlating with molecular phenotypes. By demonstrating that state space models initialized with negative real eigenvalues exhibit strong low-frequency bias, we introduce $MV_{Hybrid}$, a hybrid backbone architecture combining state space models (SSMs) with ViT. We compare five other different backbone architectures for pathology VFMs, all pretrained on identical colorectal cancer datasets using the DINOv2 self-supervised learning method. We evaluate all pretrained models using both random split and leave-one-study-out (LOSO) settings of the same biomarker dataset. In LOSO evaluation, $MV_{Hybrid}$ achieves 57% higher correlation than the best-performing ViT and shows 43% smaller performance degradation compared to random split in gene expression prediction, demonstrating superior performance and robustness, respectively. Furthermore, $MV_{Hybrid}$ shows equal or better downstream performance in classification, patch retrieval, and survival prediction tasks compared to that of ViT, showing its promise as a next-generation pathology VFM backbone. Our code is publicly available at: https://github.com/deepnoid-ai/MVHybrid.
Authors: Dingzirui Wang, Xuangliang Zhang, Keyan Xu, Qingfu Zhu, Wanxiang Che, Yang Deng
Abstract: Numerous studies have investigated the underlying mechanisms of in-context learning (ICL) effectiveness to inspire the design of related methods. However, existing work predominantly assumes the effectiveness of the demonstrations provided within ICL, while many research indicates that not all demonstrations are effective, failing to yielding any performance improvement during ICL. Therefore, in this paper, we investigate the reasons behind demonstration ineffectiveness. Our analysis is based on gradient flow and linear self-attention models. By setting the gradient flow to zero, we deduce that a demonstration becomes ineffective if its information has either been learned by the model or is irrelevant to the user query. Furthermore, we demonstrate that in multi-layer models, the disparity in effectiveness among demonstrations is amplified with layer increasing, causing the model to focus more on effective ones. Considering that current demonstration selection methods primarily focus on the relevance to the user query while overlooking the information that the model has already assimilated, we propose a novel method called GradS, which leverages gradient flow for demonstration selection. We use the magnitude of the gradient flow of the demonstration with respect to a given user query as the criterion, thereby ensuring the effectiveness of the chosen ones. We validate our derivation and GradS on four prominent LLMs across five mainstream datasets. The experimental results confirm that the disparity in effectiveness among demonstrations is magnified as the model layer increases, substantiating our derivations. Moreover, GradS achieves a relative improvement of $6.8\%$ on average over the strongest baselines, demonstrating its effectiveness.
Authors: Varun Bharti, Shashwat Jha, Dhruv Kumar, Pankaj Jalote
Abstract: Loop invariants are essential for proving the correctness of programs with loops. Developing loop invariants is challenging, and fully automatic synthesis cannot be guaranteed for arbitrary programs. Some approaches have been proposed to synthesize loop invariants using symbolic techniques and more recently using neural approaches. These approaches are able to correctly synthesize loop invariants only for subsets of standard benchmarks. In this work, we investigate whether modern, reasoning-optimized large language models can do better. We integrate OpenAI's O1, O1-mini, and O3-mini into a tightly coupled generate-and-check pipeline with the Z3 SMT solver, using solver counterexamples to iteratively guide invariant refinement. We use Code2Inv benchmark, which provides C programs along with their formal preconditions and postconditions. On this benchmark of 133 tasks, our framework achieves 100% coverage (133 out of 133), outperforming the previous best of 107 out of 133, while requiring only 1-2 model proposals per instance and 14-55 seconds of wall-clock time. These results demonstrate that LLMs possess latent logical reasoning capabilities which can help automate loop invariant synthesis. While our experiments target C-specific programs, this approach should be generalizable to other imperative languages.
Authors: Varun Bharti, Shashwat Jha, Dhruv Kumar, Pankaj Jalote
Abstract: Type annotations in Python enhance maintainability and error detection. However, generating these annotations manually is error prone and requires extra effort. Traditional automation approaches like static analysis, machine learning, and deep learning struggle with limited type vocabularies, behavioral over approximation, and reliance on large labeled datasets. In this work, we explore the use of LLMs for generating type annotations in Python. We develop a generate check repair pipeline: the LLM proposes annotations guided by a Concrete Syntax Tree representation, a static type checker (Mypy) verifies them, and any errors are fed back for iterative refinement. We evaluate four LLM variants: GPT 4oMini, GPT 4.1mini (general-purpose), and O3Mini, O4Mini (reasoning optimized), on 6000 code snippets from the ManyTypes4Py benchmark. We first measure the proportion of code snippets annotated by LLMs for which MyPy reported no errors (i.e., consistent results): GPT 4oMini achieved consistency on 65.9% of cases (34.1% inconsistent), while GPT 4.1mini, O3Mini, and O4Mini each reached approximately 88.6% consistency (around 11.4% failures). To measure annotation quality, we then compute exact-match and base-type match accuracies over all 6000 snippets: GPT 4.1mini and O3Mini perform the best, achieving up to 70.5% exact match and 79.1% base type accuracy, requiring under one repair iteration on average. Our results demonstrate that general-purpose and reasoning optimized LLMs, without any task specific fine tuning or additional training can be effective in generating consistent type annotations.They perform competitively with traditional deep learning techniques which require large labeled dataset for training. While our work focuses on Python, the pipeline can be extended to other optionally typed imperative languages like Ruby
Authors: Minghao Guo, Xi Zhu, Jingyuan Huang, Kai Mei, Yongfeng Zhang
Abstract: Graph Neural Networks (GNNs) have achieved remarkable success in graph-based learning by propagating information among neighbor nodes via predefined aggregation mechanisms. However, such fixed schemes often suffer from two key limitations. First, they cannot handle the imbalance in node informativeness -- some nodes are rich in information, while others remain sparse. Second, predefined message passing primarily leverages local structural similarity while ignoring global semantic relationships across the graph, limiting the model's ability to capture distant but relevant information. We propose Retrieval-augmented Graph Agentic Network (ReaGAN), an agent-based framework that empowers each node with autonomous, node-level decision-making. Each node acts as an agent that independently plans its next action based on its internal memory, enabling node-level planning and adaptive message propagation. Additionally, retrieval-augmented generation (RAG) allows nodes to access semantically relevant content and build global relationships in the graph. ReaGAN achieves competitive performance under few-shot in-context settings using a frozen LLM backbone without fine-tuning, showcasing the potential of agentic planning and local-global retrieval in graph learning.
Authors: Anju Rani, Daniel Ortiz-Arroyo, Petar Durdevic
Abstract: Understanding the temporal dynamics of biological growth is critical across diverse fields such as microbiology, agriculture, and biodegradation research. Although vision-language models like Contrastive Language Image Pretraining (CLIP) have shown strong capabilities in joint visual-textual reasoning, their effectiveness in capturing temporal progression remains limited. To address this, we propose CLIPTime, a multimodal, multitask framework designed to predict both the developmental stage and the corresponding timestamp of fungal growth from image and text inputs. Built upon the CLIP architecture, our model learns joint visual-textual embeddings and enables time-aware inference without requiring explicit temporal input during testing. To facilitate training and evaluation, we introduce a synthetic fungal growth dataset annotated with aligned timestamps and categorical stage labels. CLIPTime jointly performs classification and regression, predicting discrete growth stages alongside continuous timestamps. We also propose custom evaluation metrics, including temporal accuracy and regression error, to assess the precision of time-aware predictions. Experimental results demonstrate that CLIPTime effectively models biological progression and produces interpretable, temporally grounded outputs, highlighting the potential of vision-language models in real-world biological monitoring applications.
Authors: Joonas Tapaninaho, Mourad Oussala
Abstract: The training of modern large-language models requires an increasingly amount of computation power and time. Even smaller variants, such as small-language models (SLMs), take several days to train in the best-case scenarios, often requiring multiple GPUs. This paper explores methods to train and evaluate decoder-only transformer-based language models in hours instead of days/weeks. We introduces \textit{PaPaformer}, a decoder-only transformer architecture variant, whose lower-dimensional parallel paths are combined into larger model. The paper shows that these lower-dimensional paths can be trained individually with different types of training data and then combined into one larger model. This method gives the option to reduce the total number of model parameters and the training time with increasing performance. Moreover, the use of parallel path structure opens interesting possibilities to customize paths to accommodate specific task requirements.
Authors: Jens U. Kreber, Joerg Stueckler
Abstract: Articulated objects are an important type of interactable objects in everyday environments. In this paper, we propose PhysNAP, a novel diffusion model-based approach for generating articulated objects that aligns them with partial point clouds and improves their physical plausibility. The model represents part shapes by signed distance functions (SDFs). We guide the reverse diffusion process using a point cloud alignment loss computed using the predicted SDFs. Additionally, we impose non-penetration and mobility constraints based on the part SDFs for guiding the model to generate more physically plausible objects. We also make our diffusion approach category-aware to further improve point cloud alignment if category information is available. We evaluate the generative ability and constraint consistency of samples generated with PhysNAP using the PartNet-Mobility dataset. We also compare it with an unguided baseline diffusion model and demonstrate that PhysNAP can improve constraint consistency and provides a tradeoff with generative ability.
Authors: Xiang Zhang, Zhou Li, Shuangyang Li, Kai Wan, Derrick Wing Kwan Ng, Giuseppe Caire
Abstract: In decentralized federated learning (FL), multiple clients collaboratively learn a shared machine learning (ML) model by leveraging their privately held datasets distributed across the network, through interactive exchange of the intermediate model updates. To ensure data security, cryptographic techniques are commonly employed to protect model updates during aggregation. Despite growing interest in secure aggregation, existing works predominantly focus on protocol design and computational guarantees, with limited understanding of the fundamental information-theoretic limits of such systems. Moreover, optimal bounds on communication and key usage remain unknown in decentralized settings, where no central aggregator is available. Motivated by these gaps, we study the problem of decentralized secure aggregation (DSA) from an information-theoretic perspective. Specifically, we consider a network of $K$ fully-connected users, each holding a private input -- an abstraction of local training data -- who aim to securely compute the sum of all inputs. The security constraint requires that no user learns anything beyond the input sum, even when colluding with up to $T$ other users. We characterize the optimal rate region, which specifies the minimum achievable communication and secret key rates for DSA. In particular, we show that to securely compute one symbol of the desired input sum, each user must (i) transmit at least one symbol to others, (ii) hold at least one symbol of secret key, and (iii) all users must collectively hold no fewer than $K - 1$ independent key symbols. Our results establish the fundamental performance limits of DSA, providing insights for the design of provably secure and communication-efficient protocols in distributed learning systems.
Authors: Mingruo Yuan, Shuyi Zhang, Ben Kao
Abstract: Accurate confidence estimation is essential for trustworthy large language models (LLMs) systems, as it empowers the user to determine when to trust outputs and enables reliable deployment in safety-critical applications. Current confidence estimation methods for LLMs neglect the relevance between responses and contextual information, a crucial factor in output quality evaluation, particularly in scenarios where background knowledge is provided. To bridge this gap, we propose CRUX (Context-aware entropy Reduction and Unified consistency eXamination), the first framework that integrates context faithfulness and consistency for confidence estimation via two novel metrics. First, contextual entropy reduction represents data uncertainty with the information gain through contrastive sampling with and without context. Second, unified consistency examination captures potential model uncertainty through the global consistency of the generated answers with and without context. Experiments across three benchmark datasets (CoQA, SQuAD, QuAC) and two domain-specific datasets (BioASQ, EduQG) demonstrate CRUX's effectiveness, achieving the highest AUROC than existing baselines.
Authors: Francesco Panebianco, Stefano Bonfanti, Francesco Trov\`o, Michele Carminati
Abstract: The generalization capabilities of Large Language Models (LLMs) have led to their widespread deployment across various applications. However, this increased adoption has introduced several security threats, notably in the forms of jailbreaking and data leakage attacks. Additionally, Retrieval Augmented Generation (RAG), while enhancing context-awareness in LLM responses, has inadvertently introduced vulnerabilities that can result in the leakage of sensitive information. Our contributions are twofold. First, we introduce a methodology to analyze historical interaction data from an LLM system, enabling the generation of usage maps categorized by topics (including adversarial interactions). This approach further provides forensic insights for tracking the evolution of jailbreaking attack patterns. Second, we propose LeakSealer, a model-agnostic framework that combines static analysis for forensic insights with dynamic defenses in a Human-In-The-Loop (HITL) pipeline. This technique identifies topic groups and detects anomalous patterns, allowing for proactive defense mechanisms. We empirically evaluate LeakSealer under two scenarios: (1) jailbreak attempts, employing a public benchmark dataset, and (2) PII leakage, supported by a curated dataset of labeled LLM interactions. In the static setting, LeakSealer achieves the highest precision and recall on the ToxicChat dataset when identifying prompt injection. In the dynamic setting, PII leakage detection achieves an AUPRC of $0.97$, significantly outperforming baselines such as Llama Guard.
Authors: Natha\"el Da Costa, Marvin Pf\"ortner, Jon Cockayne
Abstract: Conditioning, the central operation in Bayesian statistics, is formalised by the notion of disintegration of measures. However, due to the implicit nature of their definition, constructing disintegrations is often difficult. A folklore result in machine learning conflates the construction of a disintegration with the restriction of probability density functions onto the subset of events that are consistent with a given observation. We provide a comprehensive set of mathematical tools which can be used to construct disintegrations and apply these to find densities of disintegrations on differentiable manifolds. Using our results, we provide a disturbingly simple example in which the restricted density and the disintegration density drastically disagree. Motivated by applications in approximate Bayesian inference and Bayesian inverse problems, we further study the modes of disintegrations. We show that the recently introduced notion of a "conditional mode" does not coincide in general with the modes of the conditional measure obtained through disintegration, but rather the modes of the restricted measure. We also discuss the implications of the discrepancy between the two measures in practice, advocating for the utility of both approaches depending on the modelling context.
Authors: Shantanu Thorat, Andrew Caines
Abstract: Existing AIG (AI-generated) text detectors struggle in real-world settings despite succeeding in internal testing, suggesting that they may not be robust enough. We rigorously examine the machine-learning procedure to build these detectors to address this. Most current AIG text detection datasets focus on zero-shot generations, but little work has been done on few-shot or one-shot generations, where LLMs are given human texts as an example. In response, we introduce the Diverse Adversarial Corpus of Texts Yielded from Language models (DACTYL), a challenging AIG text detection dataset focusing on one-shot/few-shot generations. We also include texts from domain-specific continued-pre-trained (CPT) language models, where we fully train all parameters using a memory-efficient optimization approach. Many existing AIG text detectors struggle significantly on our dataset, indicating a potential vulnerability to one-shot/few-shot and CPT-generated texts. We also train our own classifiers using two approaches: standard binary cross-entropy (BCE) optimization and a more recent approach, deep X-risk optimization (DXO). While BCE-trained classifiers marginally outperform DXO classifiers on the DACTYL test set, the latter excels on out-of-distribution (OOD) texts. In our mock deployment scenario in student essay detection with an OOD student essay dataset, the best DXO classifier outscored the best BCE-trained classifier by 50.56 macro-F1 score points at the lowest false positive rates for both. Our results indicate that DXO classifiers generalize better without overfitting to the test set. Our experiments highlight several areas of improvement for AIG text detectors.
Authors: Quentin Le Roux, Yannick Teglia, Teddy Furon, Philippe Loubet-Moundi
Abstract: Face Recognition Systems that operate in unconstrained environments capture images under varying conditions,such as inconsistent lighting, or diverse face poses. These challenges require including a Face Detection module that regresses bounding boxes and landmark coordinates for proper Face Alignment. This paper shows the effectiveness of Object Generation Attacks on Face Detection, dubbed Face Generation Attacks, and demonstrates for the first time a Landmark Shift Attack that backdoors the coordinate regression task performed by face detectors. We then offer mitigations against these vulnerabilities.
Authors: Chakattrai Sookkongwaree, Tattep Lakmuang, Chainarong Amornbunchornvej
Abstract: Understanding causal relationships in time series is fundamental to many domains, including neuroscience, economics, and behavioral science. Granger causality is one of the well-known techniques for inferring causality in time series. Typically, Granger causality frameworks have a strong fix-lag assumption between cause and effect, which is often unrealistic in complex systems. While recent work on variable-lag Granger causality (VLGC) addresses this limitation by allowing a cause to influence an effect with different time lags at each time point, it fails to account for the fact that causal interactions may vary not only in time delay but also across frequency bands. For example, in brain signals, alpha-band activity may influence another region with a shorter delay than slower delta-band oscillations. In this work, we formalize Multi-Band Variable-Lag Granger Causality (MB-VLGC) and propose a novel framework that generalizes traditional VLGC by explicitly modeling frequency-dependent causal delays. We provide a formal definition of MB-VLGC, demonstrate its theoretical soundness, and propose an efficient inference pipeline. Extensive experiments across multiple domains demonstrate that our framework significantly outperforms existing methods on both synthetic and real-world datasets, confirming its broad applicability to any type of time series data. Code and datasets are publicly available.
Authors: Maryam Mosleh, Marie Devlin, Ellis Solaiman
Abstract: Artificial intelligence-driven adaptive learning systems are reshaping education through data-driven adaptation of learning experiences. Yet many of these systems lack transparency, offering limited insight into how decisions are made. Most explainable AI (XAI) techniques focus on technical outputs but neglect user roles and comprehension. This paper proposes a hybrid framework that integrates traditional XAI techniques with generative AI models and user personalisation to generate multimodal, personalised explanations tailored to user needs. We redefine explainability as a dynamic communication process tailored to user roles and learning goals. We outline the framework's design, key XAI limitations in education, and research directions on accuracy, fairness, and personalisation. Our aim is to move towards explainable AI that enhances transparency while supporting user-centred experiences.
Authors: Wenxuan Wang, Zizhan Ma, Meidan Ding, Shiyi Zheng, Shengyuan Liu, Jie Liu, Jiaming Ji, Wenting Chen, Xiang Li, Linlin Shen, Yixuan Yuan
Abstract: The proliferation of Large Language Models (LLMs) in medicine has enabled impressive capabilities, yet a critical gap remains in their ability to perform systematic, transparent, and verifiable reasoning, a cornerstone of clinical practice. This has catalyzed a shift from single-step answer generation to the development of LLMs explicitly designed for medical reasoning. This paper provides the first systematic review of this emerging field. We propose a taxonomy of reasoning enhancement techniques, categorized into training-time strategies (e.g., supervised fine-tuning, reinforcement learning) and test-time mechanisms (e.g., prompt engineering, multi-agent systems). We analyze how these techniques are applied across different data modalities (text, image, code) and in key clinical applications such as diagnosis, education, and treatment planning. Furthermore, we survey the evolution of evaluation benchmarks from simple accuracy metrics to sophisticated assessments of reasoning quality and visual interpretability. Based on an analysis of 60 seminal studies from 2022-2025, we conclude by identifying critical challenges, including the faithfulness-plausibility gap and the need for native multimodal reasoning, and outlining future directions toward building efficient, robust, and sociotechnically responsible medical AI.
Authors: Banan Alkhateeb, Ellis Solaiman
Abstract: Social media platforms today strive to improve user experience through AI recommendations, yet the value of such recommendations vanishes as users do not understand the reasons behind them. This issue arises because explainability in social media is general and lacks alignment with user-specific needs. In this vision paper, we outline a user-segmented and context-aware explanation layer by proposing a visual explanation system with diverse explanation methods. The proposed system is framed by the variety of user needs and contexts, showing explanations in different visualized forms, including a technically detailed version for AI experts and a simplified one for lay users. Our framework is the first to jointly adapt explanation style (visual vs. numeric) and granularity (expert vs. lay) inside a single pipeline. A public pilot with 30 X users will validate its impact on decision-making and trust.
Authors: Shubham Kumar Nigam, Tanmay Dubey, Noel Shallum, Arnab Bhattacharya
Abstract: Legal precedent retrieval is a cornerstone of the common law system, governed by the principle of stare decisis, which demands consistency in judicial decisions. However, the growing complexity and volume of legal documents challenge traditional retrieval methods. TraceRetriever mirrors real-world legal search by operating with limited case information, extracting only rhetorically significant segments instead of requiring complete documents. Our pipeline integrates BM25, Vector Database, and Cross-Encoder models, combining initial results through Reciprocal Rank Fusion before final re-ranking. Rhetorical annotations are generated using a Hierarchical BiLSTM CRF classifier trained on Indian judgments. Evaluated on IL-PCR and COLIEE 2025 datasets, TraceRetriever addresses growing document volume challenges while aligning with practical search constraints, reliable and scalable foundation for precedent retrieval enhancing legal research when only partial case knowledge is available.
Authors: Shubham Kumar Nigam, Balaramamahanthi Deepak Patnaik, Shivam Mishra, Ajay Varghese Thomas, Noel Shallum, Kripabandhu Ghosh, Arnab Bhattacharya
Abstract: Legal Judgment Prediction (LJP) has emerged as a key area in AI for law, aiming to automate judicial outcome forecasting and enhance interpretability in legal reasoning. While previous approaches in the Indian context have relied on internal case content such as facts, issues, and reasoning, they often overlook a core element of common law systems, which is reliance on statutory provisions and judicial precedents. In this work, we propose NyayaRAG, a Retrieval-Augmented Generation (RAG) framework that simulates realistic courtroom scenarios by providing models with factual case descriptions, relevant legal statutes, and semantically retrieved prior cases. NyayaRAG evaluates the effectiveness of these combined inputs in predicting court decisions and generating legal explanations using a domain-specific pipeline tailored to the Indian legal system. We assess performance across various input configurations using both standard lexical and semantic metrics as well as LLM-based evaluators such as G-Eval. Our results show that augmenting factual inputs with structured legal knowledge significantly improves both predictive accuracy and explanation quality.
Authors: Yuxiang Wan, Ryan Devera, Wenjie Zhang, Ju Sun
Abstract: We present FMPlug, a novel plug-in framework that enhances foundation flow-matching (FM) priors for solving ill-posed inverse problems. Unlike traditional approaches that rely on domain-specific or untrained priors, FMPlug smartly leverages two simple but powerful insights: the similarity between observed and desired objects and the Gaussianity of generative flows. By introducing a time-adaptive warm-up strategy and sharp Gaussianity regularization, FMPlug unlocks the true potential of domain-agnostic foundation models. Our method beats state-of-the-art methods that use foundation FM priors by significant margins, on image super-resolution and Gaussian deblurring.
Authors: Sarah Mercer, Daniel P. Martin, Phil Swatton
Abstract: Generative agents powered by Large Language Models demonstrate human-like characteristics through sophisticated natural language interactions. Their ability to assume roles and personalities based on predefined character biographies has positioned them as cost-effective substitutes for human participants in social science research. This paper explores the validity of such persona-based agents in representing human populations; we recreate the HEXACO personality inventory experiment by surveying 310 GPT-4 powered agents, conducting factor analysis on their responses, and comparing these results to the original findings presented by Ashton, Lee, & Goldberg in 2004. Our results found 1) a coherent and reliable personality structure was recoverable from the agents' responses demonstrating partial alignment to the HEXACO framework. 2) the derived personality dimensions were consistent and reliable within GPT-4, when coupled with a sufficiently curated population, and 3) cross-model analysis revealed variability in personality profiling, suggesting model-specific biases and limitations. We discuss the practical considerations and challenges encountered during the experiment. This study contributes to the ongoing discourse on the potential benefits and limitations of using generative agents in social science research and provides useful guidance on designing consistent and representative agent personas to maximise coverage and representation of human personality traits.
Authors: Sebastian Wind, Jeta Sopa, Daniel Truhn, Mahshad Lotfinia, Tri-Thien Nguyen, Keno Bressem, Lisa Adams, Mirabela Rusu, Harald K\"ostler, Gerhard Wellein, Andreas Maier, Soroosh Tayebi Arasteh
Abstract: Clinical decision-making in radiology increasingly benefits from artificial intelligence (AI), particularly through large language models (LLMs). However, traditional retrieval-augmented generation (RAG) systems for radiology question answering (QA) typically rely on single-step retrieval, limiting their ability to handle complex clinical reasoning tasks. Here we propose an agentic RAG framework enabling LLMs to autonomously decompose radiology questions, iteratively retrieve targeted clinical evidence from Radiopaedia, and dynamically synthesize evidence-based responses. We evaluated 24 LLMs spanning diverse architectures, parameter scales (0.5B to >670B), and training paradigms (general-purpose, reasoning-optimized, clinically fine-tuned), using 104 expert-curated radiology questions from previously established RSNA-RadioQA and ExtendedQA datasets. Agentic retrieval significantly improved mean diagnostic accuracy over zero-shot prompting (73% vs. 64%; P<0.001) and conventional online RAG (73% vs. 68%; P<0.001). The greatest gains occurred in mid-sized models (e.g., Mistral Large improved from 72% to 81%) and small-scale models (e.g., Qwen 2.5-7B improved from 55% to 71%), while very large models (>200B parameters) demonstrated minimal changes (<2% improvement). Additionally, agentic retrieval reduced hallucinations (mean 9.4%) and retrieved clinically relevant context in 46% of cases, substantially aiding factual grounding. Even clinically fine-tuned models exhibited meaningful improvements (e.g., MedGemma-27B improved from 71% to 81%), indicating complementary roles of retrieval and fine-tuning. These results highlight the potential of agentic frameworks to enhance factuality and diagnostic accuracy in radiology QA, particularly among mid-sized LLMs, warranting future studies to validate their clinical utility.
Authors: Prerana Ramkumar
Abstract: Generative Adversarial Networks (GANs) have achieved realistic super-resolution (SR) of images however, they lack semantic consistency and per-pixel confidence, limiting their credibility in critical remote sensing applications such as disaster response, urban planning and agriculture. This paper introduces Semantic and Uncertainty-Aware ESRGAN (SU-ESRGAN), the first SR framework designed for satellite imagery to integrate the ESRGAN, segmentation loss via DeepLabv3 for class detail preservation and Monte Carlo dropout to produce pixel-wise uncertainty maps. The SU-ESRGAN produces results (PSNR, SSIM, LPIPS) comparable to the Baseline ESRGAN on aerial imagery. This novel model is valuable in satellite systems or UAVs that use wide field-of-view (FoV) cameras, trading off spatial resolution for coverage. The modular design allows integration in UAV data pipelines for on-board or post-processing SR to enhance imagery resulting due to motion blur, compression and sensor limitations. Further, the model is fine-tuned to evaluate its performance on cross domain applications. The tests are conducted on two drone based datasets which differ in altitude and imaging perspective. Performance evaluation of the fine-tuned models show a stronger adaptation to the Aerial Maritime Drone Dataset, whose imaging characteristics align with the training data, highlighting the importance of domain-aware training in SR-applications.
Authors: Andrea Martin, Ian R. Manchester, Luca Furieri
Abstract: In high-stakes engineering applications, optimization algorithms must come with provable worst-case guarantees over a mathematically defined class of problems. Designing for the worst case, however, inevitably sacrifices performance on the specific problem instances that often occur in practice. We address the problem of augmenting a given linearly convergent algorithm to improve its average-case performance on a restricted set of target problems - for example, tailoring an off-the-shelf solver for model predictive control (MPC) for an application to a specific dynamical system - while preserving its worst-case guarantees across the entire problem class. Toward this goal, we characterize the class of algorithms that achieve linear convergence for classes of nonsmooth composite optimization problems. In particular, starting from a baseline linearly convergent algorithm, we derive all - and only - the modifications to its update rule that maintain its convergence properties. Our results apply to augmenting legacy algorithms such as gradient descent for nonconvex, gradient-dominated functions; Nesterov's accelerated method for strongly convex functions; and projected methods for optimization over polyhedral feasibility sets. We showcase effectiveness of the approach on solving optimization problems with tight iteration budgets in application to ill-conditioned systems of linear equations and MPC for linear systems.
Authors: Julian Lemmel, Manuel Kranzl, Adam Lamine, Philipp Neubauer, Radu Grosu, Sophie Neubauer
Abstract: This paper introduces a new approach for fine-tuning the predictions of structured state space models (SSMs) at inference time using real-time recurrent learning. While SSMs are known for their efficiency and long-range modeling capabilities, they are typically trained offline and remain static during deployment. Our method enables online adaptation by continuously updating model parameters in response to incoming data. We evaluate our approach for linear-recurrent-unit SSMs using a small carbon emission dataset collected from embedded automotive hardware. Experimental results show that our method consistently reduces prediction error online during inference, demonstrating its potential for dynamic, resource-constrained environments.
Authors: Youssef Ait El Mahjoub, Jean-Michel Fourneau, Salma Alouah
Abstract: Solving Markov Decision Processes (MDPs) remains a central challenge in sequential decision-making, especially when dealing with large state spaces and long-term optimization criteria. A key step in Bellman dynamic programming algorithms is the policy evaluation, which becomes computationally demanding in infinite-horizon settings such as average-reward or discounted-reward formulations. In the context of Markov chains, aggregation and disaggregation techniques have for a long time been used to reduce complexity by exploiting structural decompositions. In this work, we extend these principles to a structured class of MDPs. We define the Single-Input Superstate Decomposable Markov Decision Process (SISDMDP), which combines Chiu's single-input decomposition with Robertazzi's single-cycle recurrence property. When a policy induces this structure, the resulting transition graph can be decomposed into interacting components with centralized recurrence. We develop an exact and efficient policy evaluation method based on this structure. This yields a scalable solution applicable to both average and discounted reward MDPs.
Authors: Jim Smith, Richard J. Preen, Andrew McCarthy, Maha Albashir, Alba Crespi-Boixader, Shahzad Mumtaz, Christian Cole, James Liley, Jost Migenda, Simon Rogers, Yola Jones
Abstract: We present SACRO-ML, an integrated suite of open source Python tools to facilitate the statistical disclosure control (SDC) of machine learning (ML) models trained on confidential data prior to public release. SACRO-ML combines (i) a SafeModel package that extends commonly used ML models to provide ante-hoc SDC by assessing the vulnerability of disclosure posed by the training regime; and (ii) an Attacks package that provides post-hoc SDC by rigorously assessing the empirical disclosure risk of a model through a variety of simulated attacks after training. The SACRO-ML code and documentation are available under an MIT license at https://github.com/AI-SDC/SACRO-ML
Authors: Hanchi Ren, Jingjing Deng, Xianghua Xie
Abstract: Federated Learning (FL) is a widely adopted privacy-preserving machine learning approach where private data remains local, enabling secure computations and the exchange of local model gradients between local clients and third-party parameter servers. However, recent findings reveal that privacy may be compromised and sensitive information potentially recovered from shared gradients. In this study, we offer detailed analysis and a novel perspective on understanding the gradient leakage problem. These theoretical works lead to a new gradient leakage defense technique that secures arbitrary model architectures using a private key-lock module. Only the locked gradient is transmitted to the parameter server for global model aggregation. Our proposed learning method is resistant to gradient leakage attacks, and the key-lock module is designed and trained to ensure that, without the private information of the key-lock module: a) reconstructing private training data from the shared gradient is infeasible; and b) the global model's inference performance is significantly compromised. We discuss the theoretical underpinnings of why gradients can leak private information and provide theoretical proof of our method's effectiveness. We conducted extensive empirical evaluations with many models on several popular benchmarks, demonstrating the robustness of our proposed approach in both maintaining model performance and defending against gradient leakage attacks.
Authors: Gaotang Li, Danai Koutra, Yujun Yan
Abstract: We address the key challenge of size-induced distribution shifts in graph neural networks (GNNs) and their impact on the generalization of GNNs to larger graphs. Existing literature operates under diverse assumptions about distribution shifts, resulting in varying conclusions about the generalizability of GNNs. In contrast to prior work, we adopt a data-driven approach to identify and characterize the types of size-induced distribution shifts and explore their impact on GNN performance from a spectral standpoint, a perspective that has been largely underexplored. Leveraging the significant variance in graph sizes in real biological datasets, we analyze biological graphs and find that spectral differences, driven by subgraph patterns (e.g., average cycle length), strongly correlate with GNN performance on larger, unseen graphs. Based on these insights, we propose three model-agnostic strategies to enhance GNNs' awareness of critical subgraph patterns, identifying size-intensive attention as the most effective approach. Extensive experiments with six GNN architectures and seven model-agnostic strategies across five datasets show that our size-intensive attention strategy significantly improves graph classification on test graphs 2 to 10 times larger than the training graphs, boosting F1 scores by up to 8% over strong baselines.
Authors: Dehua Peng, Zhipeng Gui, Wenzhang Wei, Fa Li, Jie Gui, Huayi Wu, Jianya Gong
Abstract: As a pivotal branch of machine learning, manifold learning uncovers the intrinsic low-dimensional structure within complex nonlinear manifolds in high-dimensional space for visualization, classification, clustering, and gaining key insights. Although existing techniques have achieved remarkable successes, they suffer from extensive distortions of cluster structure, which hinders the understanding of underlying patterns. Scalability issues also limit their applicability for handling large-scale data. We hence propose a sampling-based Scalable manifold learning technique that enables Uniform and Discriminative Embedding, namely SUDE, for large-scale and high-dimensional data. It starts by seeking a set of landmarks to construct the low-dimensional skeleton of the entire data, and then incorporates the non-landmarks into the learned space based on the constrained locally linear embedding (CLLE). We empirically validated the effectiveness of SUDE on synthetic datasets and real-world benchmarks, and applied it to analyze single-cell data and detect anomalies in electrocardiogram (ECG) signals. SUDE exhibits distinct advantage in scalability with respect to data size and embedding dimension, and has promising performance in cluster separation, integrity, and global structure preservation. The experiments also demonstrate notable robustness in embedding quality as the sampling rate decreases.
Authors: Jesse Hoogland, George Wang, Matthew Farrugia-Roberts, Liam Carroll, Susan Wei, Daniel Murfet
Abstract: Deep learning involves navigating a high-dimensional loss landscape over the neural network parameter space. Over the course of training, complex computational structures form and re-form inside the neural network, leading to shifts in input/output behavior. It is a priority for the science of deep learning to uncover principles governing the development of neural network structure and behavior. Drawing on the framework of singular learning theory, we propose that model development is deeply linked to degeneracy in the local geometry of the loss landscape. We investigate this link by monitoring loss landscape degeneracy throughout training, as quantified by the local learning coefficient, for a transformer language model and an in-context linear regression transformer. We show that training can be divided into distinct periods of change in loss landscape degeneracy, and that these changes in degeneracy coincide with significant changes in the internal computational structure and the input/output behavior of the transformers. This finding provides suggestive evidence that degeneracy and development are linked in transformers, underscoring the potential of a degeneracy-based perspective for understanding modern deep learning.
Authors: Dong-Hwan Jang, Sangdoo Yun, Dongyoon Han
Abstract: This paper introduces an efficient fine-tuning method for large pre-trained models, offering strong in-distribution (ID) and out-of-distribution (OOD) performance. Breaking away from traditional practices that need a multitude of fine-tuned models for averaging, our approach employs significantly fewer models to achieve final weights yet yield superior accuracy. Drawing from key insights in the weight space of fine-tuned weights, we uncover a strong link between the performance and proximity to the center of weight space. Based on this, we introduce a method that approximates a center-close weight using only two fine-tuned models, applicable during or after training. Our innovative layer-wise weight averaging technique surpasses state-of-the-art model methods such as Model Soup, utilizing only two fine-tuned models. This strategy can be aptly coined Model Stock, highlighting its reliance on selecting a minimal number of models to draw a more optimized-averaged model. We demonstrate the efficacy of Model Stock with fine-tuned models based upon pre-trained CLIP architectures, achieving remarkable performance on both ID and OOD tasks on the standard benchmarks, all while barely bringing extra computational demands. Our code and pre-trained models are available at https://github.com/naver-ai/model-stock.
Authors: Meiyu Zhong, Ravi Tandon
Abstract: Certifiable robustness gives the guarantee that small perturbations around an input to a classifier will not change the prediction. There are two approaches to provide certifiable robustness to adversarial examples: a) explicitly training classifiers with small Lipschitz constants, and b) Randomized smoothing, which adds random noise to the input to create a smooth classifier. We propose SPLITZ, a practical and novel approach which leverages the synergistic benefits of both the above ideas into a single framework. Our main idea is to split a classifier into two halves, constrain the Lipschitz constant of the first half, and smooth the second half via randomization. Motivation for SPLITZ comes from the observation that many standard deep networks exhibit heterogeneity in Lipschitz constants across layers. SPLITZ can exploit this heterogeneity while inheriting the scalability of randomized smoothing. We present a principled approach to train SPLITZ and provide theoretical analysis to derive certified robustness guarantees during inference. We present a comprehensive comparison of robustness-accuracy trade-offs and show that SPLITZ consistently improves on existing state-of-the-art approaches in the MNIST, CIFAR-10 and ImageNet datasets. For instance, with $\ell_2$ norm perturbation budget of $\epsilon=1$, SPLITZ achieves $43.2\%$ top-1 test accuracy on CIFAR-10 dataset compared to state-of-art top-1 test accuracy $39.8\%$.
Authors: Xianliang Xu, Ting Du, Wang Kong, Bin Shan, Ye Li, Zhongyi Huang
Abstract: The optimization algorithms are crucial in training physics-informed neural networks (PINNs), as unsuitable methods may lead to poor solutions. Compared to the common gradient descent (GD) algorithm, implicit gradient descent (IGD) outperforms it in handling certain multi-scale problems. In this paper, we provide convergence analysis for the IGD in training over-parameterized two-layer PINNs. We first derive the training dynamics of IGD in training two-layer PINNs. Then, over-parameterization allows us to prove that the randomly initialized IGD converges to a globally optimal solution at a linear convergence rate. Moreover, due to the distinct training dynamics of IGD compared to GD, the learning rate can be selected independently of the sample size and the least eigenvalue of the Gram matrix. Additionally, the novel approach used in our convergence analysis imposes a milder requirement on the network width. Finally, empirical results validate our theoretical findings.
Authors: Ties Robroek (IT University of Copenhagen), Neil Kim Nielsen (IT University of Copenhagen), P{\i}nar T\"oz\"un (IT University of Copenhagen)
Abstract: Training deep learning models is a repetitive and resource-intensive process. Data scientists often train several models before landing on a set of parameters (e.g., hyper-parameter tuning) and model architecture (e.g., neural architecture search), among other things that yield the highest accuracy. The computational efficiency of these training tasks depends highly on how well the training data is supplied to the training process. The repetitive nature of these tasks results in the same data processing pipelines running over and over, exacerbating the need for and costs of computational resources. In this paper, we present TensorSocket to reduce the computational needs of deep learning training by enabling simultaneous training processes to share the same data loader. TensorSocket mitigates CPU-side bottlenecks in cases where the collocated training workloads have high throughput on GPU, but are held back by lower data-loading throughput on CPU. TensorSocket achieves this by reducing redundant computations and data duplication across collocated training processes and leveraging modern GPU-GPU interconnects. While doing so, TensorSocket is able to train and balance differently-sized models and serve multiple batch sizes simultaneously and is hardware- and pipeline-agnostic in nature. Our evaluation shows that TensorSocket enables scenarios that are infeasible without data sharing, increases training throughput by up to 100%, and when utilizing cloud instances, achieves cost savings of 50% by reducing the hardware resource needs on the CPU side. Furthermore, TensorSocket outperforms the state-of-the-art solutions for shared data loading such as CoorDL and Joader; it is easier to deploy and maintain and either achieves higher or matches their throughput while requiring fewer CPU resources.
Authors: Vineet Jain, Tara Akhound-Sadegh, Siamak Ravanbakhsh
Abstract: Energy-based policies offer a flexible framework for modeling complex, multimodal behaviors in reinforcement learning (RL). In maximum entropy RL, the optimal policy is a Boltzmann distribution derived from the soft Q-function, but direct sampling from this distribution in continuous action spaces is computationally intractable. As a result, existing methods typically use simpler parametric distributions, like Gaussians, for policy representation -- limiting their ability to capture the full complexity of multimodal action distributions. In this paper, we introduce a diffusion-based approach for sampling from energy-based policies, where the negative Q-function defines the energy function. Based on this approach, we propose an actor-critic method called Diffusion Q-Sampling (DQS) that enables more expressive policy representations, allowing stable learning in diverse environments. We show that our approach enhances sample efficiency in continuous control tasks and captures multimodal behaviors, addressing key limitations of existing methods.
Authors: Zhi Wen Soi, Chenrui Fan, Aditya Shankar, Abele M\u{a}lan, Lydia Y. Chen
Abstract: Distributed time series data presents a challenge for federated learning, as clients often possess different feature sets and have misaligned time steps. Existing federated time series models are limited by the assumption of perfect temporal or feature alignment across clients. In this paper, we propose FedTDD, a novel federated time series diffusion model that jointly learns a synthesizer across clients. At the core of FedTDD is a novel data distillation and aggregation framework that reconciles the differences between clients by imputing the misaligned timesteps and features. In contrast to traditional federated learning, FedTDD learns the correlation across clients' time series through the exchange of local synthetic outputs instead of model parameters. A coordinator iteratively improves a global distiller network by leveraging shared knowledge from clients through the exchange of synthetic data. As the distiller becomes more refined over time, it subsequently enhances the quality of the clients' local feature estimates, allowing each client to then improve its local imputations for missing data using the latest, more accurate distiller. Experimental results on five datasets demonstrate FedTDD's effectiveness compared to centralized training, and the effectiveness of sharing synthetic outputs to transfer knowledge of local time series. Notably, FedTDD achieves 79.4% and 62.8% improvement over local training in Context-FID and Correlational scores.
Authors: Lucas Correia, Jan-Christoph Goos, Thomas B\"ack, Anna V. Kononova
Abstract: Benchmarking anomaly detection approaches for multivariate time series is a challenging task due to a lack of high-quality datasets. Current publicly available datasets are too small, not diverse and feature trivial anomalies, which hinders measurable progress in this research area. We propose a solution: a diverse, extensive, and non-trivial dataset generated via state-of-the-art simulation tools that reflects realistic behaviour of an automotive powertrain, including its multivariate, dynamic and variable-state properties. Additionally, our dataset represents a discrete-sequence problem, which remains unaddressed by previously-proposed solutions in literature. To cater for both unsupervised and semi-supervised anomaly detection settings, as well as time series generation and forecasting, we make different versions of the dataset available, where training and test subsets are offered in contaminated and clean versions, depending on the task. We also provide baseline results from a selection of approaches based on deterministic and variational autoencoders, as well as a non-parametric approach. As expected, the baseline experimentation shows that the approaches trained on the semi-supervised version of the dataset outperform their unsupervised counterparts, highlighting a need for approaches more robust to contaminated training data. Furthermore, results show that the threshold used can have a large influence on detection performance, hence more work needs to be invested in methods to find a suitable threshold without the need for labelled data.
Authors: Zixian Su, Jingwei Guo, Xi Yang, Qiufeng Wang, Kaizhu Huang
Abstract: Deploying deep models in real-world scenarios remains challenging due to significant performance drops under distribution shifts between training and deployment environments. Test-Time Adaptation (TTA) has recently emerged as a promising solution, enabling on-the-fly model adaptation without access to source data. However, its effectiveness degrades significantly in the presence of complex, mixed distribution shifts - common in practical settings - where multiple latent domains coexist. Adapting under such intrinsic heterogeneity, especially in unlabeled and online conditions, remains an open and underexplored challenge. In this paper, we study TTA under mixed distribution shifts and move beyond conventional homogeneous adaptation paradigms. By revisiting TTA from a frequency-domain perspective, we observe that distribution heterogeneity often manifests in Fourier space - for instance, high-frequency components tend to carry domain-specific variations. This motivates us to perform domain-aware separation using high-frequency texture cues, making diverse shift patterns more tractable. To this end, we propose FreDA, a novel Frequency-based Decentralized Adaptation framework that decomposes globally heterogeneous data into locally homogeneous components in the frequency domain. It further employs decentralized learning and augmentation strategies to robustly adapt under complex, evolving shifts. Extensive experiments across various environments (corrupted, natural, and medical) demonstrate the superiority of our proposed framework over the state-of-the-arts.
Authors: Yusheng Zhao, Xiao Luo, Haomin Wen, Zhiping Xiao, Wei Ju, Ming Zhang
Abstract: Traffic flow forecasting aims to predict future traffic flows based on the historical traffic conditions and the road network. It is an important problem in intelligent transportation systems, with a plethora of methods been proposed. Existing efforts mainly focus on capturing and utilizing spatio-temporal dependencies to predict future traffic flows. Though promising, they fall short in adapting to test-time environmental changes of traffic conditions. To tackle this challenge, we propose to introduce large language models (LLMs) to help traffic flow forecasting and design a novel method named Large Language Model Enhanced Traffic Flow Predictor (LEAF). LEAF adopts two branches, capturing different spatio-temporal relations using graph and hypergraph structures respectively. The two branches are first pre-trained individually, and during test-time, they yield different predictions. Based on these predictions, a large language model is used to select the most likely result. Then, a ranking loss is applied as the learning objective to enhance the prediction ability of the two branches. Extensive experiments on several datasets demonstrate the effectiveness of the proposed LEAF.
Authors: Eric Hirsch, Christian Friedrich
Abstract: Accurate tool wear prediction is essential for maintaining productivity and minimizing costs in machining. However, the complex nature of the tool wear process poses significant challenges to achieving reliable predictions. This study explores data-driven methods, in particular deep learning, for tool wear prediction. Traditional data-driven approaches often focus on a single process, relying on multi-sensor setups and extensive data generation, which limits generalization to new settings. Moreover, multi-sensor integration is often impractical in industrial environments. To address these limitations, this research investigates the transferability of predictive models using minimal training data, validated across two processes. Furthermore, it uses a simple setup with a single acceleration sensor to establish a low-cost data generation approach that facilitates the generalization of models to other processes via transfer learning. The study evaluates several machine learning models, including transformer-inspired convolutional neural networks (CNN), long short-term memory networks (LSTM), support vector machines (SVM), and decision trees, trained on different input formats such as feature vectors and short-time Fourier transform (STFT). The performance of the models is evaluated on two machines and on different amounts of training data, including scenarios with significantly reduced datasets, providing insight into their effectiveness under constrained data conditions. The results demonstrate the potential of specific models and configurations for effective tool wear prediction, contributing to the development of more adaptable and efficient predictive maintenance strategies in machining. Notably, the ConvNeXt model has an exceptional performance, achieving 99.1\% accuracy in identifying tool wear using data from only four milling tools operated until they are worn.
Authors: Jiayi Wang, Juan C. Alfaro, Viktor Bengs
Abstract: The label ranking problem is a supervised learning scenario in which the learner predicts a total order of the class labels for a given input instance. Recently, research has increasingly focused on the partial label ranking problem, a generalization of the label ranking problem that allows ties in the predicted orders. So far, most existing learning approaches for the partial label ranking problem rely on approximation algorithms for rank aggregation in the final prediction step. This paper explores several alternative aggregation methods for this critical step, including scoring-based and non-parametric probabilistic-based rank aggregation approaches. To enhance their suitability for the more general partial label ranking problem, the investigated methods are extended to increase the likelihood of producing ties. Experimental evaluations on standard benchmarks demonstrate that scoring-based variants consistently outperform the current state-of-the-art method in handling incomplete information. In contrast, non-parametric probabilistic-based variants fail to achieve competitive performance.
Authors: Yuanyuan Xu, Wenjie Zhang, Ying Zhang, Xuemin Lin, Xiwei Xu
Abstract: Dynamic Text-Attributed Graphs (DyTAGs) are a novel graph paradigm that captures evolving temporal events (edges) alongside rich textual attributes. Existing studies can be broadly categorized into TGNN-driven and LLM-driven approaches, both of which encode textual attributes and temporal structures for DyTAG representation. We observe that DyTAGs inherently comprise three distinct modalities: temporal, textual, and structural, often exhibiting completely disjoint distributions. However, the first two modalities are largely overlooked by existing studies, leading to suboptimal performance. To address this, we propose MoMent, a multi-modal model that explicitly models, integrates, and aligns each modality to learn node representations for link prediction. Given the disjoint nature of the original modality distributions, we first construct modality-specific features and encode them using individual encoders to capture correlations across temporal patterns, semantic context, and local structures. Each encoder generates modality-specific tokens, which are then fused into comprehensive node representations with a theoretical guarantee. To avoid disjoint subspaces of these heterogeneous modalities, we propose a dual-domain alignment loss that first aligns their distributions globally and then fine-tunes coherence at the instance level. This enhances coherent representations from temporal, textual, and structural views. Extensive experiments across seven datasets show that MoMent achieves up to 17.28% accuracy improvement and up to 31x speed-up against eight baselines.
Authors: Mingqi Yuan, Bo Li, Xin Jin, Wenjun Zeng
Abstract: Hyperparameter optimization (HPO) is a billion-dollar problem in machine learning, which significantly impacts the training efficiency and model performance. However, achieving efficient and robust HPO in deep reinforcement learning (RL) is consistently challenging due to its high non-stationarity and computational cost. To tackle this problem, existing approaches attempt to adapt common HPO techniques (e.g., population-based training or Bayesian optimization) to the RL scenario. However, they remain sample-inefficient and computationally expensive, which cannot facilitate a wide range of applications. In this paper, we propose ULTHO, an ultra-lightweight yet powerful framework for fast HPO in deep RL within single runs. Specifically, we formulate the HPO process as a multi-armed bandit with clustered arms (MABC) and link it directly to long-term return optimization. ULTHO also provides a quantified and statistical perspective to filter the HPs efficiently. We test ULTHO on benchmarks including ALE, Procgen, MiniGrid, and PyBullet. Extensive experiments demonstrate that the ULTHO can achieve superior performance with a simple architecture, contributing to the development of advanced and automated RL systems.
Authors: Leonard Waldmann, Ando Shah, Yi Wang, Nils Lehmann, Adam J. Stewart, Zhitong Xiong, Xiao Xiang Zhu, Stefan Bauer, John Chuang
Abstract: Earth observation (EO) data features diverse sensing platforms with varying spectral bands, spatial resolutions, and sensing modalities. While most prior work has constrained inputs to fixed sensors, a new class of any-sensor foundation models able to process arbitrary sensors has recently emerged. Contributing to this line of work, we propose Panopticon, an any-sensor foundation model built on the DINOv2 framework. We extend DINOv2 by (1) treating images of the same geolocation across sensors as natural augmentations, (2) subsampling channels to diversify spectral input, and (3) adding a cross attention over channels as a flexible patch embedding mechanism. By encoding the wavelength and modes of optical and synthetic aperture radar sensors, respectively, Panopticon can effectively process any combination of arbitrary channels. In extensive evaluations, we achieve state-of-the-art performance on GEO-Bench, especially on the widely-used Sentinel-1 and Sentinel-2 sensors, while out-competing other any-sensor models, as well as domain adapted fixed-sensor models on unique sensor configurations. Panopticon enables immediate generalization to both existing and future satellite platforms, advancing sensor-agnostic EO.
Authors: Juhyeong Kim, Sungyoon Choi, Youngbin Lee, Yejin Kim, Yongmin Choi, Yongjae Lee
Abstract: We propose Decision by Supervised Learning (DSL), a practical framework for robust portfolio optimization. DSL reframes portfolio construction as a supervised learning problem: models are trained to predict optimal portfolio weights, using cross-entropy loss and portfolios constructed by maximizing the Sharpe or Sortino ratio. To further enhance stability and reliability, DSL employs Deep Ensemble methods, substantially reducing variance in portfolio allocations. Through comprehensive backtesting across diverse market universes and neural architectures, shows superior performance compared to both traditional strategies and leading machine learning-based methods, including Prediction-Focused Learning and End-to-End Learning. We show that increasing the ensemble size leads to higher median returns and more stable risk-adjusted performance. The code is available at https://github.com/DSLwDE/DSLwDE.
Authors: Harvey Dam, Tripti Agarwal, Ganesh Gopalakrishnan
Abstract: Preserving topological features in learned latent spaces is a fundamental challenge in representation learning, particularly for topology-sensitive data. This paper introduces directional sign loss (DSL), an efficient, differentiable loss function that approximates the number of mismatches in the signs of finite differences between corresponding elements of two arrays. By penalizing discrepancies in critical points between input and reconstructed data, DSL encourages autoencoders and other learnable compressors to retain the topological features of the original data. We present the formulation and complexity analysis of DSL, comparing it to other non-differentiable topological measures. Experiments on multidimensional array data show that combining DSL with traditional loss functions preserves topological features more effectively than traditional losses alone. DSL serves as a differentiable, efficient proxy for common topology-based metrics, enabling topological feature preservation on previously impractical problem sizes and in a wider range of gradient-based optimization frameworks.
Authors: Anxian Liu, Junying Ma, Guang Zhang
Abstract: Financial time series forecasting in zero-shot settings is critical for investment decisions, especially during abrupt market regime shifts or in emerging markets with limited historical data. While Model-Agnostic Meta-Learning (MAML) approaches show promise, existing meta-task construction strategies often yield suboptimal performance for highly turbulent financial series. To address this, we propose a novel task-construction method that leverages learned embeddings for both meta task and also downstream predictions, enabling effective zero-shot meta-learning. Specifically, we use Gaussian Mixture Models (GMMs) to softly cluster embeddings, constructing two complementary meta-task types: intra-cluster tasks and inter-cluster tasks. By assigning embeddings to multiple latent regimes probabilistically, GMMs enable richer, more diverse meta-learning. This dual approach ensures the model can quickly adapt to local patterns while simultaneously capturing invariant cross-series features. Furthermore, we enhance inter-cluster generalization through hard task mining, which identifies robust patterns across divergent market regimes. Our method was validated using real-world financial data from high-volatility periods and multiple international markets (including emerging markets). The results demonstrate significant out-performance over existing approaches and stronger generalization in zero-shot scenarios.
Authors: Wenxin Chen, Weishen Pan, Kyra Gan, Fei Wang
Abstract: Current subgroup identification methods typically follow a two-step approach: first estimate conditional average treatment effects and then apply thresholding or rule-based procedures to define subgroups. While intuitive, this decoupled approach fails to incorporate key constraints essential for real-world clinical decision-making, such as subgroup size and propensity overlap. These constraints operate on fundamentally different axes than CATE estimation and are not naturally accommodated within existing frameworks, thereby limiting the practical applicability of these methods. We propose a unified optimization framework that directly solves the primal constrained optimization problem to identify optimal subgroups. Our key innovation is a reformulation of the constrained primal problem as an unconstrained differentiable min-max objective, solved via a gradient descent-ascent algorithm. We theoretically establish that our solution converges to a feasible and locally optimal solution. Unlike threshold-based CATE methods that apply constraints as post-hoc filters, our approach enforces them directly during optimization. The framework is model-agnostic, compatible with a wide range of CATE estimators, and extensible to additional constraints like cost limits or fairness criteria. Extensive experiments on synthetic and real-world datasets demonstrate its effectiveness in identifying high-benefit subgroups while maintaining better satisfaction of constraints.
Authors: David Ramos, Lucas Lacasa, Eusebio Valero, Gonzalo Rubio
Abstract: The main objective of this paper is to introduce a transfer learning-enhanced deep reinforcement learning (DRL) methodology that is able to optimise the geometry of any airfoil based on concomitant aerodynamic and structural integrity criteria. To showcase the method, we aim to maximise the lift-to-drag ratio $C_L/C_D$ while preserving the structural integrity of the airfoil -- as modelled by its maximum thickness -- and train the DRL agent using a list of different transfer learning (TL) strategies. The performance of the DRL agent is compared with Particle Swarm Optimisation (PSO), a traditional gradient-free optimisation method. Results indicate that DRL agents are able to perform purely aerodynamic and hybrid aerodynamic/structural shape optimisation, that the DRL approach outperforms PSO in terms of computational efficiency and aerodynamic improvement, and that the TL-enhanced DRL agent achieves performance comparable to the DRL one, while further saving substantial computational resources.
Authors: Willem Diepeveen, Jon Schwenk, Andrea Bertozzi
Abstract: We present Latent Diffeomorphic Dynamic Mode Decomposition (LDDMD), a new data reduction approach for the analysis of non-linear systems that combines the interpretability of Dynamic Mode Decomposition (DMD) with the predictive power of Recurrent Neural Networks (RNNs). Notably, LDDMD maintains simplicity, which enhances interpretability, while effectively modeling and learning complex non-linear systems with memory, enabling accurate predictions. This is exemplified by its successful application in streamflow prediction.
Authors: Yoonhyuk Choi, Jiho Choi, Chong-Kwon Kim
Abstract: Recent Graph Neural Networks (GNNs) combine spectral-spatial architectures for enhanced representation learning. However, limited attention has been paid to certified robustness, particularly regarding training strategies and underlying rationale. In this paper, we explicitly specialize each branch: the spectral network is trained to withstand l0 edge flips and capture homophilic structures, while the spatial part is designed to resist linf feature perturbations and heterophilic patterns. A context-aware gating network adaptively fuses the two representations, dynamically routing each node's prediction to the more reliable branch. This specialized adversarial training scheme uses branch-specific inner maximization (structure vs feature attacks) and a unified alignment objective. We provide theoretical guarantees: (i) expressivity of the gating mechanism beyond 1-WL, (ii) spectral-spatial frequency bias, and (iii) certified robustness with trade-off. Empirically, SpecSphere attains state-of-the-art node classification accuracy and offers tighter certified robustness on real-world benchmarks.
Authors: Patrik Kenfack, Samira Ebrahimi Kahou, Ulrich A\"ivodji
Abstract: Transformer-based tabular foundation models have recently demonstrated promising in-context learning (ICL) performance on structured data, emerging as competitive alternatives to gradient-boosted trees. However, the fairness implications of this new paradigm remain largely unexplored. We present the first investigation of fairness in tabular ICL, evaluating three recently proposed foundation models -- TabPFNv2, TabICL, and TabDPT -- on multiple benchmark datasets. To mitigate biases, we explore three pre-processing fairness-enhancing methods: correlation removal (decorrelating input features from the sensitive attribute), group-balanced sample selection (ensuring equal representation of protected groups in context examples), and uncertainty-based sample selection (prioritizing context examples with high sensitive-attribute prediction uncertainty). Our experiments show that the uncertainty-based strategy consistently improves group fairness metrics (e.g., demographic parity, equalized odds, and equal opportunity) with minimal impact on predictive accuracy. We release our code to facilitate reproducibility (https://github.com/patrikken/Fair-TabICL)
Authors: Xvyuan Liu, Xiangfei Qiu, Xingjian Wu, Zhengyu Li, Chenjuan Guo, Jilin Hu, Bin Yang
Abstract: The forecasting of irregular multivariate time series (IMTS) is a critical task in domains like healthcare and climate science. However, this task faces two significant hurdles: 1) the inherent non-uniformity and missing data in IMTS complicate the modeling of temporal dynamics, and 2) existing methods often rely on computationally expensive architectures. To address these dual challenges, we introduce APN, a general and efficient forecasting framework. At the core of APN is a novel Time-Aware Patch Aggregation (TAPA) module that introduces an aggregation-based paradigm for adaptive patching, moving beyond the limitations of fixed-span segmentation and interpolation-based methods. TAPA first learns dynamic temporal boundaries to define data-driven segments. Crucially, instead of resampling or interpolating, it directly computes patch representations via a time-aware weighted aggregation of all raw observations, where weights are determined by each observation's temporal relevance to the segment. This approach provides two key advantages: it preserves data fidelity by avoiding the introduction of artificial data points and ensures complete information coverage by design.The resulting regularized and information-rich patch representations enable the use of a lightweight query module for historical context aggregation and a simple MLP for final prediction. Extensive experiments on multiple real-world datasets demonstrate that APN establishes a new state-of-the-art, significantly outperforming existing methods in both prediction accuracy and computational efficiency.
Authors: Tinghan Ye, Amira Hijazi, Pascal Van Hentenryck
Abstract: Accurate estimation of order fulfillment time is critical for e-commerce logistics, yet traditional rule-based approaches often fail to capture the inherent uncertainties in delivery operations. This paper introduces a novel framework for distributional forecasting of order fulfillment time, leveraging Conformal Predictive Systems and Cross Venn-Abers Predictors -- model-agnostic techniques that provide rigorous coverage or validity guarantees. The proposed machine learning methods integrate granular spatiotemporal features, capturing fulfillment location and carrier performance dynamics to enhance predictive accuracy. Additionally, a cost-sensitive decision rule is developed to convert probabilistic forecasts into reliable point predictions. Experimental evaluation on a large-scale industrial dataset demonstrates that the proposed methods generate competitive distributional forecasts, while machine learning-based point predictions significantly outperform the existing rule-based system -- achieving up to 14% higher prediction accuracy and up to 75% improvement in identifying late deliveries.
Authors: Honoka Anada, Tatsuya Kaneko, Shinya Takamaeda-Yamazaki
Abstract: Federated learning (FL) enables multiple clients to collaboratively train machine learning models without sharing local data. In particular, decentralized FL (DFL), where clients exchange models without a central server, has gained attention for mitigating communication bottlenecks. Evaluating participant contributions is crucial in DFL to incentivize active participation and enhance transparency. However, existing contribution evaluation methods for FL assume centralized settings and cannot be applied directly to DFL due to two challenges: the inaccessibility of each client to non-neighboring clients' models, and the necessity to trace how contributions propagate in conjunction with peer-to-peer model exchanges over time. To address these challenges, we propose TRIP-Shapley, a novel contribution evaluation method for DFL. TRIP-Shapley formulates the clients' overall contributions by tracing the propagation of the round-wise local contributions. In this way, TRIP-Shapley accurately reflects the delayed and gradual influence propagation, as well as allowing a lightweight coordinator node to estimate the overall contributions without collecting models, but based solely on locally observable contributions reported by each client. Experiments demonstrate that TRIP-Shapley is sufficiently close to the ground-truth Shapley value, is scalable to large-scale scenarios, and remains robust in the presence of dishonest clients.
Authors: Weijie Guan, Haohui Wang, Jian Kang, Lihui Liu, Dawei Zhou
Abstract: Graph learning has been crucial to many real-world tasks, but they are often studied with a closed-world assumption, with all possible labels of data known a priori. To enable effective graph learning in an open and noisy environment, it is critical to inform the model users when the model makes a wrong prediction to in-distribution data of a known class, i.e., misclassification detection or when the model encounters out-of-distribution from novel classes, i.e., out-of-distribution detection. This paper introduces Evidential Reasoning Network (EVINET), a framework that addresses these two challenges by integrating Beta embedding within a subjective logic framework. EVINET includes two key modules: Dissonance Reasoning for misclassification detection and Vacuity Reasoning for out-of-distribution detection. Extensive experiments demonstrate that EVINET outperforms state-of-the-art methods across multiple metrics in the tasks of in-distribution classification, misclassification detection, and out-of-distribution detection. EVINET demonstrates the necessity of uncertainty estimation and logical reasoning for misclassification detection and out-of-distribution detection and paves the way for open-world graph learning. Our code and data are available at https://github.com/SSSKJ/EviNET.
Authors: Hangyu Li, Hongyue Wu, Guodong Fan, Zhen Zhang, Shizhan Chen, Zhiyong Feng
Abstract: As privacy protection gains increasing importance, more models are being trained on edge devices and subsequently merged into the central server through Federated Learning (FL). However, current research overlooks the impact of network topology, physical distance, and data heterogeneity on edge devices, leading to issues such as increased latency and degraded model performance. To address these issues, we propose a new federated learning scheme on edge devices that called Federated Learning with Encrypted Data Sharing(FedEDS). FedEDS uses the client model and the model's stochastic layer to train the data encryptor. The data encryptor generates encrypted data and shares it with other clients. The client uses the corresponding client's stochastic layer and encrypted data to train and adjust the local model. FedEDS uses the client's local private data and encrypted shared data from other clients to train the model. This approach accelerates the convergence speed of federated learning training and mitigates the negative impact of data heterogeneity, making it suitable for application services deployed on edge devices requiring rapid convergence. Experiments results show the efficacy of FedEDS in promoting model performance.
Authors: Zhiyu Zhao, Haoxuan Li, Haifeng Zhang, Jun Wang, Francesco Faccio, J\"urgen Schmidhuber, Mengyue Yang
Abstract: When building a world model, a common assumption is that the environment has a single, unchanging underlying causal rule, like applying Newton's laws to every situation. In reality, what appears as a drifting causal mechanism is often the manifestation of a fixed underlying mechanism seen through a narrow observational window. This brings about a problem that, when building a world model, even subtle shifts in policy or environment states can alter the very observed causal mechanisms. In this work, we introduce the \textbf{Meta-Causal Graph} as world models, a minimal unified representation that efficiently encodes the transformation rules governing how causal structures shift across different latent world states. A single Meta-Causal Graph is composed of multiple causal subgraphs, each triggered by meta state, which is in the latent state space. Building on this representation, we introduce a \textbf{Causality-Seeking Agent} whose objectives are to (1) identify the meta states that trigger each subgraph, (2) discover the corresponding causal relationships by agent curiosity-driven intervention policy, and (3) iteratively refine the Meta-Causal Graph through ongoing curiosity-driven exploration and agent experiences. Experiments on both synthetic tasks and a challenging robot arm manipulation task demonstrate that our method robustly captures shifts in causal dynamics and generalizes effectively to previously unseen contexts.
Authors: Ramchalam Kinattinkara Ramakrishnan, Zhaocong Yuan, Shaojie Zhuo, Chen Feng, Yicheng Lin, Chenzheng Su, Xiaopeng Zhang
Abstract: Speculative decoding generally dictates having a small, efficient draft model that is either pretrained or distilled offline to a particular target model series, for instance, Llama or Qwen models. However, within online deployment settings, there are two major challenges: 1) usage of a target model that is incompatible with the draft model; 2) expectation of latency improvements over usage and time. In this work, we propose OmniDraft, a unified framework that enables a single draft model to operate with any target model and adapt dynamically to user data. We introduce an online n-gram cache with hybrid distillation fine-tuning to address the cross-vocabulary mismatch across draft and target models; and further improve decoding speed by leveraging adaptive drafting techniques. OmniDraft is particularly suitable for on-device LLM applications where model cost, efficiency and user customization are the major points of contention. This further highlights the need to tackle the above challenges and motivates the \textit{``one drafter for all''} paradigm. We showcase the proficiency of the OmniDraft framework by performing online learning on math reasoning, coding and text generation tasks. Notably, OmniDraft enables a single Llama-68M model to pair with various target models including Vicuna-7B, Qwen2-7B and Llama3-8B models for speculative decoding; and additionally provides up to 1.5-2x speedup.
Authors: Shiyi Liu, Buwen Liang, Yuetong Fang, Zixuan Jiang, Renjing Xu
Abstract: Recent advances in AI for science have highlighted the power of contrastive learning in bridging heterogeneous biological data modalities. Building on this paradigm, we propose HIPPO (HIerarchical Protein-Protein interaction prediction across Organisms), a hierarchical contrastive framework for protein-protein interaction(PPI) prediction, where protein sequences and their hierarchical attributes are aligned through multi-tiered biological representation matching. The proposed approach incorporates hierarchical contrastive loss functions that emulate the structured relationship among functional classes of proteins. The framework adaptively incorporates domain and family knowledge through a data-driven penalty mechanism, enforcing consistency between the learned embedding space and the intrinsic hierarchy of protein functions. Experiments on benchmark datasets demonstrate that HIPPO achieves state-of-the-art performance, outperforming existing methods and showing robustness in low-data regimes. Notably, the model demonstrates strong zero-shot transferability to other species without retraining, enabling reliable PPI prediction and functional inference even in less characterized or rare organisms where experimental data are limited. Further analysis reveals that hierarchical feature fusion is critical for capturing conserved interaction determinants, such as binding motifs and functional annotations. This work advances cross-species PPI prediction and provides a unified framework for interaction prediction in scenarios with sparse or imbalanced multi-species data.
Authors: Janna Lu
Abstract: Large language models (LLMs) have demonstrated remarkable capabilities across diverse tasks, but their ability to forecast future events remains understudied. A year ago, large language models struggle to come close to the accuracy of a human crowd. I evaluate state-of-the-art LLMs on 464 forecasting questions from Metaculus, comparing their performance against human superforecasters. Frontier models achieve Brier scores that ostensibly surpass the human crowd but still significantly underperform a group of superforecasters.
Authors: Prady Saligram, Tanvir Bhathal
Abstract: Air pollution from agricultural emissions is a significant yet often overlooked contributor to environmental and public health challenges. Traditional air quality forecasting models rely on physics-based approaches, which struggle to capture complex, nonlinear pollutant interactions. In this work, we explore forecasting N$_2$O agricultural emissions through evaluating popular architectures, and proposing two novel deep learning architectures, EmissionNet (ENV) and EmissionNet-Transformer (ENT). These models leverage convolutional and transformer-based architectures to extract spatial-temporal dependencies from high-resolution emissions data
Authors: Kexin Gu Baugh, Vincent Perreault, Matthew Baugh, Luke Dickens, Katsumi Inoue, Alessandra Russo
Abstract: Neural Disjunctive Normal Form (DNF) based models are powerful and interpretable approaches to neuro-symbolic learning and have shown promising results in classification and reinforcement learning settings without prior knowledge of the tasks. However, their performance is degraded by the thresholding of the post-training symbolic translation process. We show here that part of the performance degradation during translation is due to its failure to disentangle the learned knowledge represented in the form of the networks' weights. We address this issue by proposing a new disentanglement method; by splitting nodes that encode nested rules into smaller independent nodes, we are able to better preserve the models' performance. Through experiments on binary, multiclass, and multilabel classification tasks (including those requiring predicate invention), we demonstrate that our disentanglement method provides compact and interpretable logical representations for the neural DNF-based models, with performance closer to that of their pre-translation counterparts. Our code is available at https://github.com/kittykg/disentangling-ndnf-classification.
URLs: https://github.com/kittykg/disentangling-ndnf-classification.
Authors: Martin Krutsk\'y, Gustav \v{S}\'ir, Vyacheslav Kungurtsev, Georgios Korpas
Abstract: Physics-inspired graph neural networks (PI-GNNs) have been utilized as an efficient unsupervised framework for relaxing combinatorial optimization problems encoded through a specific graph structure and loss, reflecting dependencies between the problem's variables. While the framework has yielded promising results in various combinatorial problems, we show that the performance of PI-GNNs systematically plummets with an increasing density of the combinatorial problem graphs. Our analysis reveals an interesting phase transition in the PI-GNNs' training dynamics, associated with degenerate solutions for the denser problems, highlighting a discrepancy between the relaxed, real-valued model outputs and the binary-valued problem solutions. To address the discrepancy, we propose principled alternatives to the naive strategy used in PI-GNNs by building on insights from fuzzy logic and binarized neural networks. Our experiments demonstrate that the portfolio of proposed methods significantly improves the performance of PI-GNNs in increasingly dense settings.
Authors: Yiyuan Yang, Zichuan Liu, Lei Song, Kai Ying, Zhiguang Wang, Tom Bamford, Svitlana Vyetrenko, Jiang Bian, Qingsong Wen
Abstract: Time series anomaly detection is critical across various domains, yet current approaches often limit analysis to mere binary anomaly classification without detailed categorization or further explanatory reasoning. To address these limitations, we propose a novel task, Time-series Reasoning for Anomaly (Time-RA) that transforms classical time series anomaly detection from a discriminative into a generative, reasoning-intensive task leveraging Large Language Models (LLMs). Also, we introduce the first real-world multimodal benchmark dataset, RATs40K, explicitly annotated for anomaly reasoning, comprising approximately 40,000 samples across 10 real-world domains. Each sample includes numeric time series data, contextual text information, and visual representations, each annotated with fine-grained categories (14 types for univariate anomalies and 6 for multivariate anomalies) and structured explanatory reasoning. We develop a sophisticated annotation framework utilizing ensemble-generated labels refined through GPT-4-driven feedback, ensuring accuracy and interpretability. Extensive benchmarking of LLMs and multimodal LLMs demonstrates the capabilities and limitations of current models, highlighting the critical role of supervised fine-tuning. Our dataset and task pave the way for significant advancements in interpretable time series anomaly detection and reasoning. The code (https://github.com/yyysjz1997/Time-RA) and dataset (https://huggingface.co/datasets/Time-RA/RATs40K) have been fully open-sourced to support and accelerate future research in this area.
URLs: https://github.com/yyysjz1997/Time-RA), https://huggingface.co/datasets/Time-RA/RATs40K)
Authors: Trung Nguyen, Md Masud Rana, Farjana Tasnim Mukta, Chang-Guo Zhan, Duc Duy Nguyen
Abstract: Accurate prediction of blood-brain barrier permeability (BBBP) is essential for central nervous system (CNS) drug development. While graph neural networks (GNNs) have advanced molecular property prediction, they often rely on molecular topology and neglect the three-dimensional geometric information crucial for modeling transport mechanisms. This paper introduces the geometric multi-color message-passing graph neural network (GMC-MPNN), a novel framework that enhances standard message-passing architectures by explicitly incorporating atomic-level geometric features and long-range interactions. Our model constructs weighted colored subgraphs based on atom types to capture the spatial relationships and chemical context that govern BBB permeability. We evaluated GMC-MPNN on three benchmark datasets for both classification and regression tasks, using rigorous scaffold-based splitting to ensure a robust assessment of generalization. The results demonstrate that GMC-MPNN consistently outperforms existing state-of-the-art models, achieving superior performance in both classifying compounds as permeable/non-permeable (AUC-ROC of 0.9704 and 0.9685) and in regressing continuous permeability values (RMSE of 0.4609, Pearson correlation of 0.7759). An ablation study further quantified the impact of specific atom-pair interactions, revealing that the model's predictive power derives from its ability to learn from both common and rare, but chemically significant, functional motifs. By integrating spatial geometry into the graph representation, GMC-MPNN sets a new performance benchmark and offers a more accurate and generalizable tool for drug discovery pipelines.
Authors: David McAllister, Songwei Ge, Brent Yi, Chung Min Kim, Ethan Weber, Hongsuk Choi, Haiwen Feng, Angjoo Kanazawa
Abstract: Flow-based generative models, including diffusion models, excel at modeling continuous distributions in high-dimensional spaces. In this work, we introduce Flow Policy Optimization (FPO), a simple on-policy reinforcement learning algorithm that brings flow matching into the policy gradient framework. FPO casts policy optimization as maximizing an advantage-weighted ratio computed from the conditional flow matching loss, in a manner compatible with the popular PPO-clip framework. It sidesteps the need for exact likelihood computation while preserving the generative capabilities of flow-based models. Unlike prior approaches for diffusion-based reinforcement learning that bind training to a specific sampling method, FPO is agnostic to the choice of diffusion or flow integration at both training and inference time. We show that FPO can train diffusion-style policies from scratch in a variety of continuous control tasks. We find that flow-based models can capture multimodal action distributions and achieve higher performance than Gaussian policies, particularly in under-conditioned settings.
Authors: Ambarish Singh, Romila Pradhan
Abstract: Data quality plays a pivotal role in the predictive performance of machine learning (ML) tasks - a challenge amplified by the deluge of data sources available in modern organizations. Prior work in data discovery largely focus on metadata matching, semantic similarity or identifying tables that should be joined to answer a particular query, but do not consider source quality for high performance of the downstream ML task. This paper addresses the problem of determining the best subset of data sources that must be combined to construct the underlying training dataset for a given ML task. We propose SourceGrasp and SourceSplice, frameworks designed to efficiently select a suitable subset of sources that maximizes the utility of the downstream ML model. Both the algorithms rely on the core idea that sources (or their combinations) contribute differently to the task utility, and must be judiciously chosen. While SourceGrasp utilizes a metaheuristic based on a greediness criterion and randomization, the SourceSplice framework presents a source selection mechanism inspired from gene splicing - a core concept used in protein synthesis. We empirically evaluate our algorithms on three real-world datasets and synthetic datasets and show that, with significantly fewer subset explorations, SourceSplice effectively identifies subsets of data sources leading to high task utility. We also conduct studies reporting the sensitivity of SourceSplice to the decision choices under several settings.
Authors: Felix Kronenwett, Georg Maier, Thomas L\"angle
Abstract: Sensor-based sorting systems enable the physical separation of a material stream into two fractions. The sorting decision is based on the image data evaluation of the sensors used and is carried out using actuators. Various process parameters must be set depending on the properties of the material stream, the dimensioning of the system, and the required sorting accuracy. However, continuous verification and re-adjustment are necessary due to changing requirements and material stream compositions. In this paper, we introduce an approach for optimizing, recurrently monitoring and adjusting the process parameters of a sensor-based sorting system. Based on Bayesian Optimization, Gaussian process regression models are used as surrogate models to achieve specific requirements for system behavior with the uncertainties contained therein. This method minimizes the number of necessary experiments while simultaneously considering two possible optimization targets based on the requirements for both material output streams. In addition, uncertainties are considered during determining sorting accuracies in the model calculation. We evaluated the method with three example process parameters.
Authors: Soumyadeep Dhar, Kei Sen Fong, Mehul Motani
Abstract: Distilling large neural networks into simple, human-readable symbolic formulas is a promising path toward trustworthy and interpretable AI. However, this process is often brittle, as the complex functions learned by standard networks are poor targets for symbolic discovery, resulting in low-fidelity student models. In this work, we propose a novel training paradigm to address this challenge. Instead of passively distilling a pre-trained network, we introduce a \textbf{Jacobian-based regularizer} that actively encourages the ``teacher'' network to learn functions that are not only accurate but also inherently smoother and more amenable to distillation. We demonstrate through extensive experiments on a suite of real-world regression benchmarks that our method is highly effective. By optimizing the regularization strength for each problem, we improve the $R^2$ score of the final distilled symbolic model by an average of \textbf{120\% (relative)} compared to the standard distillation pipeline, all while maintaining the teacher's predictive accuracy. Our work presents a practical and principled method for significantly improving the fidelity of interpretable models extracted from complex neural networks.
Authors: Peter Sharpe
Abstract: When numerically evaluating a function's gradient, sparsity detection can enable substantial computational speedups through Jacobian coloring and compression. However, sparsity detection techniques for black-box functions are limited, and existing finite-difference-based methods suffer from false negatives due to coincidental zero gradients. These false negatives can silently corrupt gradient calculations, leading to difficult-to-diagnose errors. We introduce NaN-propagation, which exploits the universal contamination property of IEEE 754 Not-a-Number values to trace input-output dependencies through floating-point numerical computations. By systematically contaminating inputs with NaN and observing which outputs become NaN, the method reconstructs conservative sparsity patterns that eliminate a major source of false negatives. We demonstrate this approach on an aerospace wing weight model, achieving a 1.52x speedup while uncovering dozens of dependencies missed by conventional methods -- a significant practical improvement since gradient computation is often the bottleneck in optimization workflows. The technique leverages IEEE 754 compliance to work across programming languages and math libraries without requiring modifications to existing black-box codes. Furthermore, advanced strategies such as NaN payload encoding via direct bit manipulation enable faster-than-linear time complexity, yielding speed improvements over existing black-box sparsity detection methods. Practical algorithms are also proposed to mitigate challenges from branching code execution common in engineering applications.
Authors: Yinhui Ma, Tomomasa Yamasaki, Zhehui Wang, Tao Luo, Bo Wang
Abstract: Hardware-Aware Neural Architecture Search (HW-NAS) is an efficient approach to automatically co-optimizing neural network performance and hardware energy efficiency, making it particularly useful for the development of Deep Neural Network accelerators on the edge. However, the extensive search space and high computational cost pose significant challenges to its practical adoption. To address these limitations, we propose Coflex, a novel HW-NAS framework that integrates the Sparse Gaussian Process (SGP) with multi-objective Bayesian optimization. By leveraging sparse inducing points, Coflex reduces the GP kernel complexity from cubic to near-linear with respect to the number of training samples, without compromising optimization performance. This enables scalable approximation of large-scale search space, substantially decreasing computational overhead while preserving high predictive accuracy. We evaluate the efficacy of Coflex across various benchmarks, focusing on accelerator-specific architecture. Our experimental results show that Coflex outperforms state-of-the-art methods in terms of network accuracy and Energy-Delay-Product, while achieving a computational speed-up ranging from 1.9x to 9.5x.
Authors: Jun Lu
Abstract: In 1954, Alston S. Householder published Principles of Numerical Analysis, one of the first modern treatments on matrix decomposition that favored a (block) LU decomposition-the factorization of a matrix into the product of lower and upper triangular matrices. And now, matrix decomposition has become a core technology in machine learning, largely due to the development of the backpropagation algorithm in fitting a neural network. The sole aim of this survey is to give a self-contained introduction to concepts and mathematical tools in numerical linear algebra and matrix analysis in order to seamlessly introduce matrix decomposition techniques and their applications in subsequent sections. However, we clearly realize our inability to cover all the useful and interesting results concerning matrix decomposition, given the paucity of scope to present this discussion, e.g., the separated analysis of the Euclidean space, Hermitian space, Hilbert space, and things in the complex domain. We refer the reader to literature in the field of linear algebra for a more detailed introduction to the related fields.
Authors: Soumyadip Sarkar
Abstract: Quantum generative models use the intrinsic probabilistic nature of quantum mechanics to learn and reproduce complex probability distributions. In this paper, we present an implementation of a 3-qubit quantum circuit Born machine trained to model a 3-bit Gaussian distribution using a Kullback-Leibler (KL) divergence loss and parameter-shift gradient optimization. The variational quantum circuit consists of layers of parameterized rotations and entangling gates, and is optimized such that the Born rule output distribution closely matches the target distribution. We detail the mathematical formulation of the model distribution, the KL divergence cost function, and the parameter-shift rule for gradient evaluation. Training results on a statevector simulator show that the KL divergence is minimized to near zero, and the final generated distribution aligns quantitatively with the target probabilities. We analyze the convergence behavior and discuss the implications for scalability and quantum advantage. Our results demonstrate the feasibility of small-scale quantum generative learning and provide insight into the training dynamics of quantum circuit models.
Authors: Yuchao Cai, Hanfang Yang, Yuheng Ma, Hanyuan Hang
Abstract: We consider the paradigm of unsupervised anomaly detection, which involves the identification of anomalies within a dataset in the absence of labeled examples. Though distance-based methods are top-performing for unsupervised anomaly detection, they suffer heavily from the sensitivity to the choice of the number of the nearest neighbors. In this paper, we propose a new distance-based algorithm called bagged regularized $k$-distances for anomaly detection (BRDAD), converting the unsupervised anomaly detection problem into a convex optimization problem. Our BRDAD algorithm selects the weights by minimizing the surrogate risk, i.e., the finite sample bound of the empirical risk of the bagged weighted $k$-distances for density estimation (BWDDE). This approach enables us to successfully address the sensitivity challenge of the hyperparameter choice in distance-based algorithms. Moreover, when dealing with large-scale datasets, the efficiency issues can be addressed by the incorporated bagging technique in our BRDAD algorithm. On the theoretical side, we establish fast convergence rates of the AUC regret of our algorithm and demonstrate that the bagging technique significantly reduces the computational complexity. On the practical side, we conduct numerical experiments to illustrate the insensitivity of the parameter selection of our algorithm compared with other state-of-the-art distance-based methods. Furthermore, our method achieves superior performance on real-world datasets with the introduced bagging technique compared to other approaches.
Authors: Quirin Vogel
Abstract: We prove a large deviation principle for deep neural networks with Gaussian weights and at most linearly growing activation functions, such as ReLU. This generalises earlier work, in which bounded and continuous activation functions were considered. In practice, linearly growing activation functions such as ReLU are most commonly used. We furthermore simplify previous expressions for the rate function and provide a power-series expansions for the ReLU case.
Authors: Ricardo Blum, Munir Hiabu, Enno Mammen, Joseph Theo Meyer
Abstract: Random Forests are widely claimed to capture interactions well. However, some simple examples suggest that they perform poorly in the presence of certain pure interactions that the conventional CART criterion struggles to capture during tree construction. Motivated from this, it is argued that simple alternative partitioning schemes used in the tree growing procedure can enhance identification of these interactions. In a simulation study these variants are compared to conventional Random Forests and Extremely Randomized Trees. The results validate that the modifications considered enhance the model's fitting ability in scenarios where pure interactions play a crucial role. Finally, the methods are applied to real datasets.
Authors: Xin Hao, Bahareh Nakisa, Mohmmad Naim Rastgoo, Gaoyang Pang
Abstract: Deep reinforcement Learning (DRL) offers a powerful framework for training AI agents to coordinate with human partners. However, DRL faces two critical challenges in human-AI coordination (HAIC): sparse rewards and unpredictable human behaviors. These challenges significantly limit DRL to identify effective coordination policies, due to its impaired capability of optimizing exploration and exploitation. To address these limitations, we propose an innovative behavior- and context-aware reward (BCR) for DRL, which optimizes exploration and exploitation by leveraging human behaviors and contextual information in HAIC. Our BCR consists of two components: (i) A novel dual intrinsic rewarding scheme to enhance exploration. This scheme composes an AI self-motivated intrinsic reward and a human-motivated intrinsic reward, which are designed to increase the capture of sparse rewards by a logarithmic-based strategy; and (ii) A new context-aware weighting mechanism for the designed rewards to improve exploitation. This mechanism helps the AI agent prioritize actions that better coordinate with the human partner by utilizing contextual information that can reflect the evolution of learning. Extensive simulations in the Overcooked environment demonstrate that our approach can increase the cumulative sparse rewards by approximately 20%, and improve the sample efficiency by around 38% compared to state-of-the-art baselines.
Authors: Matteo Bergamaschi, Andrea Cristofari, Vyacheslav Kungurtsev, Francesco Rinaldi
Abstract: For statistical modeling wherein the data regime is unfavorable in terms of dimensionality relative to the sample size, finding hidden sparsity in the ground truth can be critical in formulating an accurate statistical model. The so-called "l0 norm" which counts the number of non-zero components in a vector, is a strong reliable mechanism of enforcing sparsity when incorporated into an optimization problem for minimizing the fit of a given model to a set of observations. However, in big data settings wherein noisy estimates of the gradient must be evaluated out of computational necessity, the literature is scant on methods that reliably converge. In this paper we present an approach towards solving expectation objective optimization problems with cardinality constraints. We prove convergence of the underlying stochastic process, and demonstrate the performance on two Machine Learning problems.
Authors: Yaojun Zhang, Lanpeng Ji, Georgios Aivaliotis, Charles C. Taylor
Abstract: This paper proposes three types of Bayesian CART (or BCART) models for aggregate claim amount, namely, frequency-severity models, sequential models and joint models. We propose a general framework for the BCART models applicable to data with multivariate responses, which is particularly useful for the joint BCART models with a bivariate response: the number of claims and aggregate claim amount. To facilitate frequency-severity modeling, we investigate BCART models for the right-skewed and heavy-tailed claim severity data by using various distributions. We discover that the Weibull distribution is superior to gamma and lognormal distributions, due to its ability to capture different tail characteristics in tree models. Additionally, we find that sequential BCART models and joint BCART models, which incorporate dependence between the number of claims and average severity, are beneficial and thus preferable to the frequency-severity BCART models in which independence is assumed. The effectiveness of these models' performance is illustrated by carefully designed simulations and real insurance data.
Authors: Evan Rose, Hidde Lycklama, Harsh Chaudhari, Anwar Hithnawi, Alina Oprea
Abstract: Privacy-preserving machine learning (PPML) enables multiple data owners to contribute their data privately to a set of servers that run a secure multi-party computation (MPC) protocol to train a joint ML model. In these protocols, the input data remains private throughout the training process, and only the resulting model is made available. While this approach benefits privacy, it also exacerbates the risks of data poisoning, where compromised data owners induce undesirable model behavior by contributing malicious datasets. Existing MPC mechanisms can mitigate certain poisoning attacks, but these measures are not exhaustive. To complement existing poisoning defenses, we introduce UTrace: a framework for User-level Traceback of poisoning attacks in PPML. Utrace computes user responsibility scores using gradient similarity metrics aggregated across the most relevant samples in an owner's dataset. UTrace is effective at low poisoning rates and is resilient to poisoning attacks distributed across multiple data owners, unlike existing unlearning-based methods. We introduce methods for checkpointing gradients with low storage overhead, enabling traceback in the absence of data owners at deployment time. We also design several optimizations that reduce traceback time and communication in MPC. We provide a comprehensive evaluation of UTrace across four datasets from three data modalities (vision, text, and malware) and show its effectiveness against 10 poisoning attacks.
Authors: Michelle S. Lam, Fred Hohman, Dominik Moritz, Jeffrey P. Bigham, Kenneth Holstein, Mary Beth Kery
Abstract: AI policy sets boundaries on acceptable behavior for AI models, but this is challenging in the context of large language models (LLMs): how do you ensure coverage over a vast behavior space? We introduce policy maps, an approach to AI policy design inspired by the practice of physical mapmaking. Instead of aiming for full coverage, policy maps aid effective navigation through intentional design choices about which aspects to capture and which to abstract away. With Policy Projector, an interactive tool for designing LLM policy maps, an AI practitioner can survey the landscape of model input-output pairs, define custom regions (e.g., "violence"), and navigate these regions with if-then policy rules that can act on LLM outputs (e.g., if output contains "violence" and "graphic details," then rewrite without "graphic details"). Policy Projector supports interactive policy authoring using LLM classification and steering and a map visualization reflecting the AI practitioner's work. In an evaluation with 12 AI safety experts, our system helps policy designers craft policies around problematic model behaviors such as incorrect gender assumptions and handling of immediate physical safety threats.
Authors: Ryan Diaz, Adam Imdieke, Vivek Veeriah, Karthik Desingh
Abstract: Operating in unstructured environments like households requires robotic policies that are robust to out-of-distribution conditions. Although much work has been done in evaluating robustness for visuomotor policies, the robustness evaluation of a multisensory approach that includes force-torque sensing remains largely unexplored. This work introduces a novel, factor-based evaluation framework with the goal of assessing the robustness of multisensory policies in a peg-in-hole assembly task. To this end, we develop a multisensory policy framework utilizing the Perceiver IO architecture to learn the task. We investigate which factors pose the greatest generalization challenges in object assembly and explore a simple multisensory data augmentation technique to enhance out-of-distribution performance. We provide a simulation environment enabling controlled evaluation of these factors. Our results reveal that multisensory variations such as Grasp Pose present the most significant challenges for robustness, and naive unisensory data augmentation applied independently to each sensory modality proves insufficient to overcome them. Additionally, we find force-torque sensing to be the most informative modality for our contact-rich assembly task, with vision being the least informative. Finally, we briefly discuss supporting real-world experimental results. For additional experiments and qualitative results, we refer to the project webpage https://rpm-lab-umn.github.io/auginsert/ .
Authors: Zixian Su, Jingwei Guo, Xi Yang, Qiufeng Wang, Frans Coenen, Kaizhu Huang
Abstract: Medical Image Analysis (MedIA) has become indispensable in modern healthcare, enhancing clinical diagnostics and personalized treatment. Despite the remarkable advancements supported by deep learning (DL) technologies, their practical deployment faces challenges due to distribution shifts, where models trained on specific datasets underperform across others from varying hospitals, regions, or patient populations. To navigate this issue, researchers have been actively developing strategies to increase the adaptability and robustness of DL models, enabling their effective use in unfamiliar and diverse environments. This paper systematically reviews approaches that apply DL techniques to MedIA systems affected by distribution shifts. Unlike traditional categorizations based on technical specifications, our approach is grounded in the real-world operational constraints faced by healthcare institutions. Specifically, we categorize the existing body of work into Joint Training, Federated Learning, Fine-tuning, and Domain Generalization, with each method tailored to distinct scenarios caused by Data Accessibility, Privacy Concerns, and Collaborative Protocols. This perspective equips researchers with a nuanced understanding of how DL can be strategically deployed to address distribution shifts in MedIA, ensuring diverse and robust medical applications. By delving deeper into these topics, we highlight potential pathways for future research that not only address existing limitations but also push the boundaries of deployable MedIA technologies.
Authors: Xiaoling Hu, Xiangrui Zeng, Oula Puonti, Juan Eugenio Iglesias, Bruce Fischl, Yael Balbastre
Abstract: Domain randomization through synthesis is a powerful strategy to train networks that are unbiased with respect to the domain of the input images. Randomization allows networks to see a virtually infinite range of intensities and artifacts during training, thereby minimizing overfitting to appearance and maximizing generalization to unseen data. Although powerful, this approach relies on the accurate tuning of a large set of hyperparameters that govern the probabilistic distribution of the synthesized images. Instead of manually tuning these parameters, we introduce Learn2Synth, a novel procedure in which synthesis parameters are learned using a small set of real labeled data. Unlike methods that impose constraints to align synthetic data with real data (e.g., contrastive or adversarial techniques), which risk misaligning the image and its label map, we tune an augmentation engine such that a segmentation network trained on synthetic data has optimal accuracy when applied to real data. This approach allows the training procedure to benefit from real labeled examples, without ever using these real examples to train the segmentation network, which avoids biasing the network towards the properties of the training set. Specifically, we develop parametric and nonparametric strategies to enhance synthetic images in a way that improves the performance of the segmentation network. We demonstrate the effectiveness of this learning strategy on synthetic and real-world brain scans. Code is available at: https://github.com/HuXiaoling/Learn2Synth.
Authors: Taekyung Ki, Dongchan Min, Gyeongsu Chae
Abstract: With the rapid advancement of diffusion-based generative models, portrait image animation has achieved remarkable results. However, it still faces challenges in temporally consistent video generation and fast sampling due to its iterative sampling nature. This paper presents FLOAT, an audio-driven talking portrait video generation method based on flow matching generative model. Instead of a pixel-based latent space, we take advantage of a learned orthogonal motion latent space, enabling efficient generation and editing of temporally consistent motion. To achieve this, we introduce a transformer-based vector field predictor with an effective frame-wise conditioning mechanism. Additionally, our method supports speech-driven emotion enhancement, enabling a natural incorporation of expressive motions. Extensive experiments demonstrate that our method outperforms state-of-the-art audio-driven talking portrait methods in terms of visual quality, motion fidelity, and efficiency.
Authors: Junjie Luo, Abhimanyu Kumbara, Mansur Shomali, Rui Han, Anand Iyer, Ritu Agarwal, Gordon Gao
Abstract: Continuous glucose monitoring (CGM) combined with AI offers new opportunities for proactive diabetes management through real-time glucose forecasting. However, most existing models are task-specific and lack generalization across patient populations. Inspired by the autoregressive paradigm of large language models, we introduce CGM-LSM, a Transformer decoder-based Large Sensor Model (LSM) pretrained on 1.6 million CGM records from patients with different diabetes types, ages, and genders. We model patients as sequences of glucose time steps to learn latent knowledge embedded in CGM data and apply it to the prediction of glucose readings for a 2-hour horizon. Compared with prior methods, CGM-LSM significantly improves prediction accuracy and robustness: a 48.51% reduction in root mean square error in one-hour horizon forecasting and consistent zero-shot prediction performance across held-out patient groups. We analyze model performance variations across patient subgroups and prediction scenarios and outline key opportunities and challenges for advancing CGM foundation models.
Authors: Jinnan Guo, Kapil Vaswani, Andrew Paverd, Peter Pietzuch
Abstract: In federated learning (FL), data providers jointly train a model without disclosing their training data. Despite its inherent privacy benefits, a malicious data provider can simply deviate from the correct training protocol without being detected, potentially compromising the trained model. While current solutions have explored the use of trusted execution environments (TEEs) to combat such attacks, they usually assume side-channel attacks against the TEEs are out of scope. However, such side-channel attacks can undermine the security properties of TEE-based FL frameworks, not by extracting the FL data, but by leaking keys that allow the adversary to impersonate as the TEE whilst deviating arbitrarily from the correct training protocol. We describe ExclaveFL, an FL platform that provides end-to-end integrity and transparency, even in the presence of side-channel attacks on TEEs. We propose a new paradigm in which existing TEEs are used as exclaves -- integrity-protected execution environments that do not contain any secrets, making them immune to side-channel attacks. Whereas previous approaches attest the TEE itself and bind this attestation to a key held by the TEE, ExclaveFL attests individual data transformations at runtime. These runtime attestations form an attested dataflow graph, which can be checked to ensure the FL training job satisfies claims, such as deviations from the correct computation. We implement ExclaveFL by extending the popular NVFlare FL framework to use exclaves, and show experimentally that ExclaveFL introduces less than 10% overhead compared to the same FL framework without TEEs, whilst providing stronger security guarantees.
Authors: Zijiang Yan, Jianhua Pei, Hongda Wu, Hina Tabassum, Ping Wang
Abstract: This paper proposes a novel Semantic Communication (SemCom) framework for real-time adaptive-bitrate video streaming by integrating Latent Diffusion Models (LDMs) within the FFmpeg techniques. This solution addresses the challenges of high bandwidth usage, storage inefficiencies, and quality of experience (QoE) degradation associated with traditional Constant Bitrate Streaming (CBS) and Adaptive Bitrate Streaming (ABS). The proposed approach leverages LDMs to compress I-frames into a latent space, offering significant storage and semantic transmission savings without sacrificing high visual quality. While retaining B-frames and P-frames as adjustment metadata to support efficient refinement of video reconstruction at the user side, the proposed framework further incorporates state-of-the-art denoising and Video Frame Interpolation (VFI) techniques. These techniques mitigate semantic ambiguity and restore temporal coherence between frames, even in noisy wireless communication environments. Experimental results demonstrate the proposed method achieves high-quality video streaming with optimized bandwidth usage, outperforming state-of-the-art solutions in terms of QoE and resource efficiency. This work opens new possibilities for scalable real-time video streaming in 5G and future post-5G networks.
Authors: Felix Stollenwerk, Tobias Stollenwerk
Abstract: Despite their remarkable capabilities, LLMs learn word representations that exhibit the undesirable yet poorly understood feature of anisotropy. In this paper, we argue that the second moment in Adam is a cause of anisotropic embeddings, and suggest a modified optimizer called Coupled Adam to mitigate the problem. Our experiments demonstrate that Coupled Adam significantly improves the quality of embeddings, while also leading to better upstream and downstream performance on large enough datasets.
Authors: Gabriele Prato, Jerry Huang, Prasanna Parthasarathi, Shagun Sodhani, Sarath Chandar
Abstract: Large Language Models (LLMs) have emerged as highly capable systems and are increasingly being integrated into various uses. However, the rapid pace of their deployment has outpaced a comprehensive understanding of their internal mechanisms and a delineation of their capabilities and limitations. A desired attribute of an intelligent system is its ability to recognize the scope of its own knowledge. To investigate whether LLMs embody this characteristic, we develop a benchmark designed to challenge these models to enumerate all information they possess on specific topics. This benchmark evaluates whether the models recall excessive, insufficient, or the precise amount of information, thereby indicating their awareness of their own knowledge. Our findings reveal that all tested LLMs, given sufficient scale, demonstrate an understanding of how much they know about specific topics. While different architectures exhibit varying rates of this capability's emergence, the results suggest that awareness of knowledge may be a generalizable attribute of LLMs. Further research is needed to confirm this potential and fully elucidate the underlying mechanisms.
Authors: N. Richardson, C. Bosch, R. P. Adams
Abstract: Optical computing systems provide an alternate hardware model which appears to be aligned with the demands of neural network workloads. However, the challenge of implementing energy efficient nonlinearities in optics -- a key requirement for realizing neural networks -- is a conspicuous missing link. In this work we introduce a novel method to achieve nonlinear computation in fully linear media. Our method can operate at low power and requires only the ability to drive the optical system at a data-dependent spatial position. Leveraging this positional encoding, we formulate a fully automated, topology-optimization-based hardware design framework for extremely specialized optical neural networks, drawing on modern advancements in optimization and machine learning. We evaluate our optical designs on machine learning classification tasks: demonstrating significant improvements over linear methods, and competitive performance when compared to standard artificial neural networks.
Authors: Rui Melo, Claudia Mamede, Andre Catarino, Rui Abreu, Henrique Lopes Cardoso
Abstract: Software vulnerabilities such as buffer overflows and SQL injections are a major source of security breaches. Traditional methods for vulnerability detection remain essential but are limited by high false positive rates, scalability issues, and reliance on manual effort. These constraints have driven interest in AI-based approaches to automated vulnerability detection and secure code generation. While Large Language Models (LLMs) have opened new avenues for classification tasks, their complexity and opacity pose challenges for interpretability and deployment. Sparse Autoencoder offer a promising solution to this problem. We explore whether SAEs can serve as a lightweight, interpretable alternative for bug detection in Java functions. We evaluate the effectiveness of SAEs when applied to representations from GPT-2 Small and Gemma 2B, examining their capacity to highlight buggy behaviour without fine-tuning the underlying LLMs. We found that SAE-derived features enable bug detection with an F1 score of up to 89%, consistently outperforming fine-tuned transformer encoder baselines. Our work provides the first empirical evidence that SAEs can be used to detect software bugs directly from the internal representations of pretrained LLMs, without any fine-tuning or task-specific supervision. Code available at https://github.com/rufimelo99/SAE-Java-Bug-Detection
Authors: Sakshi Arya
Abstract: We study sequential decision-making in batched nonparametric contextual bandits, where actions are selected over a finite horizon divided into a small number of batches. Motivated by constraints in domains such as medicine and marketing -- where online feedback is limited -- we propose a nonparametric algorithm that combines adaptive k-nearest neighbor (k-NN) regression with the upper confidence bound (UCB) principle. Our method, BaNk-UCB, is fully nonparametric, adapts to the context dimension, and is simple to implement. Unlike prior work relying on parametric or binning-based estimators, BaNk-UCB uses local geometry to estimate rewards and adaptively balances exploration and exploitation. We provide near-optimal regret guarantees under standard Lipschitz smoothness and margin assumptions, using a theoretically motivated batch schedule that balances regret across batches and achieves minimax-optimal rates. Empirical evaluations on synthetic and real-world datasets demonstrate that BaNk-UCB consistently outperforms binning-based baselines.
Authors: Maida Wang, Xiao Xue, Peter V. Coveney
Abstract: Learning the behaviour of chaotic systems remains challenging due to instability in long-term predictions and difficulties in accurately capturing invariant statistical properties. While quantum machine learning offers a promising route to efficiently capture physical properties from high-dimensional data, its practical deployment is hindered by current hardware noise and limited scalability. Here, we introduce a quantum-informed machine learning framework for learning partial differential equations, with an application focus on chaotic systems. A quantum circuit Born machine is employed to learn the invariant properties of chaotic dynamical systems, achieving substantial memory efficiency by representing these complex physical statistics with a compact set of trainable circuit parameters. This approach reduces the data storage requirement by over two orders of magnitude compared to the raw simulation data. The resulting statistical quantum-informed prior is then incorporated into a Koopman-based auto-regressive model to address issues such as gradient vanishing or explosion, while maintaining long-term statistical fidelity. The framework is evaluated on three representative systems: the Kuramoto-Sivashinsky equation, two-dimensional Kolmogorov flow and turbulent channel flow. In all cases, the quantum-informed model achieves superior performance compared to its classical counterparts without quantum priors. This hybrid architecture offers a practical route for learning dynamical systems using near-term quantum hardware.
Authors: Federico Echenique, Alireza Fallah, Michael I. Jordan
Abstract: We propose a general methodology for recovering preference parameters from data on choices and response times. Our methods yield estimates with fast ($1/n$ for $n$ data points) convergence rates when specialized to the popular Drift Diffusion Model (DDM), but are broadly applicable to generalizations of the DDM as well as to alternative models of decision making that make use of response time data. The paper develops an empirical application to an experiment on intertemporal choice, showing that the use of response times delivers predictive accuracy and matters for the estimation of economically relevant parameters.
Authors: Hongzhe Bi, Lingxuan Wu, Tianwei Lin, Hengkai Tan, Zhizhong Su, Hang Su, Jun Zhu
Abstract: Imitation learning for robotic manipulation faces a fundamental challenge: the scarcity of large-scale, high-quality robot demonstration data. Recent robotic foundation models often pre-train on cross-embodiment robot datasets to increase data scale, while they face significant limitations as the diverse morphologies and action spaces across different robot embodiments make unified training challenging. In this paper, we present H-RDT (Human to Robotics Diffusion Transformer), a novel approach that leverages human manipulation data to enhance robot manipulation capabilities. Our key insight is that large-scale egocentric human manipulation videos with paired 3D hand pose annotations provide rich behavioral priors that capture natural manipulation strategies and can benefit robotic policy learning. We introduce a two-stage training paradigm: (1) pre-training on large-scale egocentric human manipulation data, and (2) cross-embodiment fine-tuning on robot-specific data with modular action encoders and decoders. Built on a diffusion transformer architecture with 2B parameters, H-RDT uses flow matching to model complex action distributions. Extensive evaluations encompassing both simulation and real-world experiments, single-task and multitask scenarios, as well as few-shot learning and robustness assessments, demonstrate that H-RDT outperforms training from scratch and existing state-of-the-art methods, including Pi0 and RDT, achieving significant improvements of 13.9% and 40.5% over training from scratch in simulation and real-world experiments, respectively. The results validate our core hypothesis that human manipulation data can serve as a powerful foundation for learning bimanual robotic manipulation policies.