Authors: Zijian Zhang, Rong Wang, Shiyang Li, Yuebo Luo, Mingyi Hong, Caiwen Ding
Abstract: Developing efficient CUDA kernels is increasingly critical for AI applications such as large-scale LLM training. However, manual kernel design is both costly and time-consuming, motivating automatic approaches that leverage LLMs for code generation. Existing methods for automatic kernel generation, however, often produce low-efficiency kernels, incur high computational overhead, and fail to generalize across settings. In this work, we propose CudaForge, a training-free multi-agent workflow for CUDA kernel generation and optimization. Our workflow is inspired by the iterative workflow of human experts, which contains steps such as developing initial kernels, testing correctness, analyzing hardware feedback, and iterative improvement. More specifically, CudaForge employs two LLM agents: a Coder and a Judge, that iteratively generate, correct, and optimize CUDA kernels, while integrating hardware feedback such as Nsight Compute (NCU) metrics. In extensive evaluations, we show that CudaForge, by leveraging base models like OpenAI-o3, achieves 97.6\% correctness of generated kernels and an average 1.68$\times$ speedup over PyTorch baselines, substantially surpassing state-of-the-art models including OpenAI-o3 and Kevin on KernelBench. Beyond accuracy and speed, CudaForge demonstrates strong generalization across GPUs (A100, RTX 6000, 4090, 3090) and base models (OpenAI-o3, GPT-5, gpt-oss-120B, Claude-Sonnet-4, QwQ-32B), while maintaining high efficiency. In particular, generating an optimized kernel takes about 26.5 minutes on one RTX6000 and incurs about \$ 0.3 API cost, which is significantly cheaper than existing agentic work that costs 6 H100 hours and \$ 5 API cost per kernel. Our results highlight that multi-agent, training-free workflows can enable cost-effective, generalizable, and high-performance CUDA kernel optimization. Code available at https://github.com/OptimAI-Lab/CudaForge
Authors: Ruibo Hou, Shiyu Teng, Jiaqing Liu, Shurong Chai, Yinhao Li, Lanfen Lin, Yen-Wei Chen
Abstract: Multimodal deep learning has shown promise in depression detection by integrating text, audio, and video signals. Recent work leverages sentiment analysis to enhance emotional understanding, yet suffers from high computational cost, domain mismatch, and static knowledge limitations. To address these issues, we propose a novel Retrieval-Augmented Generation (RAG) framework. Given a depression-related text, our method retrieves semantically relevant emotional content from a sentiment dataset and uses a Large Language Model (LLM) to generate an Emotion Prompt as an auxiliary modality. This prompt enriches emotional representation and improves interpretability. Experiments on the AVEC 2019 dataset show our approach achieves state-of-the-art performance with CCC of 0.593 and MAE of 3.95, surpassing previous transfer learning and multi-task learning baselines.
Authors: Doulaye Demb\'el\'e
Abstract: Classification is a machine learning method used in many practical applications: text mining, handwritten character recognition, face recognition, pattern classification, scene labeling, computer vision, natural langage processing. A classifier prediction results and training set information are often used to get a contingency table which is used to quantify the method quality through an evaluation measure. Such measure, typically a numerical value, allows to choose a suitable method among several. Many evaluation measures available in the literature are less accurate for a dataset with imbalanced classes. In this paper, the eigenvalues entropy is used as an evaluation measure for a binary or a multi-class problem. For a binary problem, relations are given between the eigenvalues and some commonly used measures, the sensitivity, the specificity, the area under the operating receiver characteristic curve and the Gini index. A by-product result of this paper is an estimate of the confusion matrix to deal with the curse of the imbalanced classes. Various data examples are used to show the better performance of the proposed evaluation measure over the gold standard measures available in the literature.
Authors: Zhiwen Li, Cheuk Hin Ho, Lok Ming Lui
Abstract: Traditional methods for high-dimensional diffeomorphic mapping often struggle with the curse of dimensionality. We propose a mesh-free learning framework designed for $n$-dimensional mapping problems, seamlessly combining variational principles with quasi-conformal theory. Our approach ensures accurate, bijective mappings by regulating conformality distortion and volume distortion, enabling robust control over deformation quality. The framework is inherently compatible with gradient-based optimization and neural network architectures, making it highly flexible and scalable to higher-dimensional settings. Numerical experiments on both synthetic and real-world medical image data validate the accuracy, robustness, and effectiveness of the proposed method in complex registration scenarios.
Authors: Ahmet Erdem Pamuk, Emir Kaan \"Ozdemir, \c{S}uayp Talha Kocabay
Abstract: Large language models (LLMs) are increasingly trained with classical optimization techniques like AdamW to improve convergence and generalization. However, the mechanisms by which quantum-inspired methods enhance classical training remain underexplored. We introduce Superpositional Gradient Descent (SGD), a novel optimizer linking gradient updates with quantum superposition by injecting quantum circuit perturbations. We present a mathematical framework and implement hybrid quantum-classical circuits in PyTorch and Qiskit. On synthetic sequence classification and large-scale LLM fine-tuning, SGD converges faster and yields lower final loss than AdamW. Despite promising results, scalability and hardware constraints limit adoption. Overall, this work provides new insights into the intersection of quantum computing and deep learning, suggesting practical pathways for leveraging quantum principles to control and enhance model behavior.
Authors: Seungwoo Yoo, Kyeongmin Yeo, Jisung Hwang, Minhyuk Sung
Abstract: We introduce Neural Green's Function, a neural solution operator for linear partial differential equations (PDEs) whose differential operators admit eigendecompositions. Inspired by Green's functions, the solution operators of linear PDEs that depend exclusively on the domain geometry, we design Neural Green's Function to imitate their behavior, achieving superior generalization across diverse irregular geometries and source and boundary functions. Specifically, Neural Green's Function extracts per-point features from a volumetric point cloud representing the problem domain and uses them to predict a decomposition of the solution operator, which is subsequently applied to evaluate solutions via numerical integration. Unlike recent learning-based solution operators, which often struggle to generalize to unseen source or boundary functions, our framework is, by design, agnostic to the specific functions used during training, enabling robust and efficient generalization. In the steady-state thermal analysis of mechanical part geometries from the MCB dataset, Neural Green's Function outperforms state-of-the-art neural operators, achieving an average error reduction of 13.9\% across five shape categories, while being up to 350 times faster than a numerical solver that requires computationally expensive meshing.
Authors: Yeqiu Chen, Ziyan Liu, Hong Wang
Abstract: Solving large-scale Generalized Eigenvalue Problems (GEPs) is a fundamental yet computationally prohibitive task in science and engineering. As a promising direction, contour integral (CI) methods, such as the CIRR algorithm, offer an efficient and parallelizable framework. However, their performance is critically dependent on the selection of integration contours -- improper selection without reliable prior knowledge of eigenvalue distribution can incur significant computational overhead and compromise numerical accuracy. To address this challenge, we propose DeepContour, a novel hybrid framework that integrates a deep learning-based spectral predictor with Kernel Density Estimation for principled contour design. Specifically, DeepContour first employs a Fourier Neural Operator (FNO) to rapidly predict the spectral distribution of a given GEP. Subsequently, Kernel Density Estimation (KDE) is applied to the predicted spectrum to automatically and systematically determine proper integration contours. Finally, these optimized contours guide the CI solver to efficiently find the desired eigenvalues. We demonstrate the effectiveness of our method on diverse challenging scientific problems. In our main experiments, DeepContour accelerates GEP solving across multiple datasets, achieving up to a 5.63$\times$ speedup. By combining the predictive power of deep learning with the numerical rigor of classical solvers, this work pioneers an efficient and robust paradigm for tackling difficult generalized eigenvalue involving matrices of high dimension.
Authors: Qingyue Long, Can Rong, Tong Li, Yong Li
Abstract: Human trajectory data is crucial in urban planning, traffic engineering, and public health. However, directly using real-world trajectory data often faces challenges such as privacy concerns, data acquisition costs, and data quality. A practical solution to these challenges is trajectory generation, a method developed to simulate human mobility behaviors. Existing trajectory generation methods mainly focus on capturing individual movement patterns but often overlook the influence of population distribution on trajectory generation. In reality, dynamic population distribution reflects changes in population density across different regions, significantly impacting individual mobility behavior. Thus, we propose a novel trajectory generation framework based on a diffusion model, which integrates the dynamic population distribution constraints to guide high-fidelity generation outcomes. Specifically, we construct a spatial graph to enhance the spatial correlation of trajectories. Then, we design a dynamic population distribution aware denoising network to capture the spatiotemporal dependencies of human mobility behavior as well as the impact of population distribution in the denoising process. Extensive experiments show that the trajectories generated by our model can resemble real-world trajectories in terms of some critical statistical metrics, outperforming state-of-the-art algorithms by over 54%.
Authors: Haoming Wang, Wei Gao
Abstract: Image generation models are usually personalized in practical uses in order to better meet the individual users' heterogeneous needs, but most personalized models lack explainability about how they are being personalized. Such explainability can be provided via visual features in generated images, but is difficult for human users to understand. Explainability in natural language is a better choice, but the existing approaches to explainability in natural language are limited to be coarse-grained. They are unable to precisely identify the multiple aspects of personalization, as well as the varying levels of personalization in each aspect. To address such limitation, in this paper we present a new technique, namely \textbf{FineXL}, towards \textbf{Fine}-grained e\textbf{X}plainability in natural \textbf{L}anguage for personalized image generation models. FineXL can provide natural language descriptions about each distinct aspect of personalization, along with quantitative scores indicating the level of each aspect of personalization. Experiment results show that FineXL can improve the accuracy of explainability by 56\%, when different personalization scenarios are applied to multiple types of image generation models.
Authors: Yirong Zeng, Xiao Ding, Yutai Hou, Yuxian Wang, Li Du, Juyi Dai, Qiuyang Ding, Duyu Tang, Dandan Tu, Weiwen Liu, Bing Qin, Ting Liu
Abstract: Training tool-augmented LLMs has emerged as a promising approach to enhancing language models' capabilities for complex tasks. The current supervised fine-tuning paradigm relies on constructing extensive domain-specific datasets to train models. However, this approach often struggles to generalize effectively to unfamiliar or intricate tool-use scenarios. Recently, reinforcement learning (RL) paradigm can endow LLMs with superior reasoning and generalization abilities. In this work, we address a key question: Can the pure RL be used to effectively elicit a model's intrinsic reasoning capabilities and enhance the tool-agnostic generalization? We propose a dynamic generalization-guided reward design for rule-based RL, which progressively shifts rewards from exploratory to exploitative tool-use patterns. Based on this design, we introduce the Tool-Zero series models. These models are trained to enable LLMs to autonomously utilize general tools by directly scaling up RL from Zero models (i.e., base models without post-training). Experimental results demonstrate that our models achieve over 7% performance improvement compared to both SFT and RL-with-SFT models under the same experimental settings. These gains are consistently replicated across cross-dataset and intra-dataset evaluations, validating the effectiveness and robustness of our methods.
Authors: Hasan Tutar, Caner Erden, \"Umit \c{S}ent\"urk
Abstract: The determination of sample size in qualitative research has traditionally relied on the subjective and often ambiguous principle of data saturation, which can lead to inconsistencies and threaten methodological rigor. This study introduces a new, systematic model based on machine learning (ML) to make this process more objective. Utilizing a dataset derived from five fundamental qualitative research approaches - namely, Case Study, Grounded Theory, Phenomenology, Narrative Research, and Ethnographic Research - we developed an ensemble learning model. Ten critical parameters, including research scope, information power, and researcher competence, were evaluated using an ordinal scale and used as input features. After thorough preprocessing and outlier removal, multiple ML algorithms were trained and compared. The K-Nearest Neighbors (KNN), Gradient Boosting (GB), Random Forest (RF), XGBoost, and Decision Tree (DT) algorithms showed the highest explanatory power (Test R2 ~ 0.85), effectively modeling the complex, non-linear relationships involved in qualitative sampling decisions. Feature importance analysis confirmed the vital roles of research design type and information power, providing quantitative validation of key theoretical assumptions in qualitative methodology. The study concludes by proposing a conceptual framework for a web-based computational application designed to serve as a decision support system for qualitative researchers, journal reviewers, and thesis advisors. This model represents a significant step toward standardizing sample size justification, enhancing transparency, and strengthening the epistemological foundation of qualitative inquiry through evidence-based, systematic decision-making.
Authors: Abdelaziz Bounhar, Hadi Abdine, Evan Dufraisse, Ahmad Chamma, Amr Mohamed, Dani Bouch, Michalis Vazirgiannis, Guokan Shang
Abstract: Large language models (LLMs) trained for step-by-step reasoning often become excessively verbose, raising inference cost. Standard Reinforcement Learning with Verifiable Rewards (RLVR) pipelines filter out ``easy'' problems for training efficiency, leaving the model to train primarily on harder problems that require longer reasoning chains. This skews the output length distribution upward, resulting in a \textbf{model that conflates ``thinking longer'' with ``thinking better''}. In this work, we show that retaining and modestly up-weighting moderately easy problems acts as an implicit length regularizer. Exposing the model to solvable short-chain tasks constrains its output distribution and prevents runaway verbosity. The result is \textbf{\emph{emergent brevity for free}}: the model learns to solve harder problems without inflating the output length, \textbf{ despite the absence of any explicit length penalization}. RLVR experiments using this approach on \textit{Qwen3-4B-Thinking-2507} (with a 16k token limit) achieve baseline pass@1 AIME25 accuracy while generating solutions that are, on average, nearly twice as short. The code is available at \href{https://github.com/MBZUAI-Paris/Frugal-AI}{GitHub}, with datasets and models on \href{https://huggingface.co/collections/MBZUAI-Paris/k2-think-mini-68dcfa8b114686a4bd3dc2bc}{Hugging Face}.
URLs: https://github.com/MBZUAI-Paris/Frugal-AI, https://huggingface.co/collections/MBZUAI-Paris/k2-think-mini-68dcfa8b114686a4bd3dc2bc
Authors: Tiberiu Musat
Abstract: Grokking is a puzzling phenomenon in neural networks where full generalization occurs only after a substantial delay following the complete memorization of the training data. Previous research has linked this delayed generalization to representation learning driven by weight decay, but the precise underlying dynamics remain elusive. In this paper, we argue that post-memorization learning can be understood through the lens of constrained optimization: gradient descent effectively minimizes the weight norm on the zero-loss manifold. We formally prove this in the limit of infinitesimally small learning rates and weight decay coefficients. To further dissect this regime, we introduce an approximation that decouples the learning dynamics of a subset of parameters from the rest of the network. Applying this framework, we derive a closed-form expression for the post-memorization dynamics of the first layer in a two-layer network. Experiments confirm that simulating the training process using our predicted gradients reproduces both the delayed generalization and representation learning characteristic of grokking.
Authors: Guillaume Tejedor (BDTLN), Veronika Peralta (BDTLN), Nicolas Labroche (LIFAT, BDTLN), Patrick Marcel (LIFO, Pamda), H\'el\`ene Blasco (UT), Hugo Alarcan (CHRU Tours)
Abstract: Amyotrophic lateral sclerosis (ALS) is a severe disease with a typical survival of 3-5 years after symptom onset. Current treatments offer only limited life extension, and the variability in patient responses highlights the need for personalized care. However, research is hindered by small, heterogeneous cohorts, sparse longitudinal data, and the lack of a clear definition for clinically meaningful patient clusters. Existing clustering methods remain limited in both scope and number. To address this, we propose a clustering approach that groups sequences using a disease progression declarative score. Our approach integrates medical expertise through multiple descriptive variables, investigating several distance measures combining such variables, both by reusing off-the-shelf distances and employing a weak-supervised learning method. We pair these distances with clustering methods and benchmark them against state-of-the-art techniques. The evaluation of our approach on a dataset of 353 ALS patients from the University Hospital of Tours, shows that our method outperforms state-of-the-art methods in survival analysis while achieving comparable silhouette scores. In addition, the learned distances enhance the relevance and interpretability of results for medical experts.
Authors: Zihan Li, Mingyang Wan, Mingyu Gao, Zhongshan Chen, Xiangke Wang, Feifan Zhang
Abstract: Covalent organic frameworks (COFs) are promising adsorbents for gas adsorption and separation, while identifying the optimal structures among their vast design space requires efficient high-throughput screening. Conventional machine-learning predictors rely heavily on specific gas-related features. However, these features are time-consuming and limit scalability, leading to inefficiency and labor-intensive processes. Herein, a universal COFs adsorption prediction framework (COFAP) is proposed, which can extract multi-modal structural and chemical features through deep learning, and fuse these complementary features via cross-modal attention mechanism. Without Henry coefficients or adsorption heat, COFAP sets a new SOTA by outperforming previous approaches on hypoCOFs dataset. Based on COFAP, we also found that high-performing COFs for separation concentrate within a narrow range of pore size and surface area. A weight-adjustable prioritization scheme is also developed to enable flexible, application-specific ranking of candidate COFs for researchers. Superior efficiency and accuracy render COFAP directly deployable in crystalline porous materials.
Authors: Md Abrar Hasnat, Md Jobayer, Md. Mehedi Hasan Shawon, Md. Golam Rabiul Alam
Abstract: Cardiovascular disease (CVD) remains a critical global health concern, demanding reliable and interpretable predictive models for early risk assessment. This study presents a large-scale analysis using the Heart Disease Health Indicators Dataset, developing a strategically weighted ensemble model that combines tree-based methods (LightGBM, XGBoost) with a Convolutional Neural Network (CNN) to predict CVD risk. The model was trained on a preprocessed dataset of 229,781 patients where the inherent class imbalance was managed through strategic weighting and feature engineering enhanced the original 22 features to 25. The final ensemble achieves a statistically significant improvement over the best individual model, with a Test AUC of 0.8371 (p=0.003) and is particularly suited for screening with a high recall of 80.0%. To provide transparency and clinical interpretability, surrogate decision trees and SHapley Additive exPlanations (SHAP) are used. The proposed model delivers a combination of robust predictive performance and clinical transparency by blending diverse learning architectures and incorporating explainability through SHAP and surrogate decision trees, making it a strong candidate for real-world deployment in public health screening.
Authors: Prasanth K K, Shubham Sharma
Abstract: Standard Recurrent Neural Networks, including LSTMs, struggle to model long-range dependencies, particularly in sequences containing noisy or misleading information. We propose a new architectural principle, Output-Conditioned Gating, which enables a model to perform self-reflection by modulating its internal memory gates based on its own past inferences. This creates a stabilizing feedback loop that enhances memory retention. Our final model, the EchoLSTM, integrates this principle with an attention mechanism. We evaluate the EchoLSTM on a series of challenging benchmarks. On a custom-designed Distractor Signal Task, the EchoLSTM achieves 69.0% accuracy, decisively outperforming a standard LSTM baseline by 33 percentage points. Furthermore, on the standard ListOps benchmark, the EchoLSTM achieves performance competitive with a modern Transformer model, 69.8% vs. 71.8%, while being over 5 times more parameter-efficient. A final Trigger Sensitivity Test provides qualitative evidence that our model's self-reflective mechanism leads to a fundamentally more robust memory system.
Authors: Manuel A. Hernandez Alonso, Michael Depass, Stephan Quessy, Numa Dancause, Ignasi Cos
Abstract: Electroencephalography (EEG) and local field potentials (LFP) are two widely used techniques to record electrical activity from the brain. These signals are used in both the clinical and research domains for multiple applications. However, most brain data recordings suffer from a myriad of artifacts and noise sources other than the brain itself. Thus, a major requirement for their use is proper and, given current volumes of data, a fully automatized conditioning. As a means to this end, here we introduce an unsupervised, multipurpose EEG/LFP preprocessing method, the NeuroClean pipeline. In addition to its completeness and reliability, NeuroClean is an unsupervised series of algorithms intended to mitigate reproducibility issues and biases caused by human intervention. The pipeline is designed as a five-step process, including the common bandpass and line noise filtering, and bad channel rejection. However, it incorporates an efficient independent component analysis with an automatic component rejection based on a clustering algorithm. This machine learning classifier is used to ensure that task-relevant information is preserved after each step of the cleaning process. We used several data sets to validate the pipeline. NeuroClean removed several common types of artifacts from the signal. Moreover, in the context of motor tasks of varying complexity, it yielded more than 97% accuracy (vs. a chance-level of 33.3%) in an optimized Multinomial Logistic Regression model after cleaning the data, compared to the raw data, which performed at 74% accuracy. These results show that NeuroClean is a promising pipeline and workflow that can be applied to future work and studies to achieve better generalization and performance on machine learning pipelines.
Authors: Donghee Lee, Hye-Sung Lee, Jaeok Yi
Abstract: We present the bulk-boundary decomposition as a new framework for understanding the training dynamics of deep neural networks. Starting from the stochastic gradient descent formulation, we show that the Lagrangian can be reorganized into a data-independent bulk term and a data-dependent boundary term. The bulk captures the intrinsic dynamics set by network architecture and activation functions, while the boundary reflects stochastic interactions from training samples at the input and output layers. This decomposition exposes the local and homogeneous structure underlying deep networks. As a natural extension, we develop a field-theoretic formulation of neural dynamics based on this decomposition.
Authors: Aditya Sridhar, Nish Sinnadurai, Sean Lie, Vithursan Thangarasa
Abstract: Speculative decoding accelerates LLMs by using a lightweight draft model to generate tokens autoregressively before verifying them in parallel with a larger target model. However, determining the optimal number of tokens to draft remains a key challenge limiting the approach's effectiveness. Dynamic speculative decoding aims to intelligently decide how many tokens to draft to achieve maximum speedups. Existing methods often rely on hand-tuned, sensitive thresholds (e.g., token entropy), which are costly to set and generalize poorly across models and domains. We propose TapOut, an online, training-free, plug-and-play algorithm for dynamic speculation policy selection using multi-armed bandits. Our approach employs a meta-algorithm that selects among multiple parameter-free dynamic speculation strategies based on past reward and exploration. We conduct extensive experiments across diverse model pairs and datasets, showing that TapOut achieves competitive or superior speedups compared to well-established dynamic speculation baselines without any hyperparameter tuning.
Authors: Daniel Aarao Reis Arturi, Eric Zhang, Andrew Ansah, Kevin Zhu, Ashwinee Panda, Aishwarya Balwani
Abstract: Recent work has discovered that large language models can develop broadly misaligned behaviors after being fine-tuned on narrowly harmful datasets, a phenomenon known as emergent misalignment (EM). However, the fundamental mechanisms enabling such harmful generalization across disparate domains remain poorly understood. In this work, we adopt a geometric perspective to study EM and demonstrate that it exhibits a fundamental cross-task linear structure in how harmful behavior is encoded across different datasets. Specifically, we find a strong convergence in EM parameters across tasks, with the fine-tuned weight updates showing relatively high cosine similarities, as well as shared lower-dimensional subspaces as measured by their principal angles and projection overlaps. Furthermore, we also show functional equivalence via linear mode connectivity, wherein interpolated models across narrow misalignment tasks maintain coherent, broadly misaligned behavior. Our results indicate that EM arises from different narrow tasks discovering the same set of shared parameter directions, suggesting that harmful behaviors may be organized into specific, predictable regions of the weight landscape. By revealing this fundamental connection between parametric geometry and behavioral outcomes, we hope our work catalyzes further research on parameter space interpretability and weight-based interventions.
Authors: Rathin Chandra Shit
Abstract: Catastrophic forgetting is one of the fundamental issues of continual learning because neural networks forget the tasks learned previously when trained on new tasks. The proposed framework is a new path-coordinated framework of continual learning that unites the Neural Tangent Kernel (NTK) theory of principled plasticity bounds, statistical validation by Wilson confidence intervals, and evaluation of path quality by the use of multiple metrics. Experimental evaluation shows an average accuracy of 66.7% at the cost of 23.4% catastrophic forgetting on Split-CIFAR10, a huge improvement over the baseline and competitive performance achieved, which is very close to state-of-the-art results. Further, it is found out that NTK condition numbers are predictive indicators of learning capacity limits, showing the existence of a critical threshold at condition number $>10^{11}$. It is interesting to note that the proposed strategy shows a tendency of lowering forgetting as the sequence of tasks progresses (27% to 18%), which is a system stabilization. The framework validates 80% of discovered paths with a rigorous statistical guarantee and maintains 90-97% retention on intermediate tasks. The core capacity limits of the continual learning environment are determined in the analysis, and actionable insights to enhance the adaptive regularization are offered.
Authors: Duc A. Tran, Dung Truong, Duy Le
Abstract: Submodular maximization is an optimization problem benefiting many machine learning applications, where we seek a small subset best representing an extremely large dataset. We focus on the federated setting where the data are locally owned by decentralized clients who have their own definitions for the quality of representability. This setting requires repetitive aggregation of local information computed by the clients. While the main motivation is to respect the privacy and autonomy of the clients, the federated setting is vulnerable to client misbehaviors: malicious clients might share fake information. An analogy is backdoor attack in conventional federated learning, but our challenge differs freshly due to the unique characteristics of submodular maximization. We propose RobustFSM, a federated submodular maximization solution that is robust to various practical client attacks. Its performance is substantiated with an empirical evaluation study using real-world datasets. Numerical results show that the solution quality of RobustFSM substantially exceeds that of the conventional federated algorithm when attacks are severe. The degree of this improvement depends on the dataset and attack scenarios, which can be as high as 200%
Authors: Elham Gholamzadeh, Kajal Singla, Nico Scherf
Abstract: Predicting interspecies interactions is a key challenge in microbial ecology, as these interactions are critical to determining the structure and activity of microbial communities. In this work, we used data on monoculture growth capabilities, interactions with other species, and phylogeny to predict a negative or positive effect of interactions. More precisely, we used one of the largest available pairwise interaction datasets to train our models, comprising over 7,500 interactions be- tween 20 species from two taxonomic groups co-cultured under 40 distinct carbon conditions, with a primary focus on the work of Nestor et al.[28 ]. In this work, we propose Graph Neural Networks (GNNs) as a powerful classifier to predict the direction of the effect. We construct edge-graphs of pairwise microbial interactions in order to leverage shared information across individual co-culture experiments, and use GNNs to predict modes of interaction. Our model can not only predict binary interactions (positive/negative) but also classify more complex interaction types such as mutualism, competition, and parasitism. Our initial results were encouraging, achieving an F1-score of 80.44%. This significantly outperforms comparable methods in the literature, including conventional Extreme Gradient Boosting (XGBoost) models, which reported an F1-score of 72.76%.
Authors: M. Z. Haider, M. U. Ghouri, Tayyaba Noreen, M. Salman
Abstract: Rare events such as financial crashes, climate extremes, and biological anomalies are notoriously difficult to model due to their scarcity and heavy-tailed distributions. Classical deep generative models often struggle to capture these rare occurrences, either collapsing low-probability modes or producing poorly calibrated uncertainty estimates. In this work, we propose the Quantum-Enhanced Generative Model (QEGM), a hybrid classical-quantum framework that integrates deep latent-variable models with variational quantum circuits. The framework introduces two key innovations: (1) a hybrid loss function that jointly optimizes reconstruction fidelity and tail-aware likelihood, and (2) quantum randomness-driven noise injection to enhance sample diversity and mitigate mode collapse. Training proceeds via a hybrid loop where classical parameters are updated through backpropagation while quantum parameters are optimized using parameter-shift gradients. We evaluate QEGM on synthetic Gaussian mixtures and real-world datasets spanning finance, climate, and protein structure. Results demonstrate that QEGM reduces tail KL divergence by up to 50 percent compared to state-of-the-art baselines (GAN, VAE, Diffusion), while improving rare-event recall and coverage calibration. These findings highlight the potential of QEGM as a principled approach for rare-event prediction, offering robustness beyond what is achievable with purely classical methods.
Authors: Bozhi You, Irene Wang, Zelal Su Mustafaoglu, Abhinav Jangda, Ang\'elica Moreira, Roshan Dathathri, Divya Mahajan, Keshav Pingali
Abstract: Bad charactors when submitting to arXiv: Attention is a fundamental building block of large language models (LLMs), so there have been many efforts to implement it efficiently. For example, FlashAttention leverages tiling and kernel fusion to optimize attention. Recently, a number of variants of attention have been introduced to enhance model quality or efficiency. Supporting them efficiently remains difficult since they usually require specialized kernels or hand-tuned implementations. FlexAttention recently addressed part of this gap by using static programming templates to support FlashAttention-like kernels for a subset of attention variants. In this paper, we introduce Flashlight, a compiler-native framework within the PyTorch ecosystem that automatically generates fused, FlashAttention-style kernels for arbitrary attention-based programs, without relying on static templates or predefined kernel specializations. Flashlight leverages PyTorch's compilation workflow to fuse and tile attention computations transparently, enabling efficient execution for diverse attention patterns. Not only does it support all variants expressible in the FlexAttention model but it also handles more general, data-dependent attention formulations that are beyond the capabilities of FlexAttention. Our results show that Flashlight produces kernels with competitive or superior performance to FlexAttention, while offering the flexibility of native PyTorch code, enabling developers to rapidly explore new attention models without sacrificing performance.
Authors: Vivswan Shah, Randy Cogill, Hanwei Yue, Gopinath Chennupati, Rinat Khaziev
Abstract: Fine-tuning LLMs for classification typically maps inputs directly to labels. We ask whether attaching brief explanations to each label during fine-tuning yields better models. We evaluate conversational response quality along three axes: naturalness, comprehensiveness, and on-topic adherence, each rated on 5-point scales. Using ensemble-generated data from multiple LLMs, we fine-tune a 7B-parameter model and test across six diverse conversational datasets. Across 18 dataset, task settings, label-plus-explanation training outperforms label-only baselines. A central and unexpected result concerns random tokens. We replace human-written explanations with text that is syntactically incoherent yet vocabulary-aligned with the originals (e.g., shuffled or bag-of-words variants). Despite lacking semantics, these pseudo-explanations still improve accuracy over label-only training and often narrow much of the gap to true explanations. The effect persists across datasets and training seeds, indicating that gains arise less from meaning than from structure: the extra token budget encourages richer intermediate computation and acts as a regularizer that reduces over-confident shortcuts. Internal analyses support this view: explanation-augmented models exhibit higher activation entropy in intermediate layers alongside sharper predictive mass at the output layer, consistent with increased deliberation before decision. Overall, explanation-augmented fine-tuning, whether with genuine rationales or carefully constructed random token sequences, improves accuracy and reliability for LLM classification while clarifying how token-level scaffolding shapes computation during inference.
Authors: Hamidreza Sadeghsalehi
Abstract: Objective gait analysis using wearable sensors and AI is critical for managing neurological and orthopedic conditions. However, models are vulnerable to hidden dataset biases, and task-specific sensor optimization remains a challenge. We propose a multi-stream attention-based deep learning framework that functions as both a sensor optimizer and an automated data auditor. Applied to the Voisard et al. (2025) multi-cohort gait dataset on four clinical tasks (PD, OA, CVA screening; PD vs CVA differential), the model's attention mechanism quantitatively discovered a severe dataset confound. For OA and CVA screening, tasks where bilateral assessment is clinically essential, the model assigned more than 70 percent attention to the Right Foot while statistically ignoring the Left Foot (less than 0.1 percent attention, 95 percent CI [0.0-0.1]). This was not a clinical finding but a direct reflection of a severe laterality bias (for example, 15 of 15 right-sided OA) in the public dataset. The primary contribution of this work is methodological, demonstrating that an interpretable framework can automatically audit dataset integrity. As a secondary finding, the model proposes novel, data-driven sensor synergies (for example, Head plus Foot for PD screening) as hypotheses for future optimized protocols.
Authors: Nimrod Megiddo, Segev Wasserkrug, Orit Davidovich, Shimrit Shtern
Abstract: The paper is about developing a solver for maximizing a real-valued function of binary variables. The solver relies on an algorithm that estimates the optimal objective-function value of instances from the underlying distribution of objectives and their respective sub-instances. The training of the estimator is based on an inequality that facilitates the use of the expected total deviation from optimality conditions as a loss function rather than the objective-function itself. Thus, it does not calculate values of policies, nor does it rely on solved instances.
Authors: Jucheng Shen, Yeonju Ro
Abstract: Masked diffusion language models (MDLMs) are becoming competitive with their autoregressive counterparts but typically decode with fixed steps and sequential unmasking. To accelerate decoding, recent work such as Fast-dLLM enables parallel decoding via a static global confidence threshold, yet we observe strong block- and step-wise confidence fluctuations and, within a dataset, near-identical confidence trajectories across inputs as measured by cosine similarity. Motivated by these observations, we introduce One-Shot Dynamic Thresholding (OSDT), which calibrates thresholds on a single sequence and applies them to subsequent inputs with negligible overhead. On GPQA, GSM8K, and HumanEval, OSDT attains superior accuracy-throughput trade-offs (+24% tokens/s on GSM8K at the best accuracy, +45% on GPQA with comparable accuracy, and +50% on HumanEval with a modest accuracy gap). Beyond these results, our findings suggest broader opportunities to leverage reusable task-level confidence signatures for more general-purpose algorithmic and systems innovations in diffusion decoding.
Authors: S\'ekou-Oumar Kaba, Kusha Sareen, Daniel Levy, Siamak Ravanbakhsh
Abstract: Effectively leveraging prior knowledge of a system's physics is crucial for applications of machine learning to scientific domains. Previous approaches mostly focused on incorporating physical insights at the architectural level. In this paper, we propose a framework to leverage physical information directly into the loss function for prediction and generative modeling tasks on systems like molecules and spins. We derive energy loss functions assuming that each data sample is in thermal equilibrium with respect to an approximate energy landscape. By using the reverse KL divergence with a Boltzmann distribution around the data, we obtain the loss as an energy difference between the data and the model predictions. This perspective also recasts traditional objectives like MSE as energy-based, but with a physically meaningless energy. In contrast, our formulation yields physically grounded loss functions with gradients that better align with valid configurations, while being architecture-agnostic and computationally efficient. The energy loss functions also inherently respect physical symmetries. We demonstrate our approach on molecular generation and spin ground-state prediction and report significant improvements over baselines.
Authors: Stefan F. Schouten, Peter Bloem
Abstract: Contrast-Consistent Search (CCS) is an unsupervised probing method able to test whether large language models represent binary features, such as sentence truth, in their internal activations. While CCS has shown promise, its two-term objective has been only partially understood. In this work, we revisit CCS with the aim of clarifying its mechanisms and extending its applicability. We argue that what should be optimized for, is relative contrast consistency. Building on this insight, we reformulate CCS as an eigenproblem, yielding closed-form solutions with interpretable eigenvalues and natural extensions to multiple variables. We evaluate these approaches across a range of datasets, finding that they recover similar performance to CCS, while avoiding problems around sensitivity to random initialization. Our results suggest that relativizing contrast consistency not only improves our understanding of CCS but also opens pathways for broader probing and mechanistic interpretability methods.
Authors: Lancelot Da Costa, Sanjeev Namjoshi, Mohammed Abbas Ansari, Bernhard Sch\"olkopf
Abstract: The field of world modeling is fragmented, with researchers developing bespoke architectures that rarely build upon each other. We propose a framework that specifies the natural building blocks for structured world models based on the fundamental stochastic processes that any world model must capture: discrete processes (logic, symbols) and continuous processes (physics, dynamics); the world model is then defined by the hierarchical composition of these building blocks. We examine Hidden Markov Models (HMMs) and switching linear dynamical systems (sLDS) as natural building blocks for discrete and continuous modeling--which become partially-observable Markov decision processes (POMDPs) and controlled sLDS when augmented with actions. This modular approach supports both passive modeling (generation, forecasting) and active control (planning, decision-making) within the same architecture. We avoid the combinatorial explosion of traditional structure learning by largely fixing the causal architecture and searching over only four depth parameters. We review practical expressiveness through multimodal generative modeling (passive) and planning from pixels (active), with performance competitive to neural approaches while maintaining interpretability. The core outstanding challenge is scalable joint structure-parameter learning; current methods finesse this by cleverly growing structure and parameters incrementally, but are limited in their scalability. If solved, these natural building blocks could provide foundational infrastructure for world modeling, analogous to how standardized layers enabled progress in deep learning.
Authors: Kishansingh Rajput, Malachi Schram, Brian Sammuli, Sen Lin
Abstract: Machine Learning (ML) is poised to play a pivotal role in the development and operation of next-generation fusion devices. Fusion data shows non-stationary behavior with distribution drifts, resulted by both experimental evolution and machine wear-and-tear. ML models assume stationary distribution and fail to maintain performance when encountered with such non-stationary data streams. Online learning techniques have been leveraged in other domains, however it has been largely unexplored for fusion applications. In this paper, we present an application of online learning to continuously adapt to drifting data stream for prediction of Toroidal Field (TF) coils deflection at the DIII-D fusion facility. The results demonstrate that online learning is critical to maintain ML model performance and reduces error by 80% compared to a static model. Moreover, traditional online learning can suffer from short-term performance degradation as ground truth is not available before making the predictions. As such, we propose an uncertainty guided online ensemble method to further improve the performance. The Deep Gaussian Process Approximation (DGPA) technique is leveraged for calibrated uncertainty estimation and the uncertainty values are then used to guide a meta-algorithm that produces predictions based on an ensemble of learners trained on different horizon of historical data. The DGPA also provides uncertainty estimation along with the predictions for decision makers. The online ensemble and the proposed uncertainty guided online ensemble reduces predictions error by about 6%, and 10% respectively over standard single model based online learning.
Authors: Rodrigo Mendoza-Smith
Abstract: Shapley data valuation provides a principled, axiomatic framework for assigning importance to individual datapoints, and has gained traction in dataset curation, pruning, and pricing. However, it is a combinatorial measure that requires evaluating marginal utility across all subsets of the data, making it computationally infeasible at scale. We propose a geometric alternative based on statistical leverage scores, which quantify each datapoint's structural influence in the representation space by measuring how much it extends the span of the dataset and contributes to the effective dimensionality of the training problem. We show that our scores satisfy the dummy, efficiency, and symmetry axioms of Shapley valuation and that extending them to \emph{ridge leverage scores} yields strictly positive marginal gains that connect naturally to classical A- and D-optimal design criteria. We further show that training on a leverage-sampled subset produces a model whose parameters and predictive risk are within $O(\varepsilon)$ of the full-data optimum, thereby providing a rigorous link between data valuation and downstream decision quality. Finally, we conduct an active learning experiment in which we empirically demonstrate that ridge-leverage sampling outperforms standard baselines without requiring access gradients or backward passes.
Authors: Arjun Rao, Marc Ru{\ss}wurm, Konstantin Klemmer, Esther Rolf
Abstract: Within the context of representation learning for Earth observation, geographic Implicit Neural Representations (INRs) embed low-dimensional location inputs (longitude, latitude) into high-dimensional embeddings, through models trained on geo-referenced satellite, image or text data. Despite the common aim of geographic INRs to distill Earth's data into compact, learning-friendly representations, we lack an understanding of how much information is contained in these Earth representations, and where that information is concentrated. The intrinsic dimension of a dataset measures the number of degrees of freedom required to capture its local variability, regardless of the ambient high-dimensional space in which it is embedded. This work provides the first study of the intrinsic dimensionality of geographic INRs. Analyzing INRs with ambient dimension between 256 and 512, we find that their intrinsic dimensions fall roughly between 2 and 10 and are sensitive to changing spatial resolution and input modalities during INR pre-training. Furthermore, we show that the intrinsic dimension of a geographic INR correlates with downstream task performance and can capture spatial artifacts, facilitating model evaluation and diagnostics. More broadly, our work offers an architecture-agnostic, label-free metric of information content that can enable unsupervised evaluation, model selection, and pre-training design across INRs.
Authors: Xinyuan Song, Jiaye Teng, Ziye Ma
Abstract: In this paper we study how the choice of loss functions of non-convex optimization problems affects their robustness and optimization landscape, through the study of noisy matrix sensing. In traditional regression tasks, mean squared error (MSE) loss is a common choice, but it can be unreliable for non-Gaussian or heavy-tailed noise. To address this issue, we adopt a robust loss based on nonparametric regression, which uses a kernel-based estimate of the residual density and maximizes the estimated log-likelihood. This robust formulation coincides with the MSE loss under Gaussian errors but remains stable under more general settings. We further examine how this robust loss reshapes the optimization landscape by analyzing the upper-bound of restricted isometry property (RIP) constants for spurious local minima to disappear. Through theoretical and empirical analysis, we show that this new loss excels at handling large noise and remains robust across diverse noise distributions. This work offers initial insights into enhancing the robustness of machine learning tasks through simply changing the loss, guided by an intuitive and broadly applicable analytical framework.
Authors: Xuheng Li, Quanquan Gu
Abstract: Variance-dependent regret bounds have received increasing attention in recent studies on contextual bandits. However, most of these studies are focused on upper confidence bound (UCB)-based bandit algorithms, while sampling based bandit algorithms such as Thompson sampling are still understudied. The only exception is the LinVDTS algorithm (Xu et al., 2023), which is limited to linear reward function and its regret bound is not optimal with respect to the model dimension. In this paper, we present FGTSVA, a variance-aware Thompson Sampling algorithm for contextual bandits with general reward function with optimal regret bound. At the core of our analysis is an extension of the decoupling coefficient, a technique commonly used in the analysis of Feel-good Thompson sampling (FGTS) that reflects the complexity of the model space. With the new decoupling coefficient denoted by $\mathrm{dc}$, FGTS-VA achieves the regret of $\tilde{O}(\sqrt{\mathrm{dc}\cdot\log|\mathcal{F}|\sum_{t=1}^T\sigma_t^2}+\mathrm{dc})$, where $|\mathcal{F}|$ is the size of the model space, $T$ is the total number of rounds, and $\sigma_t^2$ is the subgaussian norm of the noise (e.g., variance when the noise is Gaussian) at round $t$. In the setting of contextual linear bandits, the regret bound of FGTSVA matches that of UCB-based algorithms using weighted linear regression (Zhou and Gu, 2022).
Authors: Yasaman Torabi, Shahram Shirani, James P. Reilly
Abstract: Early identification of abnormal physiological patterns is essential for the timely detection of cardiac disease. This work introduces a hybrid quantum-classical convolutional neural network (QCNN) designed to classify S3 and murmur abnormalities in heart sound signals. The approach transforms one-dimensional phonocardiogram (PCG) signals into compact two-dimensional images through a combination of wavelet feature extraction and adaptive threshold compression methods. We compress the cardiac-sound patterns into an 8-pixel image so that only 8 qubits are needed for the quantum stage. Preliminary results on the HLS-CMDS dataset demonstrate 93.33% classification accuracy on the test set and 97.14% on the train set, suggesting that quantum models can efficiently capture temporal-spectral correlations in biomedical signals. To our knowledge, this is the first application of a QCNN algorithm for bioacoustic signal processing. The proposed method represents an early step toward quantum-enhanced diagnostic systems for resource-constrained healthcare environments.
Authors: Yi Luo, Haochen Zhao, Xiao Liang, Yiwei Liu, Yuye Zhang, Xinyu Li, Jianxin Wang
Abstract: Drug synergy prediction is a critical task in the development of effective combination therapies for complex diseases, including cancer. Although existing methods have shown promising results, they often operate as black-box predictors that rely predominantly on statistical correlations between drug characteristics and results. To address this limitation, we propose CausalDDS, a novel framework that disentangles drug molecules into causal and spurious substructures, utilizing the causal substructure representations for predicting drug synergy. By focusing on causal sub-structures, CausalDDS effectively mitigates the impact of redundant features introduced by spurious substructures, enhancing the accuracy and interpretability of the model. In addition, CausalDDS employs a conditional intervention mechanism, where interventions are conditioned on paired molecular structures, and introduces a novel optimization objective guided by the principles of sufficiency and independence. Extensive experiments demonstrate that our method outperforms baseline models, particularly in cold start and out-of-distribution settings. Besides, CausalDDS effectively identifies key substructures underlying drug synergy, providing clear insights into how drug combinations work at the molecular level. These results underscore the potential of CausalDDS as a practical tool for predicting drug synergy and facilitating drug discovery.
Authors: Abdullah Almansour, Ozan Tonguz
Abstract: Machine Learning (ML) models are extensively used in various applications due to their significant advantages over traditional learning methods. However, the developed ML models often underperform when deployed in the real world due to the well-known distribution shift problem. This problem can lead to a catastrophic outcomes when these decision-making systems have to operate in high-risk applications. Many researchers have previously studied this problem in ML, known as distribution shift problem, using statistical techniques (such as Kullback-Leibler, Kolmogorov-Smirnov Test, Wasserstein distance, etc.) to quantify the distribution shift. In this letter, we show that using Characteristic Function (CF) as a frequency domain approach is a powerful alternative for measuring the distribution shift in high-dimensional space and for domain adaptation.
Authors: Bart{\l}omiej Ma{\l}kus, Szymon Bobek, Grzegorz J. Nalepa
Abstract: Time series data is one of the most popular data modalities in critical domains such as industry and medicine. The demand for algorithms that not only exhibit high accuracy but also offer interpretability is crucial in such fields, as decisions made there bear significant consequences. In this paper, we present ProtoTSNet, a novel approach to interpretable classification of multivariate time series data, through substantial enhancements to the ProtoPNet architecture. Our method is tailored to overcome the unique challenges of time series analysis, including capturing dynamic patterns and handling varying feature significance. Central to our innovation is a modified convolutional encoder utilizing group convolutions, pre-trainable as part of an autoencoder and designed to preserve and quantify feature importance. We evaluated our model on 30 multivariate time series datasets from the UEA archive, comparing our approach with existing explainable methods as well as non-explainable baselines. Through comprehensive evaluation and ablation studies, we demonstrate that our approach achieves the best performance among ante-hoc explainable methods while maintaining competitive performance with non-explainable and post-hoc explainable approaches, providing interpretable results accessible to domain experts.
Authors: Yuzhuang Pian, Taiyu Wang, Shiqi Zhang, Rui Xu, Yonghong Liu
Abstract: Accurate air quality forecasts are vital for public health alerts, exposure assessment, and emissions control. In practice, observational data are often missing in varying proportions and patterns due to collection and transmission issues. These incomplete spatiotemporal records impede reliable inference and risk assessment and can lead to overconfident extrapolation. To address these challenges, we propose an end to end framework, the channel gated learning unit based spatiotemporal bayesian neural field (CGLUBNF). It uses Fourier features with a graph attention encoder to capture multiscale spatial dependencies and seasonal temporal dynamics. A channel gated learning unit, equipped with learnable activations and gated residual connections, adaptively filters and amplifies informative features. Bayesian inference jointly optimizes predictive distributions and parameter uncertainty, producing point estimates and calibrated prediction intervals. We conduct a systematic evaluation on two real world datasets, covering four typical missing data patterns and comparing against five state of the art baselines. CGLUBNF achieves superior prediction accuracy and sharper confidence intervals. In addition, we further validate robustness across multiple prediction horizons and analysis the contribution of extraneous variables. This research lays a foundation for reliable deep learning based spatio-temporal forecasting with incomplete observations in emerging sensing paradigms, such as real world vehicle borne mobile monitoring.
Authors: Kevin Valencia, Thilina Balasooriya, Xihaier Luo, Shinjae Yoo, David Keetae Park
Abstract: Multimodal spatiotemporal learning on real-world experimental data is constrained by two challenges: within-modality measurements are sparse, irregular, and noisy (QA/QC artifacts) but cross-modally correlated; the set of available modalities varies across space and time, shrinking the usable record unless models can adapt to arbitrary subsets at train and test time. We propose OmniField, a continuity-aware framework that learns a continuous neural field conditioned on available modalities and iteratively fuses cross-modal context. A multimodal crosstalk block architecture paired with iterative cross-modal refinement aligns signals prior to the decoder, enabling unified reconstruction, interpolation, forecasting, and cross-modal prediction without gridding or surrogate preprocessing. Extensive evaluations show that OmniField consistently outperforms eight strong multimodal spatiotemporal baselines. Under heavy simulated sensor noise, performance remains close to clean-input levels, highlighting robustness to corrupted measurements.
Authors: Fan Feng, Phillip Lippe, Sara Magliacane
Abstract: Agents that understand objects and their interactions can learn policies that are more robust and transferable. However, most object-centric RL methods factor state by individual objects while leaving interactions implicit. We introduce the Factored Interactive Object-Centric World Model (FIOC-WM), a unified framework that learns structured representations of both objects and their interactions within a world model. FIOC-WM captures environment dynamics with disentangled and modular representations of object interactions, improving sample efficiency and generalization for policy learning. Concretely, FIOC-WM first learns object-centric latents and an interaction structure directly from pixels, leveraging pre-trained vision encoders. The learned world model then decomposes tasks into composable interaction primitives, and a hierarchical policy is trained on top: a high level selects the type and order of interactions, while a low level executes them. On simulated robotic and embodied-AI benchmarks, FIOC-WM improves policy-learning sample efficiency and generalization over world-model baselines, indicating that explicit, modular interaction learning is crucial for robust control.
Authors: Costin-Andrei Oncescu, Qingyang Wu, Wai Tong Chung, Robert Wu, Bryan Gopal, Junxiong Wang, Tri Dao, Ben Athiwaratkun
Abstract: An increasing number of LLMs employ Mixture-of-Experts (MoE) architectures where the feed-forward layer is replaced by a pool of experts and each token only activates a small subset of them. During autoregressive generation, these models often enter a memory-bound regime even for moderate batch sizes because the average expert load grows more slowly than in an equivalent dense feedforward layer. Consequently, MoE latency is governed by the number of activated experts. We introduce a framework for dynamically re-routing token-to-expert mapping to lower this number (and thus, the decode latency) while preserving a comparable quality. Our best results use a batch-aware routing that works by having tokens piggyback experts that have already been loaded into memory due to being crucial to other tokens within the same batch. Empirically, we evaluate our method on the Qwen3-30B and Qwen3-235B models with a batch size of $16$. Without any statistically significant loss in accuracy, our approach achieves latency reductions of $39\%$ and $15\%$ in the MoE layer decode latency, respectively.
Authors: Hikaru Homma, Jun Ohkubo
Abstract: Initialization of neural network parameters, such as weights and biases, has a crucial impact on learning performance; if chosen well, we can even avoid the need for additional training with backpropagation. For example, algorithms based on the ridgelet transform or the SWIM (sampling where it matters) concept have been proposed for initialization. On the other hand, it is well-known that neural networks tend to learn coarse information in the earlier layers. The feature is called spectral bias. In this work, we investigate the effects of utilizing information on the spectral bias in the initialization of neural networks. Hence, we propose a framework that adjusts the scale factors in the SWIM algorithm to capture low-frequency components in the early-stage hidden layers and to represent high-frequency components in the late-stage hidden layers. Numerical experiments on a one-dimensional regression task and the MNIST classification task demonstrate that the proposed method outperforms the conventional initialization algorithms. This work clarifies the importance of intrinsic spectral properties in learning neural networks, and the finding yields an effective parameter initialization strategy that enhances their training performance.
Authors: Ayoub Ghriss
Abstract: Probabilistic relaxations of graph cuts offer a differentiable alternative to spectral clustering, enabling end-to-end and online learning without eigendecompositions, yet prior work centered on RatioCut and lacked general guarantees and principled gradients. We present a unified probabilistic framework that covers a wide class of cuts, including Normalized Cut. Our framework provides tight analytic upper bounds on expected discrete cuts via integral representations and Gauss hypergeometric functions with closed-form forward and backward. Together, these results deliver a rigorous, numerically stable foundation for scalable, differentiable graph partitioning covering a wide range of clustering and contrastive learning objectives.
Authors: Yuheng Zhao, Yu-Hu Yan, Kfir Yehuda Levy, Peng Zhao
Abstract: Smoothness is known to be crucial for acceleration in offline optimization, and for gradient-variation regret minimization in online learning. Interestingly, these two problems are actually closely connected -- accelerated optimization can be understood through the lens of gradient-variation online learning. In this paper, we investigate online learning with H\"older smooth functions, a general class encompassing both smooth and non-smooth (Lipschitz) functions, and explore its implications for offline optimization. For (strongly) convex online functions, we design the corresponding gradient-variation online learning algorithm whose regret smoothly interpolates between the optimal guarantees in smooth and non-smooth regimes. Notably, our algorithms do not require prior knowledge of the H\"older smoothness parameter, exhibiting strong adaptivity over existing methods. Through online-to-batch conversion, this gradient-variation online adaptivity yields an optimal universal method for stochastic convex optimization under H\"older smoothness. However, achieving universality in offline strongly convex optimization is more challenging. We address this by integrating online adaptivity with a detection-based guess-and-check procedure, which, for the first time, yields a universal offline method that achieves accelerated convergence in the smooth regime while maintaining near-optimal convergence in the non-smooth one.
Authors: Ziyi Wang, Lijian Jiang
Abstract: Data assimilation (DA) has increasingly emerged as a critical tool for state estimation across a wide range of applications. It is signiffcantly challenging when the governing equations of the underlying dynamics are unknown. To this end, various machine learning approaches have been employed to construct a surrogate state transition model in a supervised learning framework, which relies on pre-computed training datasets. However, it is often infeasible to obtain noise-free ground-truth state sequences in practice. To address this challenge, we propose a novel method that integrates reinforcement learning with ensemble-based Bayesian ffltering methods, enabling the learning of surrogate state transition model for unknown dynamics directly from noisy observations, without using true state trajectories. Speciffcally, we treat the process for computing maximum likelihood estimation of surrogate model parameters as a sequential decision-making problem, which can be formulated as a discretetime Markov decision process (MDP). Under this formulation, learning the surrogate transition model is equivalent to ffnding an optimal policy of the MDP, which can be effectively addressed using reinforcement learning techniques. Once the model is trained offfine, state estimation can be performed in the online stage using ffltering methods based on the learned dynamics. The proposed framework accommodates a wide range of observation scenarios, including nonlinear and partially observed measurement models. A few numerical examples demonstrate that the proposed method achieves superior accuracy and robustness in high-dimensional settings.
Authors: Kuan-Cheng Chen, Samuel Yen-Chi Chen, Chen-Yu Liu, Kin K. Leung
Abstract: The rapid growth of industrial Internet of Things (IIoT) systems has created new challenges for anomaly detection in high-dimensional, multivariate time-series, where privacy, scalability, and communication efficiency are critical. Classical federated learning approaches mitigate privacy concerns by enabling decentralized training, but they often struggle with highly non-linear decision boundaries and imbalanced anomaly distributions. To address this gap, we propose a Federated Quantum Kernel Learning (FQKL) framework that integrates quantum feature maps with federated aggregation to enable distributed, privacy-preserving anomaly detection across heterogeneous IoT networks. In our design, quantum edge nodes locally compute compressed kernel statistics using parameterized quantum circuits and share only these summaries with a central server, which constructs a global Gram matrix and trains a decision function (e.g., Fed-QSVM). Experimental results on synthetic IIoT benchmarks demonstrate that FQKL achieves superior generalization in capturing complex temporal correlations compared to classical federated baselines, while significantly reducing communication overhead. This work highlights the promise of quantum kernels in federated settings, advancing the path toward scalable, robust, and quantum-enhanced intelligence for next-generation IoT infrastructures.
Authors: Fengjuan Wang, Zhiyi Su, Xingzhu Hu, Cheng Wang, Mou Sun
Abstract: Training large Mixture-of-Experts (MoE) models remains computationally prohibitive due to their extreme compute and memory demands. Although low-precision training promises to accelerate computation and reduce memory footprint, existing implementations still rely on BF16-dominated dataflows with frequent quantize-dequantize (Q/DQ) conversions. These redundant casts erode much of FP8's theoretical efficiency. However, naively removing these casts by keeping dataflows entirely in FP8 introduces double quantization error: tensors quantized along different dimensions accumulate inconsistent scaling factors, degrading numerical stability. We propose FP8-Flow-MoE, an FP8 training recipe featuring a quantization-consistent FP8-centric dataflow with a scaling-aware transpose and fused FP8 operators that streamline computation and eliminate explicit cast operations from 12 to 2. Evaluations on a 671B-parameter MoE model demonstrate up to 21\% higher throughput and 16.5 GB lower memory usage per GPU compared to BF16 and na\"ive FP8 baselines, while maintaining stable convergence. We provide a plug-and-play FP8 recipe compatible with TransformerEngine and Megatron-LM, which will be open-sourced soon.
Authors: Aman Sharma, Paras Chopra
Abstract: We revisit test-time scaling for language model reasoning and ask a fundamental question: at equal token budget and compute, is it better to run multiple independent chains in parallel, or to run fewer chains that iteratively refine through sequential steps? Through comprehensive evaluation across 5 state-of-the-art open source models and 3 challenging reasoning benchmarks, we find that sequential scaling where chains explicitly build upon previous attempts consistently outperforms the dominant parallel self-consistency paradigm in 95.6% of configurations with gains in accuracy upto 46.7%. Further, we introduce inverse-entropy weighted voting, a novel training-free method to further boost the accuracy of sequential scaling. By weighing answers in proportion to the inverse entropy of their reasoning chains, we increase our success rate over parallel majority and establish it as the optimal test-time scaling strategy. Our findings fundamentally challenge the parallel reasoning orthodoxy that has dominated test-time scaling since Wang et al.'s self-consistency decoding (Wang et al., 2022), positioning sequential refinement as the robust default for modern LLM reasoning and necessitating a paradigm shift in how we approach inference-time optimization.
Authors: Jueye Zhang, Chao Yang, Youfang Lai, Kai-Wen Li, Wenting Yan, Yunzhou Xia, Haimei Zhang, Jingjing Zhou, Gen Yang, Chen Lin, Tian Li, Yibao Zhang
Abstract: Head-and-neck cancer (HNC) planning is difficult because multiple critical organs-at-risk (OARs) are close to complex targets. Intensity-modulated carbon-ion therapy (IMCT) offers superior dose conformity and OAR sparing but remains slow due to relative biological effectiveness (RBE) modeling, leading to laborious, experience-based, and often suboptimal tuning of many treatment-planning parameters (TPPs). Recent deep learning (DL) methods are limited by data bias and plan feasibility, while reinforcement learning (RL) struggles to efficiently explore the exponentially large TPP search space. We propose a scalable multi-agent RL (MARL) framework for parallel tuning of 45 TPPs in IMCT. It uses a centralized-training decentralized-execution (CTDE) QMIX backbone with Double DQN, Dueling DQN, and recurrent encoding (DRQN) for stable learning in a high-dimensional, non-stationary environment. To enhance efficiency, we (1) use compact historical DVH vectors as state inputs, (2) apply a linear action-to-value transform mapping small discrete actions to uniform parameter adjustments, and (3) design an absolute, clinically informed piecewise reward aligned with plan scores. A synchronous multi-process worker system interfaces with the PHOENIX TPS for parallel optimization and accelerated data collection. On a head-and-neck dataset (10 training, 10 testing), the method tuned 45 parameters simultaneously and produced plans comparable to or better than expert manual ones (relative plan score: RL $85.93\pm7.85%$ vs Manual $85.02\pm6.92%$), with significant (p-value $<$ 0.05) improvements for five OARs. The framework efficiently explores high-dimensional TPP spaces and generates clinically competitive IMCT plans through direct TPS interaction, notably improving OAR sparing.
Authors: Tianle Pu, Zijie Geng, Haoyang Liu, Shixuan Liu, Jie Wang, Li Zeng, Chao Chen, Changjun Fan
Abstract: Mixed-Integer Linear Programming (MILP) is a fundamental and powerful framework for modeling complex optimization problems across diverse domains. Recently, learning-based methods have shown great promise in accelerating MILP solvers by predicting high-quality solutions. However, most existing approaches are developed and evaluated in single-domain settings, limiting their ability to generalize to unseen problem distributions. This limitation poses a major obstacle to building scalable and general-purpose learning-based solvers. To address this challenge, we introduce RoME, a domain-Robust Mixture-of-Experts framework for predicting MILP solutions across domains. RoME dynamically routes problem instances to specialized experts based on learned task embeddings. The model is trained using a two-level distributionally robust optimization strategy: inter-domain to mitigate global shifts across domains, and intra-domain to enhance local robustness by introducing perturbations on task embeddings. We reveal that cross-domain training not only enhances the model's generalization capability to unseen domains but also improves performance within each individual domain by encouraging the model to capture more general intrinsic combinatorial patterns. Specifically, a single RoME model trained on three domains achieves an average improvement of 67.7% then evaluated on five diverse domains. We further test the pretrained model on MIPLIB in a zero-shot setting, demonstrating its ability to deliver measurable performance gains on challenging real-world instances where existing learning-based approaches often struggle to generalize.
Authors: Fidan Karimova, Tong Chen, Yu Yang, Shazia Sadiq
Abstract: Predicting crimes in urban environments is crucial for public safety, yet existing prediction methods often struggle to align the knowledge across diverse cities that vary dramatically in data availability of specific crime types. We propose HYpernetwork-enhanced Spatial Temporal Learning (HYSTL), a framework that can effectively train a unified, stronger crime predictor without assuming identical crime types in different cities' records. In HYSTL, instead of parameterising a dedicated predictor per crime type, a hypernetwork is designed to dynamically generate parameters for the prediction function conditioned on the crime type of interest. To bridge the semantic gap between different crime types, a structured crime knowledge graph is built, where the learned representations of crimes are used as the input to the hypernetwork to facilitate parameter generation. As such, when making predictions for each crime type, the predictor is additionally guided by its intricate association with other relevant crime types. Extensive experiments are performed on two cities with non-overlapping crime types, and the results demonstrate HYSTL outperforms state-of-the-art baselines.
Authors: David Nabergoj, Erik \v{S}trumbelj
Abstract: Preconditioning is a key component of MCMC algorithms that improves sampling efficiency by facilitating exploration of geometrically complex target distributions through an invertible map. While linear preconditioners are often sufficient for moderately complex target distributions, recent work has explored nonlinear preconditioning with invertible neural networks as components of normalizing flows (NFs). However, empirical and theoretical studies show that overparameterized NF preconditioners can degrade sampling efficiency and fit quality. Moreover, existing NF-based approaches do not adapt their architectures to the target distribution. Related work outside of MCMC similarly finds that suitably parameterized NFs can achieve comparable or superior performance with substantially less training time or data. We propose a factorized preconditioning architecture that reduces NF complexity by combining a linear component with a conditional NF, improving adaptability to target geometry. The linear preconditioner is applied to dimensions that are approximately Gaussian, as estimated from warmup samples, while the conditional NF models more complex dimensions. Our method yields significantly better tail samples on two complex synthetic distributions and consistently better performance on a sparse logistic regression posterior across varying likelihood and prior strengths. It also achieves higher effective sample sizes on hierarchical Bayesian model posteriors with weak likelihoods and strong funnel geometries. This approach is particularly relevant for hierarchical Bayesian model analyses with limited data and could inform current theoretical and software strides in neural MCMC design.
Authors: Zhuodi Cai, Ziyu Xu, Juan Pampin
Abstract: We introduce a lightweight, real-time motion recognition system that enables synergic human-machine performance through wearable IMU sensor data, MiniRocket time-series classification, and responsive multimedia control. By mapping dancer-specific movement to sound through somatic memory and association, we propose an alternative approach to human-machine collaboration, one that preserves the expressive depth of the performing body while leveraging machine learning for attentive observation and responsiveness. We demonstrate that this human-centered design reliably supports high accuracy classification (<50 ms latency), offering a replicable framework to integrate dance-literate machines into creative, educational, and live performance contexts.
Authors: Qingyun Sun, Jiayi Luo, Haonan Yuan, Xingcheng Fu, Hao Peng, Jianxin Li, Philip S. Yu
Abstract: Graph neural networks have shown remarkable success in exploiting the spatial and temporal patterns on dynamic graphs. However, existing GNNs exhibit poor generalization ability under distribution shifts, which is inevitable in dynamic scenarios. As dynamic graph generation progresses amid evolving latent non-stationary environments, it is imperative to explore their effects on out-of-distribution (OOD) generalization. This paper proposes a novel Evolving Graph Learning framework for OOD generalization (EvoOOD) by environment-aware invariant pattern recognition. Specifically, we first design an environment sequential variational auto-encoder to model environment evolution and infer the underlying environment distribution. Then, we introduce a mechanism for environment-aware invariant pattern recognition, tailored to address environmental diversification through inferred distributions. Finally, we conduct fine-grained causal interventions on individual nodes using a mixture of instantiated environment samples. This approach helps to distinguish spatio-temporal invariant patterns for OOD prediction, especially in non-stationary environments. Experimental results demonstrate the superiority of EvoGOOD on both real-world and synthetic dynamic datasets under distribution shifts. To the best of our knowledge, it is the first attempt to study the dynamic graph OOD generalization problem from the environment evolution perspective.
Authors: Rohan Wandre, Yash Gajewar, Namrata Patel, Vivek Dhalkari
Abstract: Retrieval-Augmented Generation (RAG) has emerged as the dominant paradigm for grounding large language model outputs in verifiable evidence. However, as modern AI agents transition from static knowledge bases to continuous multimodal streams encompassing text, images, video, and audio, two critical challenges arise: maintaining index freshness without prohibitive re-indexing costs, and preserving cross-modal semantic consistency across heterogeneous embedding spaces. We present LUMA-RAG, a lifelong multimodal agent architecture featuring three key innovations: (i) a streaming, multi-tier memory system that dynamically spills embeddings from a hot HNSW tier to a compressed IVFPQ tier under strict memory budgets; (ii) a streaming CLAP->CLIP alignment bridge that maintains cross-modal consistency through incremental orthogonal Procrustes updates; and (iii) stability-aware retrieval telemetry providing Safe@k guarantees by jointly bounding alignment drift and quantization error. Experiments demonstrate robust text-to-image retrieval (Recall@10 = 0.94), graceful performance degradation under product quantization offloading, and provably stable audio-to-image rankings (Safe@1 = 1.0), establishing LUMA-RAG as a practical framework for production multimodal RAG systems.
Authors: Rohith Shinoj Kumar, Rushdeep Dinda, Aditya Tyagi, Annappa B., Naveen Kumar M. R
Abstract: Early detection of heart arrhythmia can prevent severe future complications in cardiac patients. While manual diagnosis still remains the clinical standard, it relies heavily on visual interpretation and is inherently subjective. In recent years, deep learning has emerged as a powerful tool to automate arrhythmia detection, offering improved accuracy, consistency, and efficiency. Several variants of convolutional and recurrent neural network architectures have been widely explored to capture spatial and temporal patterns in physiological signals. However, despite these advancements, current models often struggle to generalize well in real-world scenarios, especially when dealing with small or noisy datasets, which are common challenges in biomedical applications. In this paper, a novel CNN-H-Infinity-LSTM architecture is proposed to identify arrhythmic heart signals from heart sound recordings. This architecture introduces trainable parameters inspired by the H-Infinity filter from control theory, enhancing robustness and generalization. Extensive experimentation on the PhysioNet CinC Challenge 2016 dataset, a public benchmark of heart audio recordings, demonstrates that the proposed model achieves stable convergence and outperforms existing benchmarks, with a test accuracy of 99.42% and an F1 score of 98.85%.
Authors: Gennaro Guidone, Luca Monegaglia, Elia Raimondi, Han Wang, Mattia Bianchi, Florian D\"orfler
Abstract: We present a novel decentralized algorithm for coverage control in unknown spatial environments modeled by Gaussian Processes (GPs). To trade-off between exploration and exploitation, each agent autonomously determines its trajectory by minimizing a local cost function. Inspired by the GP-UCB (Upper Confidence Bound for GPs) acquisition function, the proposed cost combines the expected locational cost with a variance-based exploration term, guiding agents toward regions that are both high in predicted density and model uncertainty. Compared to previous work, our algorithm operates in a fully decentralized fashion, relying only on local observations and communication with neighboring agents. In particular, agents periodically update their inducing points using a greedy selection strategy, enabling scalable online GP updates. We demonstrate the effectiveness of our algorithm in simulation.
Authors: Virgile Dine, Teddy Furon, Charly Faure
Abstract: We formulate the machine unlearning problem as a general constrained optimization problem. It unifies the first-order methods from the approximate machine unlearning literature. This paper then introduces the concept of feasible updates as the model's parameter update directions that help with unlearning while not degrading the utility of the initial model. Our design of feasible updates is based on masking, \ie\ a careful selection of the model's parameters worth updating. It also takes into account the estimation noise of the gradients when processing each batch of data to offer a statistical guarantee to derive locally feasible updates. The technique can be plugged in, as an add-on, to any first-order approximate unlearning methods. Experiments with computer vision classifiers validate this approach.
Authors: Thomas Sanchez, Pedro M. Gordaliza, Meritxell Bach Cuadra
Abstract: Machine learning methods are increasingly applied in medical imaging, yet many reported improvements lack statistical robustness: recent works have highlighted that small but significant performance gains are highly likely to be false positives. However, these analyses do not take \emph{underspecification} into account -- the fact that models achieving similar validation scores may behave differently on unseen data due to random initialization or training dynamics. Here, we extend a recent statistical framework modeling false outperformance claims to include underspecification as an additional variance component. Our simulations demonstrate that even modest seed variability ($\sim1\%$) substantially increases the evidence required to support superiority claims. Our findings underscore the need for explicit modeling of training variance when validating medical imaging systems.
Authors: Xuan-Truong Quan, Xuan-Son Quan, Duc Do Minh, Vinh Nguyen Van
Abstract: Knowledge graph embedding (KGE) has become a fundamental technique for representation learning on multi-relational data. Many seminal models, such as TransE, operate in an unbounded Euclidean space, which presents inherent limitations in modeling complex relations and can lead to inefficient training. In this paper, we propose Spherical Knowledge Graph Embedding (SKGE), a model that challenges this paradigm by constraining entity representations to a compact manifold: a hypersphere. SKGE employs a learnable, non-linear Spherization Layer to map entities onto the sphere and interprets relations as a hybrid translate-then-project transformation. Through extensive experiments on three benchmark datasets, FB15k-237, CoDEx-S, and CoDEx-M, we demonstrate that SKGE consistently and significantly outperforms its strong Euclidean counterpart, TransE, particularly on large-scale benchmarks such as FB15k-237 and CoDEx-M, demonstrating the efficacy of the spherical geometric prior. We provide an in-depth analysis to reveal the sources of this advantage, showing that this geometric constraint acts as a powerful regularizer, leading to comprehensive performance gains across all relation types. More fundamentally, we prove that the spherical geometry creates an "inherently hard negative sampling" environment, naturally eliminating trivial negatives and forcing the model to learn more robust and semantically coherent representations. Our findings compellingly demonstrate that the choice of manifold is not merely an implementation detail but a fundamental design principle, advocating for geometric priors as a cornerstone for designing the next generation of powerful and stable KGE models.
Authors: Mohammad Sadegh Eshaghi, Cosmin Anitescu, Navid Valizadeh, Yizheng Wang, Xiaoying Zhuang, Timon Rabczuk
Abstract: Partial differential equations (PDEs) underpin quantitative descriptions across the physical sciences and engineering, yet high-fidelity simulation remains a major computational bottleneck for many-query, real-time, and design tasks. Data-driven surrogates can be strikingly fast but are often unreliable when applied outside their training distribution. Here we introduce Neural Operator Warm Starts (NOWS), a hybrid strategy that harnesses learned solution operators to accelerate classical iterative solvers by producing high-quality initial guesses for Krylov methods such as conjugate gradient and GMRES. NOWS leaves existing discretizations and solver infrastructures intact, integrating seamlessly with finite-difference, finite-element, isogeometric analysis, finite volume method, etc. Across our benchmarks, the learned initialization consistently reduces iteration counts and end-to-end runtime, resulting in a reduction of the computational time of up to 90 %, while preserving the stability and convergence guarantees of the underlying numerical algorithms. By combining the rapid inference of neural operators with the rigor of traditional solvers, NOWS provides a practical and trustworthy approach to accelerate high-fidelity PDE simulations.
Authors: Rajan Das Gupta, Md Kishor Morol, Nafiz Fahad, Md Tanzib Hosain, Sumaya Binte Zilani Choya, Md Jakir Hossen
Abstract: As the global burden of Alzheimer's disease (AD) continues to grow, early and accurate detection has become increasingly critical, especially in regions with limited access to advanced diagnostic tools. We propose BRAINS (Biomedical Retrieval-Augmented Intelligence for Neurodegeneration Screening) to address this challenge. This novel system harnesses the powerful reasoning capabilities of Large Language Models (LLMs) for Alzheimer's detection and monitoring. BRAINS features a dual-module architecture: a cognitive diagnostic module and a case-retrieval module. The Diagnostic Module utilizes LLMs fine-tuned on cognitive and neuroimaging datasets -- including MMSE, CDR scores, and brain volume metrics -- to perform structured assessments of Alzheimer's risk. Meanwhile, the Case Retrieval Module encodes patient profiles into latent representations and retrieves similar cases from a curated knowledge base. These auxiliary cases are fused with the input profile via a Case Fusion Layer to enhance contextual understanding. The combined representation is then processed with clinical prompts for inference. Evaluations on real-world datasets demonstrate BRAINS effectiveness in classifying disease severity and identifying early signs of cognitive decline. This system not only shows strong potential as an assistive tool for scalable, explainable, and early-stage Alzheimer's disease detection, but also offers hope for future applications in the field.
Authors: Ronald Katende
Abstract: We propose a unified information-geometric framework that formalizes understanding in learning as a trade-off between informativeness and geometric simplicity. An encoder phi is evaluated by U(phi) = I(phi(X); Y) - beta * C(phi), where C(phi) penalizes curvature and intrinsic dimensionality, enforcing smooth, low-complexity manifolds. Under mild manifold and regularity assumptions, we derive non-asymptotic bounds showing that generalization error scales with intrinsic dimension while curvature controls approximation stability, directly linking geometry to sample efficiency. To operationalize this theory, we introduce the Variational Geometric Information Bottleneck (V-GIB), a variational estimator that unifies mutual-information compression and curvature regularization through tractable geometric proxies such as the Hutchinson trace, Jacobian norms, and local PCA. Experiments across synthetic manifolds, few-shot settings, and real-world datasets (Fashion-MNIST, CIFAR-10) reveal a robust information-geometry Pareto frontier, stable estimators, and substantial gains in interpretive efficiency. Fractional-data experiments on CIFAR-10 confirm that curvature-aware encoders maintain predictive power under data scarcity, validating the predicted efficiency-curvature law. Overall, V-GIB provides a principled and measurable route to representations that are geometrically coherent, data-efficient, and aligned with human-understandable structure.
Authors: Changhao Miao, Yuntian Zhang, Tongyu Wu, Fang Deng, Chen Chen
Abstract: The capacitated location-routing problems (CLRPs) are classical problems in combinatorial optimization, which require simultaneously making location and routing decisions. In CLRPs, the complex constraints and the intricate relationships between various decisions make the problem challenging to solve. With the emergence of deep reinforcement learning (DRL), it has been extensively applied to address the vehicle routing problem and its variants, while the research related to CLRPs still needs to be explored. In this paper, we propose the DRL with heterogeneous query (DRLHQ) to solve CLRP and open CLRP (OCLRP), respectively. We are the first to propose an end-to-end learning approach for CLRPs, following the encoder-decoder structure. In particular, we reformulate the CLRPs as a markov decision process tailored to various decisions, a general modeling framework that can be adapted to other DRL-based methods. To better handle the interdependency across location and routing decisions, we also introduce a novel heterogeneous querying attention mechanism designed to adapt dynamically to various decision-making stages. Experimental results on both synthetic and benchmark datasets demonstrate superior solution quality and better generalization performance of our proposed approach over representative traditional and DRL-based baselines in solving both CLRP and OCLRP.
Authors: Munib Mesinovic, Max Buhlan, Tingting Zhu
Abstract: Healthcare artificial intelligence systems routinely fail when deployed across institutions, with documented performance drops and perpetuation of discriminatory patterns embedded in historical data. This brittleness stems, in part, from learning statistical associations rather than causal mechanisms. Causal graph neural networks address this triple crisis of distribution shift, discrimination, and inscrutability by combining graph-based representations of biomedical data with causal inference principles to learn invariant mechanisms rather than spurious correlations. This Review examines methodological foundations spanning structural causal models, disentangled causal representation learning, and techniques for interventional prediction and counterfactual reasoning on graphs. We analyse applications demonstrating clinical value across psychiatric diagnosis through brain network analysis, cancer subtyping via multi-omics causal integration, continuous physiological monitoring with mechanistic interpretation, and drug recommendation correcting prescription bias. These advances establish foundations for patient-specific Causal Digital Twins, enabling in silico clinical experimentation, with integration of large language models for hypothesis generation and causal graph neural networks for mechanistic validation. Substantial barriers remain, including computational requirements precluding real-time deployment, validation challenges demanding multi-modal evidence triangulation beyond cross-validation, and risks of causal-washing where methods employ causal terminology without rigorous evidentiary support. We propose tiered frameworks distinguishing causally-inspired architectures from causally-validated discoveries and identify critical research priorities making causal rather than purely associational claims.
Authors: Hortence Nana, Andreas Athanasopoulos, Christos Dimitrakakis
Abstract: We study a many-to-one matching problem, such as the college admission problem, where each college can admit multiple students. Unlike classical models, colleges evaluate sets of students through non-linear utility functions that capture diversity between them. In this setting, we show that classical stable matchings may fail to exist. To address this, we propose alternative solution concepts based on Rawlsian fairness, aiming to maximize the minimum utility across colleges. We design both deterministic and stochastic algorithms that iteratively improve the outcome of the worst-off college, offering a practical approach to fair allocation when stability cannot be guaranteed.
Authors: Mathieu Chevalley, Arash Mehrjou, Patrick Schwab
Abstract: We investigate theoretical guarantees for the false-negative rate (FNR) -- the fraction of true causal edges whose orientation is not recovered, under single-variable random interventions and an $\epsilon$-interventional faithfulness assumption that accommodates latent confounding. For sparse Erd\H{o}s--R\'enyi directed acyclic graphs, where the edge probability scales as $p_e = \Theta(1/d)$, we show that the FNR concentrates around its mean at rate $O(\frac{\log d}{\sqrt d})$, implying that large deviations above the expected error become exponentially unlikely as dimensionality increases. This concentration ensures that derived upper bounds hold with high probability in large-scale settings. Extending the analysis to generalized Barab\'asi--Albert graphs reveals an even stronger phenomenon: when the degree exponent satisfies $\gamma > 3$, the deviation width scales as $O(d^{\beta - \frac{1}{2}})$ with $\beta = 1/(\gamma - 1) < \frac{1}{2}$, and hence vanishes in the limit. This demonstrates that realistic scale-free topologies intrinsically regularize causal discovery, reducing variability in orientation error. These finite-dimension results provide the first dimension-adaptive, faithfulness-robust guarantees for causal structure recovery, and challenge the intuition that high dimensionality and network heterogeneity necessarily hinder accurate discovery. Our simulation results corroborate these theoretical predictions, showing that the FNR indeed concentrates and often vanishes in practice as dimensionality grows.
Authors: Yixiu Mao, Yun Qu, Qi Wang, Xiangyang Ji
Abstract: Offline reinforcement learning (RL) suffers from extrapolation errors induced by out-of-distribution (OOD) actions. To address this, offline RL algorithms typically impose constraints on action selection, which can be systematically categorized into density, support, and sample constraints. However, we show that each category has inherent limitations: density and sample constraints tend to be overly conservative in many scenarios, while the support constraint, though least restrictive, faces challenges in accurately modeling the behavior policy. To overcome these limitations, we propose a new neighborhood constraint that restricts action selection in the Bellman target to the union of neighborhoods of dataset actions. Theoretically, the constraint not only bounds extrapolation errors and distribution shift under certain conditions, but also approximates the support constraint without requiring behavior policy modeling. Moreover, it retains substantial flexibility and enables pointwise conservatism by adapting the neighborhood radius for each data point. In practice, we employ data quality as the adaptation criterion and design an adaptive neighborhood constraint. Building on an efficient bilevel optimization framework, we develop a simple yet effective algorithm, Adaptive Neighborhood-constrained Q learning (ANQ), to perform Q learning with target actions satisfying this constraint. Empirically, ANQ achieves state-of-the-art performance on standard offline RL benchmarks and exhibits strong robustness in scenarios with noisy or limited data.
Authors: Lukas Fehring, Marcel Wever, Maximilian Splieth\"over, Leona Hennig, Henning Wachsmuth, Marius Lindauer
Abstract: Hyperparameter optimization (HPO), for example, based on Bayesian optimization (BO), supports users in designing models well-suited for a given dataset. HPO has proven its effectiveness on several applications, ranging from classical machine learning for tabular data to deep neural networks for computer vision and transformers for natural language processing. However, HPO still sometimes lacks acceptance by machine learning experts due to its black-box nature and limited user control. Addressing this, first approaches have been proposed to initialize BO methods with expert knowledge. However, these approaches do not allow for online steering during the optimization process. In this paper, we introduce a novel method that enables repeated interventions to steer BO via user input, specifying expert knowledge and user preferences at runtime of the HPO process in the form of prior distributions. To this end, we generalize an existing method, $\pi$BO, preserving theoretical guarantees. We also introduce a misleading prior detection scheme, which allows protection against harmful user inputs. In our experimental evaluation, we demonstrate that our method can effectively incorporate multiple priors, leveraging informative priors, whereas misleading priors are reliably rejected or overcome. Thereby, we achieve competitiveness to unperturbed BO.
Authors: Gilad Karpel, Ruida Zhou, Shoham Sabach, Mohammad Ghavamzadeh
Abstract: Proximal Policy Optimization (PPO) is widely regarded as one of the most successful deep reinforcement learning algorithms, known for its robustness and effectiveness across a range of problems. The PPO objective encourages the importance ratio between the current and behavior policies to move to the "right" direction -- starting from importance sampling ratios equal to 1, increasing the ratios for actions with positive advantages and decreasing those with negative advantages. A clipping function is introduced to prevent over-optimization when updating the importance ratio in these "right" direction regions. Many PPO variants have been proposed to extend its success, most of which modify the objective's behavior by altering the clipping in the "right" direction regions. However, due to randomness in the rollouts and stochasticity of the policy optimization, we observe that the ratios frequently move to the "wrong" direction during the PPO optimization. This is a key factor hindering the improvement of PPO, but it has been largely overlooked. To address this, we propose the Directional-Clamp PPO algorithm (DClamp-PPO), which further penalizes the actions going to the strict "wrong" direction regions, where the advantage is positive (negative) and importance ratio falls below (above) $1 - \beta$ ($1+\beta$), for a tunable parameter $\beta \in (0, 1)$. The penalty is by enforcing a steeper loss slope, i.e., a clamp, in those regions. We demonstrate that DClamp-PPO consistently outperforms PPO, as well as its variants, by focusing on modifying the objective's behavior in the "right" direction, across various MuJoCo environments, using different random seeds. The proposed method is shown, both theoretically and empirically, to better avoid "wrong" direction updates while keeping the importance ratio closer to 1.
Authors: Chitro Majumdar, Sergio Scandizzo, Ratanlal Mahanta, Avradip Mandal, Swarnendu Bhattacharjee
Abstract: We introduce Omega^2, a Large Language Model-driven framework for corporate credit scoring that combines structured financial data with advanced machine learning to improve predictive reliability and interpretability. Our study evaluates Omega^2 on a multi-agency dataset of 7,800 corporate credit ratings drawn from Moody's, Standard & Poor's, Fitch, and Egan-Jones, each containing detailed firm-level financial indicators such as leverage, profitability, and liquidity ratios. The system integrates CatBoost, LightGBM, and XGBoost models optimized through Bayesian search under temporal validation to ensure forward-looking and reproducible results. Omega^2 achieved a mean test AUC above 0.93 across agencies, confirming its ability to generalize across rating systems and maintain temporal consistency. These results show that combining language-based reasoning with quantitative learning creates a transparent and institution-grade foundation for reliable corporate credit-risk assessment.
Authors: Nadia Daoudi, Ivan Alfonso, Jordi Cabot
Abstract: The development of smart systems (i.e., systems enhanced with AI components) has thrived thanks to the rapid advancements in neural networks (NNs). A wide range of libraries and frameworks have consequently emerged to support NN design and implementation. The choice depends on factors such as available functionalities, ease of use, documentation and community support. After adopting a given NN framework, organizations might later choose to switch to another if performance declines, requirements evolve, or new features are introduced. Unfortunately, migrating NN implementations across libraries is challenging due to the lack of migration approaches specifically tailored for NNs. This leads to increased time and effort to modernize NNs, as manual updates are necessary to avoid relying on outdated implementations and ensure compatibility with new features. In this paper, we propose an approach to automatically migrate neural network code across deep learning frameworks. Our method makes use of a pivot NN model to create an abstraction of the NN prior to migration. We validate our approach using two popular NN frameworks, namely PyTorch and TensorFlow. We also discuss the challenges of migrating code between the two frameworks and how they were approached in our method. Experimental evaluation on five NNs shows that our approach successfully migrates their code and produces NNs that are functionally equivalent to the originals. Artefacts from our work are available online.
Authors: Mohammed Al-Jaff, Giovanni Luca Marchetti, Michael C Welle, Jens Lundell, Mats G. Gustafsson, Gustav Eje Henter, Hossein Azizpour, Danica Kragic
Abstract: Idempotent Generative Networks (IGNs) are deep generative models that also function as local data manifold projectors, mapping arbitrary inputs back onto the manifold. They are trained to act as identity operators on the data and as idempotent operators off the data manifold. However, IGNs suffer from mode collapse, mode dropping, and training instability due to their objectives, which contain adversarial components and can cause the model to cover the data manifold only partially -- an issue shared with generative adversarial networks. We introduce Non-Adversarial Idempotent Generative Networks (NAIGNs) to address these issues. Our loss function combines reconstruction with the non-adversarial generative objective of Implicit Maximum Likelihood Estimation (IMLE). This improves on IGN's ability to restore corrupted data and generate new samples that closely match the data distribution. We moreover demonstrate that NAIGNs implicitly learn the distance field to the data manifold, as well as an energy-based model.
Authors: David Kattermann, Lothar Sebastian Krapp
Abstract: We study computable probably approximately correct (CPAC) learning, where learners are required to be computable functions. It had been previously observed that the Fundamental Theorem of Statistical Learning, which characterizes PAC learnability by finiteness of the Vapnik-Chervonenkis (VC-)dimension, no longer holds in this framework. Recent works recovered analogs of the Fundamental Theorem in the computable setting, for instance by introducing an effective VC-dimension. Guided by this, we investigate the connection between CPAC learning and recursively enumerable representable (RER) classes, whose members can be algorithmically listed. Our results show that the effective VC-dimensions can take arbitrary values above the traditional one, even for RER classes, which creates a whole family of (non-)examples for various notions of CPAC learning. Yet the two dimensions coincide for classes satisfying sufficiently strong notions of CPAC learning. We then observe that CPAC learnability can also be characterized via containment of RER classes that realize the same samples. Furthermore, it is shown that CPAC learnable classes satisfying a unique identification property are necessarily RER. Finally, we establish that agnostic learnability can be guaranteed for RER classes, by considering the relaxed notion of nonuniform CPAC learning.
Authors: Tiziano Balaconi, Aldo Glielmo, Marco Taboga
Abstract: We introduce GasRL, a simulator that couples a calibrated representation of the natural gas market with a model of storage-operator policies trained with deep reinforcement learning (RL). We use it to analyse how optimal stockpile management affects equilibrium prices and the dynamics of demand and supply. We test various RL algorithms and find that Soft Actor Critic (SAC) exhibits superior performance in the GasRL environment: multiple objectives of storage operators - including profitability, robust market clearing and price stabilisation - are successfully achieved. Moreover, the equilibrium price dynamics induced by SAC-derived optimal policies have characteristics, such as volatility and seasonality, that closely match those of real-world prices. Remarkably, this adherence to the historical distribution of prices is obtained without explicitly calibrating the model to price data. We show how the simulator can be used to assess the effects of EU-mandated minimum storage thresholds. We find that such thresholds have a positive effect on market resilience against unanticipated shifts in the distribution of supply shocks. For example, with unusually large shocks, market disruptions are averted more often if a threshold is in place.
Authors: Oleksiy Ostapenko, Luke Kumar, Raymond Li, Denis Kocetkov, Joel Lamy-Poirier, Shruthan Radhakrishna, Soham Parikh, Shambhavi Mishra, Sebastien Paquet, Srinivas Sunkara, Val\'erie B\'ecaert, Sathwik Tejaswi Madhusudhan, Torsten Scholak
Abstract: Large Language Models (LLMs) achieve remarkable reasoning capabilities through transformer architectures with attention mechanisms. However, transformers suffer from quadratic time and memory complexity in the attention module (MHA) and require caching key-value states during inference, which severely limits throughput and scalability. High inference throughput is critical for agentic tasks, long-context reasoning, efficient deployment under high request loads, and more efficient test-time compute scaling. State Space Models (SSMs) such as Mamba offer a promising alternative with linear inference complexity and a constant memory footprint via recurrent computation with fixed-size hidden states. In this technical report we introduce the Apriel-H1 family of hybrid LLMs that combine transformer attention and SSM sequence mixers for efficient reasoning at 15B model size. These models are obtained through incremental distillation from a pretrained reasoning transformer, Apriel-Nemotron-15B-Thinker, progressively replacing less critical attention layers with linear Mamba blocks. We release multiple post-distillation variants of Apriel-H1-15B-Thinker with different SSM-to-MHA ratios and analyse how reasoning performance degrades as more Mamba layers replace MHA. Additionally, we release a 30/50 hybrid variant of Apriel-H1, further fine-tuned on a supervised dataset of reasoning traces, achieving over 2x higher inference throughput when deployed in the production-ready vLLM environment, with minimal degradation in reasoning performance. This shows that distilled hybrid SSM-Transformer architectures can deliver substantial efficiency gains over the pretrained transformer equivalent without substantially compromising the reasoning quality.
Authors: Lihan Xu, Yanjie Dong, Gang Wang, Runhao Zeng, Xiaoyi Fan, Xiping Hu
Abstract: We investigate robust federated learning, where a group of workers collaboratively train a shared model under the orchestration of a central server in the presence of Byzantine adversaries capable of arbitrary and potentially malicious behaviors. To simultaneously enhance communication efficiency and robustness against such adversaries, we propose a Byzantine-resilient Nesterov-Accelerated Federated Learning (Byrd-NAFL) algorithm. Byrd-NAFL seamlessly integrates Nesterov's momentum into the federated learning process alongside Byzantine-resilient aggregation rules to achieve fast and safeguarding convergence against gradient corruption. We establish a finite-time convergence guarantee for Byrd-NAFL under non-convex and smooth loss functions with relaxed assumption on the aggregated gradients. Extensive numerical experiments validate the effectiveness of Byrd-NAFL and demonstrate the superiority over existing benchmarks in terms of convergence speed, accuracy, and resilience to diverse Byzantine attack strategies.
Authors: Cooper Simpson, Stephen Becker, Alireza Doostan
Abstract: Focusing on implicit neural representations, we present a novel in situ training protocol that employs limited memory buffers of full and sketched data samples, where the sketched data are leveraged to prevent catastrophic forgetting. The theoretical motivation for our use of sketching as a regularizer is presented via a simple Johnson-Lindenstrauss-informed result. While our methods may be of wider interest in the field of continual learning, we specifically target in situ neural compression using implicit neural representation-based hypernetworks. We evaluate our method on a variety of complex simulation data in two and three dimensions, over long time horizons, and across unstructured grids and non-Cartesian geometries. On these tasks, we show strong reconstruction performance at high compression rates. Most importantly, we demonstrate that sketching enables the presented in situ scheme to approximately match the performance of the equivalent offline method.
Authors: Giacomo Camposampiero, Pietro Barbiero, Michael Hersche, Roger Wattenhofer, Abbas Rahimi
Abstract: Compositional generalization-a key open challenge in modern machine learning-requires models to predict unknown combinations of known concepts. However, assessing compositional generalization remains a fundamental challenge due to the lack of standardized evaluation protocols and the limitations of current benchmarks, which often favor efficiency over rigor. At the same time, general-purpose vision architectures lack the necessary inductive biases, and existing approaches to endow them compromise scalability. As a remedy, this paper introduces: 1) a rigorous evaluation framework that unifies and extends previous approaches while reducing computational requirements from combinatorial to constant; 2) an extensive and modern evaluation on the status of compositional generalization in supervised vision backbones, training more than 5000 models; 3) Attribute Invariant Networks, a class of models establishing a new Pareto frontier in compositional generalization, achieving a 23.43% accuracy improvement over baselines while reducing parameter overhead from 600% to 16% compared to fully disentangled counterparts.
Authors: Georgios Tzannetos, Parameswaran Kamalaruban, Adish Singla
Abstract: Training agents to operate under strict constraints during deployment, such as limited resource budgets or stringent safety requirements, presents significant challenges, especially when these constraints render the task complex. In this work, we propose a curriculum learning strategy that gradually tightens constraints during training, enabling the agent to incrementally master the deployment requirements. Inspired by self-paced learning techniques in unconstrained reinforcement learning (RL), our approach facilitates a smoother transition to challenging environments by initially training on simplified versions of the constraints and progressively introducing the full deployment conditions. We provide a theoretical analysis using an RL agent in a binary-tree Markov Decision Process (MDP) to demonstrate that our curriculum strategy can accelerate training relative to a baseline approach that imposes the trajectory constraints from the outset. Moreover, we empirically validate the effectiveness and generality of our method across both RL and large language model (LLM) agents in diverse settings, including a binary-tree MDP, a multi-task navigation domain, and a math reasoning task with two benchmarks. These results highlight the potential of curriculum design in enhancing the efficiency and performance of agents operating under complex trajectory constraints during deployment. Moreover, when applied to LLMs, our strategy enables compression of output chain-of-thought tokens, achieving a substantial inference speedup on consumer hardware, demonstrating its effectiveness for resource-constrained deployment.
Authors: Adia Khalid, Alina Deriyeva, Benjamin Paassen
Abstract: Knowledge tracing (KT) models are a crucial basis for pedagogical decision-making, namely which task to select next for a learner and when to stop teaching a particular skill. Given the high stakes of pedagogical decisions, KT models are typically required to be interpretable, in the sense that they should implement an explicit model of human learning and provide explicit estimates of learners' abilities. However, to our knowledge, no study to date has investigated whether the interpretability of KT models actually helps human teachers to make teaching decisions. We address this gap. First, we perform a simulation study to show that, indeed, decisions based on interpretable KT models achieve mastery faster compared to decisions based on a non-interpretable model. Second, we repeat the study but ask $N=12$ human teachers to make the teaching decisions based on the information provided by KT models. As expected, teachers rate interpretable KT models higher in terms of usability and trustworthiness. However, the number of tasks needed until mastery hardly differs between KT models. This suggests that the relationship between model interpretability and teacher decisions is not straightforward: teachers do not solely rely on KT models to make decisions and further research is needed to investigate how learners and teachers actually understand and use KT models.
Authors: Ilies Chibane, Thomas George, Pierre Nodet, Vincent Lemaire
Abstract: Mislabeled data is a pervasive issue that undermines the performance of machine learning systems in real-world applications. An effective approach to mitigate this problem is to detect mislabeled instances and subject them to special treatment, such as filtering or relabeling. Automatic mislabeling detection methods typically rely on training a base machine learning model and then probing it for each instance to obtain a trust score that each provided label is genuine or incorrect. The properties of this base model are thus of paramount importance. In this paper, we investigate the impact of calibrating this model. Our empirical results show that using calibration methods improves the accuracy and robustness of mislabeled instance detection, providing a practical and effective solution for industrial applications.
Authors: Lejs Deen Behric, Liang Zhang, Bingcong Li, Kiran Koshy Thekumparampil
Abstract: Zeroth-order or derivative-free optimization (MeZO) is an attractive strategy for finetuning large language models (LLMs) because it eliminates the memory overhead of backpropagation. However, it converges slowly due to the inherent curse of dimensionality when searching for descent directions in the high-dimensional parameter space of billion-scale LLMs. We propose ConMeZO, a novel zeroth-order optimizer that accelerates convergence by adaptive directional sampling. Instead of drawing the direction uniformly at random, ConMeZO restricts the sampling to a cone centered around a momentum estimate. This concentrates the search in directions where the true gradient is more likely to lie and thus reduces the effect of high dimensions. We prove that ConMeZO achieves the same worst-case convergence rate as MeZO. Empirically, when finetuning LLMs on natural language tasks, ConMeZO is up to 2X faster than MeZO while retaining the low-memory footprint of zeroth-order methods.
Authors: Xun Wang, Zhuoran Li, Yanshan Lin, Hai Zhong, Longbo Huang
Abstract: Training a team of agents from scratch in multi-agent reinforcement learning (MARL) is highly inefficient, much like asking beginners to play a symphony together without first practicing solo. Existing methods, such as offline or transferable MARL, can ease this burden, but they still rely on costly multi-agent data, which often becomes the bottleneck. In contrast, solo experiences are far easier to obtain in many important scenarios, e.g., collaborative coding, household cooperation, and search-and-rescue. To unlock their potential, we propose Solo-to-Collaborative RL (SoCo), a framework that transfers solo knowledge into cooperative learning. SoCo first pretrains a shared solo policy from solo demonstrations, then adapts it for cooperation during multi-agent training through a policy fusion mechanism that combines an MoE-like gating selector and an action editor. Experiments across diverse cooperative tasks show that SoCo significantly boosts the training efficiency and performance of backbone algorithms. These results demonstrate that solo demonstrations provide a scalable and effective complement to multi-agent data, making cooperative learning more practical and broadly applicable.
Authors: Saeed Razavikia, Jos\'e Mairton Barros Da Silva Junior, Carlo Fischione
Abstract: Recently, the ChannelComp framework has proposed digital over-the-air computation by designing digital modulations that enable the computation of arbitrary functions. Unlike traditional analog over-the-air computation, which is restricted to nomographic functions, ChannelComp enables a broader range of computational tasks while maintaining compatibility with digital communication systems. This framework is intended for applications that favor local information processing over the mere acquisition of data. However, ChannelComp is currently designed for scalar function computation, while numerous data-centric applications necessitate vector-based computations, and it is susceptible to channel fading. In this work, we introduce a generalization of the ChannelComp framework, called VecComp, by integrating ChannelComp with multiple-antenna technology. This generalization not only enables vector function computation but also ensures scalability in the computational complexity, which increases only linearly with the vector dimension. As such, VecComp remains computationally efficient and robust against channel impairments, making it suitable for high-dimensional, data-centric applications. We establish a non-asymptotic upper bound on the mean squared error of VecComp, affirming its computation efficiency under fading channel conditions. Numerical experiments show the effectiveness of VecComp in improving the computation of vector functions and fading compensation over noisy and fading multiple-access channels.
Authors: Bum Chul Kwon, Ben Shapira, Moshiko Raboh, Shreyans Sethi, Shruti Murarka, Joseph A Morrone, Jianying Hu, Parthasarathy Suryanarayanan
Abstract: The chemical space of drug-like molecules is vast, motivating the development of generative models that must learn broad chemical distributions, enable conditional generation by capturing structure-property representations, and provide fast molecular generation. Meeting the objectives depends on modeling choices, including the probabilistic modeling approach, the conditional generative formulation, the architecture, and the molecular input representation. To address the challenges, we present STAR-VAE (Selfies-encoded, Transformer-based, AutoRegressive Variational Auto Encoder), a scalable latent-variable framework with a Transformer encoder and an autoregressive Transformer decoder. It is trained on 79 million drug-like molecules from PubChem, using SELFIES to guarantee syntactic validity. The latent-variable formulation enables conditional generation: a property predictor supplies a conditioning signal that is applied consistently to the latent prior, the inference network, and the decoder. Our contributions are: (i) a Transformer-based latent-variable encoder-decoder model trained on SELFIES representations; (ii) a principled conditional latent-variable formulation for property-guided generation; and (iii) efficient finetuning with low-rank adapters (LoRA) in both encoder and decoder, enabling fast adaptation with limited property and activity data. On the GuacaMol and MOSES benchmarks, our approach matches or exceeds baselines, and latent-space analyses reveal smooth, semantically structured representations that support both unconditional exploration and property-aware generation. On the Tartarus benchmarks, the conditional model shifts docking-score distributions toward stronger predicted binding. These results suggest that a modernized, scale-appropriate VAE remains competitive for molecular generation when paired with principled conditioning and parameter-efficient finetuning.
Authors: Xinghan Li, Haodong Wen, Kaifeng Lyu
Abstract: Despite the popularity of the Adam optimizer in practice, most theoretical analyses study Stochastic Gradient Descent (SGD) as a proxy for Adam, and little is known about how the solutions found by Adam differ. In this paper, we show that Adam implicitly reduces a unique form of sharpness measure shaped by its adaptive updates, leading to qualitatively different solutions from SGD. More specifically, when the training loss is small, Adam wanders around the manifold of minimizers and takes semi-gradients to minimize this sharpness measure in an adaptive manner, a behavior we rigorously characterize through a continuous-time approximation using stochastic differential equations. We further demonstrate how this behavior differs from that of SGD in a well-studied setting: when training overparameterized models with label noise, SGD has been shown to minimize the trace of the Hessian matrix, $\tr(\mH)$, whereas we prove that Adam minimizes $\tr(\Diag(\mH)^{1/2})$ instead. In solving sparse linear regression with diagonal linear networks, this distinction enables Adam to achieve better sparsity and generalization than SGD. Finally, our analysis framework extends beyond Adam to a broad class of adaptive gradient methods, including RMSProp, Adam-mini, Adalayer and Shampoo, and provides a unified perspective on how these adaptive optimizers reduce sharpness, which we hope will offer insights for future optimizer design.
Authors: Andras Ferenczi, Sutapa Samanta, Dagen Wang, Todd Hodges
Abstract: Federated learning (FL) is a widely used method for training machine learning (ML) models in a scalable way while preserving privacy (i.e., without centralizing raw data). Prior research shows that the risk of exposing sensitive data increases cumulatively as the number of iterations where a client's updates are included in the aggregated model increase. Attackers can launch membership inference attacks (MIA; deciding whether a sample or client participated), property inference attacks (PIA; inferring attributes of a client's data), and model inversion attacks (MI; reconstructing inputs), thereby inferring client-specific attributes and, in some cases, reconstructing inputs. In this paper, we mitigate risk by substantially reducing per client exposure using a quantum computing-inspired quadratic unconstrained binary optimization (QUBO) formulation that selects a small subset of client updates most relevant for each training round. In this work, we focus on two threat vectors: (i) information leakage by clients during training and (ii) adversaries who can query or obtain the global model. We assume a trusted central server and do not model server compromise. This method also assumes that the server has access to a validation/test set with global data distribution. Experiments on the MNIST dataset with 300 clients in 20 rounds showed a 95.2% per-round and 49% cumulative privacy exposure reduction, with 147 clients' updates never being used during training while maintaining in general the full-aggregation accuracy or even better. The method proved to be efficient at lower scale and more complex model as well. A CINIC-10 dataset-based experiment with 30 clients resulted in 82% per-round privacy improvement and 33% cumulative privacy.
Authors: Mayank Jobanputra, Nils Philipp Walter, Maitrey Mehta, Blerta Veseli, Evan Parker Kelly Chapple, Yifan Wang, Sneha Chetani, Ellie Pavlick, Antonio Vergari, Vera Demberg
Abstract: We present a systematic study of subtraction in large language models (LLMs). While prior benchmarks emphasize addition and multiplication, subtraction has received comparatively little attention despite being structurally distinct as a non-commutative operation. We evaluate eight pretrained LLMs spanning four families on addition and subtraction problems. Our experiments reveal that subtraction accuracy lags behind addition by a wide margin. We find that the errors for ($a-b$) are concentrated in cases where ($a
Authors: Nicolas Riccieri Gardin Assumpcao, Leandro Villas
Abstract: Federated Learning (FL) is a distributed training paradigm wherein participants collaborate to build a global model while ensuring the privacy of the involved data, which remains stored on participant devices. However, proposals aiming to ensure such privacy also make it challenging to protect against potential attackers seeking to compromise the training outcome. In this context, we present Fast, Private, and Protected (FPP), a novel approach that aims to safeguard federated training while enabling secure aggregation to preserve data privacy. This is accomplished by evaluating rounds using participants' assessments and enabling training recovery after an attack. FPP also employs a reputation-based mechanism to mitigate the participation of attackers. We created a dockerized environment to validate the performance of FPP compared to other approaches in the literature (FedAvg, Power-of-Choice, and aggregation via Trimmed Mean and Median). Our experiments demonstrate that FPP achieves a rapid convergence rate and can converge even in the presence of malicious participants performing model poisoning attacks.
Authors: Aditya Tanna, Pratinav Seth, Mohamed Bouadi, Utsav Avaiya, Vinay Kumar Sankarapu
Abstract: Tabular foundation models represent a growing paradigm in structured data learning, extending the benefits of large-scale pretraining to tabular domains. However, their adoption remains limited due to heterogeneous preprocessing pipelines, fragmented APIs, inconsistent fine-tuning procedures, and the absence of standardized evaluation for deployment-oriented metrics such as calibration and fairness. We present TabTune, a unified library that standardizes the complete workflow for tabular foundation models through a single interface. TabTune provides consistent access to seven state-of-the-art models supporting multiple adaptation strategies, including zero-shot inference, meta-learning, supervised fine-tuning (SFT), and parameter-efficient fine-tuning (PEFT). The framework automates model-aware preprocessing, manages architectural heterogeneity internally, and integrates evaluation modules for performance, calibration, and fairness. Designed for extensibility and reproducibility, TabTune enables consistent benchmarking of adaptation strategies of tabular foundation models. The library is open source and available at https://github.com/Lexsi-Labs/TabTune .
Authors: Morgan Allen, Paul Savala
Abstract: In Major League Baseball, strategy and planning are major factors in determining the outcome of a game. Previous studies have aided this by building machine learning models for predicting the winning team of any given game. We extend this work by training a comprehensive set of machine learning models using a common dataset. In addition, we relate the win probabilities produced by these models to win strength as measured by score differential. In doing so we show that the most common machine learning models do indeed demonstrate a relationship between predicted win probability and the strength of the win. Finally, we analyze the results of using predicted win probabilities as a decision making mechanism on run-line betting. We demonstrate positive returns when utilizing appropriate betting strategies, and show that naive use of machine learning models for betting lead to significant loses.
Authors: Hakob Tamazyan, Ani Vanyan, Alvard Barseghyan, Anna Khosrovyan, Evan Shelhamer, Hrant Khachatrian
Abstract: The number and diversity of remote sensing satellites grows over time, while the vast majority of labeled data comes from older satellites. As the foundation models for Earth observation scale up, the cost of (re-)training to support new satellites grows too, so the generalization capabilities of the models towards new satellites become increasingly important. In this work we introduce GeoCrossBench, an extension of the popular GeoBench benchmark with a new evaluation protocol: it tests the in-distribution performance; generalization to satellites with no band overlap; and generalization to satellites with additional bands with respect to the training set. We also develop a self-supervised extension of ChannelViT, ChiViT, to improve its cross-satellite performance. First, we show that even the best foundation models for remote sensing (DOFA, TerraFM) do not outperform general purpose models like DINOv3 in the in-distribution setting. Second, when generalizing to new satellites with no band overlap, all models suffer 2-4x drop in performance, and ChiViT significantly outperforms the runner-up DINOv3. Third, the performance of all tested models drops on average by 5-25\% when given additional bands during test time. Finally, we show that fine-tuning just the last linear layer of these models using oracle labels from all bands can get relatively consistent performance across all satellites, highlighting that the benchmark is far from being saturated. We publicly release the code and the datasets to encourage the development of more future-proof remote sensing models with stronger cross-satellite generalization.
Authors: Abhishek Panigrahi, Bingbin Liu, Sadhika Malladi, Sham Kakade, Surbhi Goel
Abstract: Knowledge distillation is an efficient strategy to use data generated by large "teacher" language models to train smaller capable "student" models, but selecting the optimal teacher for a specific student-task combination requires expensive trial-and-error. We propose a lightweight score called GRACE to quantify how effective a teacher will be for post-training a student model. GRACE measures distributional properties of the student's gradients without access to a verifier, teacher logits, teacher internals, or test data. From an information-theoretic perspective, GRACE connects to leave-one-out stability of gradient-based algorithms, which controls the generalization performance of the distilled students. On GSM8K and MATH, GRACE correlates strongly (up to 86% Spearman correlation) with the performance of the distilled LLaMA and OLMo students. In particular, training a student using the GRACE-selected teacher can improve the performance by up to 7.4% over naively using the best-performing teacher. Further, GRACE can provide guidance on crucial design choices in distillation, including (1) the best temperature to use when generating from the teacher, (2) the best teacher to use given a size constraint, and (3) the best teacher to use within a specific model family. Altogether, our findings demonstrate that GRACE can efficiently and effectively identify a strongly compatible teacher for a given student and provide fine-grained guidance on how to perform distillation.
Authors: Amrita Ghosh, Mugdha Sarkar, Ying-Jer Kao, Pochung Chen
Abstract: We propose the use of the ``spin-opstring", derived from Stochastic Series Expansion Quantum Monte Carlo (QMC) simulations as machine learning (ML) input data. It offers a compact, memory-efficient representation of QMC simulation cells, combining the initial state with an operator string that encodes the state's evolution through imaginary time. Using supervised ML, we demonstrate the input's effectiveness in capturing both conventional and topological phase transitions, and in a regression task to predict non-local observables. We also demonstrate the capability of spin-opstring data in transfer learning by training models on one quantum system and successfully predicting on another, as well as showing that models trained on smaller system sizes generalize well to larger ones. Importantly, we illustrate a clear advantage of spin-opstring over conventional spin configurations in the accurate prediction of a quantum phase transition. Finally, we show how the inherent structure of spin-opstring provides an elegant framework for the interpretability of ML predictions. Using two state-of-the-art interpretability techniques, Layer-wise Relevance Propagation and SHapley Additive exPlanations, we show that the ML models learn and rely on physically meaningful features from the input data. Together, these findings establish the spin-opstring as a broadly-applicable and interpretable input format for ML in quantum many-body physics.
Authors: Ching-Chih Sung, Shuntaro Suzuki, Francis Pingfan Chien, Komei Sugiura, Yu Tsao
Abstract: Clarifying the neural basis of speech intelligibility is critical for computational neuroscience and digital speech processing. Recent neuroimaging studies have shown that intelligibility modulates cortical activity beyond simple acoustics, primarily in the superior temporal and inferior frontal gyri. However, previous studies have been largely confined to clean speech, leaving it unclear whether the brain employs condition-invariant neural codes across diverse listening environments. To address this gap, we propose a novel architecture built upon a deep state space model for decoding intelligibility from fMRI signals, specifically tailored to their high-dimensional temporal structure. We present the first attempt to decode intelligibility across acoustically distinct conditions, showing our method significantly outperforms classical approaches. Furthermore, region-wise analysis highlights contributions from auditory, frontal, and parietal regions, and cross-condition transfer indicates the presence of condition-invariant neural codes, thereby advancing understanding of abstract linguistic representations in the brain.
Authors: Toby Barter, Zheng Gao, Eva Christodoulaki, Jing Chen, John Cartlidge
Abstract: Bond markets respond differently to macroeconomic news compared to equity markets, yet most sentiment models, including FinBERT, are trained primarily on general financial or equity news data. This mismatch is important because bond prices often move in the opposite direction to economic optimism, making general or equity-based sentiment tools potentially misleading. In this paper, we introduce BondBERT, a transformer-based language model fine-tuned on bond-specific news. BondBERT can act as the perception and reasoning component of a financial decision-support agent, providing sentiment signals that integrate with forecasting models. It is a generalisable framework for adapting transformers to low-volatility, domain-inverse sentiment tasks by compiling and cleaning 30,000 UK bond market articles (2018--2025) for training, validation, and testing. We compare BondBERT's sentiment predictions against FinBERT, FinGPT, and Instruct-FinGPT using event-based correlation, up/down accuracy analyses, and LSTM forecasting across ten UK sovereign bonds. We find that BondBERT consistently produces positive correlations with bond returns, achieves higher alignment and forecasting accuracy than the three baseline models, with lower normalised RMSE and higher information coefficient. These results demonstrate that domain-specific sentiment adaptation better captures fixed income dynamics, bridging a gap between NLP advances and bond market analytics.
Authors: Christian Schiffer, Zeynep Boztoprak, Jan-Oliver Kropp, Julia Th\"onni{\ss}en, Katia Berr, Hannah Spitzer, Katrin Amunts, Timo Dickscheid
Abstract: To study how the human brain works, we need to explore the organization of the cerebral cortex and its detailed cellular architecture. We introduce CytoNet, a foundation model that encodes high-resolution microscopic image patches of the cerebral cortex into highly expressive feature representations, enabling comprehensive brain analyses. CytoNet employs self-supervised learning using spatial proximity as a powerful training signal, without requiring manual labelling. The resulting features are anatomically sound and biologically relevant. They encode general aspects of cortical architecture and unique brain-specific traits. We demonstrate top-tier performance in tasks such as cortical area classification, cortical layer segmentation, cell morphology estimation, and unsupervised brain region mapping. As a foundation model, CytoNet offers a consistent framework for studying cortical microarchitecture, supporting analyses of its relationship with other structural and functional brain features, and paving the way for diverse neuroscientific investigations.
Authors: Etash Guha, Tianxiao Jiang, Andrew Deng, Jian Zhang, Muthu Annamalai
Abstract: Mapping a dataflow-graph of an ML model onto a reconfigurable system is difficult, as different mappings have different throughputs and consume resource constraints differently. To solve this, a model to evaluate the throughput of mappings is necessary as measuring throughput completely is expensive. Many use a hand-designed analytical model, relying on proxy features or intuition, introducing error. We provide a Learned Approach that predicts throughput 31%-52% more accurately over a variety of graphs. In addition, our approach shows no accuracy degradation after removing performance annotations. We show that using this approach results in 5.6% faster compiled graphs.
Authors: Elaina Rohlfing, Azim Ahmadzadeh, V Aparna
Abstract: Solar-flare forecasting has been extensively researched yet remains an open problem. In this paper, we investigate the contributions of elastic distance measures for detecting patterns in the solar-flare dataset, SWAN-SF. We employ a simple $k$-medoids clustering algorithm to evaluate the effectiveness of advanced, high-dimensional distance metrics. Our results show that, despite thorough optimization, none of the elastic distances outperform Euclidean distance by a significant margin. We demonstrate that, although elastic measures have shown promise for univariate time series, when applied to the multivariate time series of SWAN-SF, characterized by the high stochasticity of solar activity, they effectively collapse to Euclidean distance. We conduct thousands of experiments and present both quantitative and qualitative evidence supporting this finding.
Authors: HM Shadman Tabib, Md. Hasnaen Adil, Ayesha Rahman, Ahmmad Nur Swapnil, Maoyejatun Hasana, Ahmed Hossain Chowdhury, A. B. M. Alim Al Islam
Abstract: Access to clinical multi-channel EEG remains limited in many regions worldwide. We present NEUROSKY-EPI, the first open dataset of single-channel, consumer-grade EEG for epilepsy, collected in a South Asian clinical setting along with rich contextual metadata. To explore its utility, we introduce EmbedCluster, a patient-stratification pipeline that transfers representations from EEGNet models trained on clinical data and enriches them with contextual autoencoder embeddings, followed by unsupervised clustering of patients based on EEG patterns. Results show that low-cost, single-channel data can support meaningful stratification. Beyond algorithmic performance, we emphasize human-centered concerns such as deployability in resource-constrained environments, interpretability for non-specialists, and safeguards for privacy, inclusivity, and bias. By releasing the dataset and code, we aim to catalyze interdisciplinary research across health technology, human-computer interaction, and machine learning, advancing the goal of affordable and actionable EEG-based epilepsy care.
Authors: Robyn Wyrick
Abstract: As artificial intelligence (AI) advances toward superhuman capabilities, aligning these systems with human values becomes increasingly critical. Current alignment strategies rely largely on externally specified constraints that may prove insufficient against future super-intelligent AI capable of circumventing top-down controls. This research investigates whether artificial neural networks (ANNs) can develop patterns analogous to biological mirror neurons cells that activate both when performing and observing actions, and how such patterns might contribute to intrinsic alignment in AI. Mirror neurons play a crucial role in empathy, imitation, and social cognition in humans. The study therefore asks: (1) Can simple ANNs develop mirror-neuron patterns? and (2) How might these patterns contribute to ethical and cooperative decision-making in AI systems? Using a novel Frog and Toad game framework designed to promote cooperative behaviors, we identify conditions under which mirror-neuron patterns emerge, evaluate their influence on action circuits, introduce the Checkpoint Mirror Neuron Index (CMNI) to quantify activation strength and consistency, and propose a theoretical framework for further study. Our findings indicate that appropriately scaled model capacities and self/other coupling foster shared neural representations in ANNs similar to biological mirror neurons. These empathy-like circuits support cooperative behavior and suggest that intrinsic motivations modeled through mirror-neuron dynamics could complement existing alignment techniques by embedding empathy-like mechanisms directly within AI architectures.
Authors: Fangbing Liu, Pengfei Duan, Wen Li, Yi He
Abstract: Recent advancements have demonstrated the great potential of flow matching-based Multimodal Large Language Models (MLLMs) in image editing. However, state-of-the-art works like BAGEL face limitations, including detail degradation, content inconsistency, and inefficiency due to their reliance on random noise initialization. To address these issues, we propose LGCC, a novel framework with two key components: Local Gaussian Noise Coupling (LGNC) and Content Consistency Loss (CCL). LGNC preserves spatial details by modeling target image embeddings and their locally perturbed counterparts as coupled pairs, while CCL ensures semantic alignment between edit instructions and image modifications, preventing unintended content removal. By integrating LGCC with the BAGEL pre-trained model via curriculum learning, we significantly reduce inference steps, improving local detail scores on I2EBench by 1.60% and overall scores by 0.53%. LGCC achieves 3x -- 5x speedup for lightweight editing and 2x for universal editing, requiring only 40% -- 50% of the inference time of BAGEL or Flux. These results demonstrate LGCC's ability to preserve detail, maintain contextual integrity, and enhance inference speed, offering a cost-efficient solution without compromising editing quality.
Authors: Wenzhe Fan, Ning Yan, Masood Mortazavi
Abstract: Planning has been a cornerstone of artificial intelligence for solving complex problems, and recent progress in LLM-based multi-agent frameworks have begun to extend this capability. However, the role of human-like memory within these frameworks remains largely unexplored. Understanding how agents coordinate through memory is critical for natural language planning, where iterative reasoning, constraint tracking, and error correction drive the success. Inspired by working memory model in cognitive psychology, we present EvoMem, a multi-agent framework built on a dual-evolving memory mechanism. The framework consists of three agents (Constraint Extractor, Verifier, and Actor) and two memory modules: Constraint Memory (CMem), which evolves across queries by storing task-specific rules and constraints while remains fixed within a query, and Query-feedback Memory (QMem), which evolves within a query by accumulating feedback across iterations for solution refinement. Both memory modules are reset at the end of each query session. Evaluations on trip planning, meeting planning, and calendar scheduling show consistent performance improvements, highlighting the effectiveness of EvoMem. This success underscores the importance of memory in enhancing multi-agent planning.
Authors: Leonardo C\'azares-Trejo, Marco Loreto-Silva, Huziel E. Sauceda
Abstract: Noncovalent interactions--vdW dispersion, hydrogen/halogen bonding, ion-$\pi$, and $\pi$-stacking--govern structure, dynamics, and emergent phenomena in materials and molecular systems, yet accurately learning them alongside covalent forces remains a core challenge for machine-learned force fields (MLFFs). This challenge is acute for global models that use Coulomb-matrix (CM) descriptors compared under Euclidean/Frobenius metrics in multifragment settings. We show that the mismatch between predominantly covalent force labels and the CM's overrepresentation of intermolecular features biases single-model training and degrades force-field fidelity. To address this, we introduce \textit{$\Delta$-sGDML}, a scale-aware formulation within the sGDML framework that explicitly decouples intra- and intermolecular physics by training fragment-specific models alongside a dedicated binding model, then composing them at inference. Across benzene dimers, host-guest complexes (C$_{60}$@buckycatcher, NO$_3^-$@i-corona[6]arene), benzene-water, and benzene-Na$^+$, \mbox{$\Delta$-sGDML} delivers consistent gains over a single global model, with fragment-resolved force-error reductions up to \textbf{75\%}, without loss of energy accuracy. Furthermore, molecular-dynamics simulations further confirm that the $\Delta$-model yields a reliable force field for C$_{60}$@buckycatcher, producing stable trajectories across a wide range of temperatures (10-400~K), unlike the single global model, which loses stability above $\sim$200~K. The method offers a practical route to homogenize per-fragment errors and recover reliable noncovalent physics in global MLFFs.
Authors: Eleonora Villa, Golam Mohiuddin Shaifullah, Andrea Possenti, Carmelita Carbone
Abstract: We present a detailed study of Bayesian inference workflows for pulsar timing array data with a focus on enhancing efficiency, robustness and speed through the use of normalizing flow-based nested sampling. Building on the Enterprise framework, we integrate the i-nessai sampler and benchmark its performance on realistic, simulated datasets. We analyze its computational scaling and stability, and show that it achieves accurate posteriors and reliable evidence estimates with substantially reduced runtime, by up to three orders of magnitude depending on the dataset configuration, with respect to conventional single-core parallel-tempering MCMC analyses. These results highlight the potential of flow-based nested sampling to accelerate PTA analyses while preserving the quality of the inference.
Authors: Eleonora Villa, Luigi D'Amico, Aldo Barca, Fatima Modica Bittordo, Francesco Al\`i, Massimo Meneghetti, Luca Naso
Abstract: Pulsar Timing Arrays provide a powerful framework to measure low-frequency gravitational waves, but accuracy and robustness of the results are challenged by complex noise processes that must be accurately modeled. Standard PTA analyses assign fixed uniform noise priors to each pulsar, an approach that can introduce systematic biases when combining the array. To overcome this limitation, we adopt a hierarchical Bayesian modeling strategy in which noise priors are parametrized by higher-level hyperparameters. We further address the challenge posed by the correlations between hyperparameters and physical noise parameters, focusing on those describing red noise and dispersion measure variations. To decorrelate these quantities, we introduce an orthogonal reparametrization of the hierarchical model implemented with Normalizing Flows. We also employ i-nessai, a flow-guided nested sampler, to efficiently explore the resulting higher-dimensional parameter space. We apply our method to a minimal 3-pulsar case study, performing a simultaneous inference of noise and SGWB parameters. Despite the limited dataset, the results consistently show that the hierarchical treatment constrains the noise parameters more tightly and partially alleviates the red-noise-SGWB degeneracy, while the orthogonal reparametrization further enhances parameter independence without affecting the correlations intrinsic to the power-law modeling of the physical processes involved.
Authors: Yifan F. Zhang, Sarang Gopalakrishnan
Abstract: We prove that the Gibbs states of classical, and commuting-Pauli, Hamiltonians are stable under weak local decoherence: i.e., we show that the effect of the decoherence can be locally reversed. In particular, our conclusions apply to finite-temperature equilibrium critical points and ordered low-temperature phases. In these systems the unconditional spatio-temporal correlations are long-range, and local (e.g., Metropolis) dynamics exhibits critical slowing down. Nevertheless, our results imply the existence of local "decoders" that undo the decoherence, when the decoherence strength is below a critical value. An implication of these results is that thermally stable quantum memories have a threshold against decoherence that remains nonzero as one approaches the critical temperature. Analogously, in diffusion models, stability of data distributions implies the existence of computationally-efficent local denoisers in the late-time generation dynamics.
Authors: Pradyun Hebbar, Thandikire Madula, Vinicius Mikuni, Benjamin Nachman, Nadav Outmezguine, Inbar Savoray
Abstract: Physical symmetries provide a strong inductive bias for constructing functions to analyze data. In particular, this bias may improve robustness, data efficiency, and interpretability of machine learning models. However, building machine learning models that explicitly respect symmetries can be difficult due to the dedicated components required. Moreover, real-world experiments may not exactly respect fundamental symmetries at the level of finite granularities and energy thresholds. In this work, we explore an alternative approach to create symmetry-aware machine learning models. We introduce soft constraints that allow the model to decide the importance of added symmetries during the learning process instead of enforcing exact symmetries. We investigate two complementary approaches, one that penalizes the model based on specific transformations of the inputs and one inspired by group theory and infinitesimal transformations of the inputs. Using top quark jet tagging and Lorentz equivariance as examples, we observe that the addition of the soft constraints leads to more robust performance while requiring negligible changes to current state-of-the-art models.
Authors: Karol Radziszewski, Micha{\l} Szpunar, Piotr Ociepka, Mateusz Buczy\'nski
Abstract: We present a scalable recommender system implementation based on RippleNet, tailored for the media domain with a production deployment in Onet.pl, one of Poland's largest online media platforms. Our solution addresses the cold-start problem for newly published content by integrating content-based item embeddings into the knowledge propagation mechanism of RippleNet, enabling effective scoring of previously unseen items. The system architecture leverages Amazon SageMaker for distributed training and inference, and Apache Airflow for orchestrating data pipelines and model retraining workflows. To ensure high-quality training data, we constructed a comprehensive golden dataset consisting of user and item features and a separate interaction table, all enabling flexible extensions and integration of new signals.
Authors: Jinchao Feng, Charles Kulick, Sui Tang
Abstract: We develop a Gaussian process framework for learning interaction kernels in multi-species interacting particle systems from trajectory data. Such systems provide a canonical setting for multiscale modeling, where simple microscopic interaction rules generate complex macroscopic behaviors. While our earlier work established a Gaussian process approach and convergence theory for single-species systems, and later extended to second-order models with alignment and energy-type interactions, the multi-species setting introduces new challenges: heterogeneous populations interact both within and across species, the number of unknown kernels grows, and asymmetric interactions such as predator-prey dynamics must be accommodated. We formulate the learning problem in a nonparametric Bayesian setting and establish rigorous statistical guarantees. Our analysis shows recoverability of the interaction kernels, provides quantitative error bounds, and proves statistical optimality of posterior estimators, thereby unifying and generalizing previous single-species theory. Numerical experiments confirm the theoretical predictions and demonstrate the effectiveness of the proposed approach, highlighting its advantages over existing kernel-based methods. This work contributes a complete statistical framework for data-driven inference of interaction laws in multi-species systems, advancing the broader multiscale modeling program of connecting microscopic particle dynamics with emergent macroscopic behavior.
Authors: Renos Zabounidis, Aditya Golatkar, Michael Kleinman, Alessandro Achille, Wei Xia, Stefano Soatto
Abstract: We propose Re-FORC, an adaptive reward prediction method that, given a context, enables prediction of the expected future rewards as a function of the number of future thinking tokens. Re-FORC trains a lightweight adapter on reasoning models, demonstrating improved prediction with longer reasoning and larger models. Re-FORC enables: 1) early stopping of unpromising reasoning chains, reducing compute by 26% while maintaining accuracy, 2) optimized model and thinking length selection that achieves 4% higher accuracy at equal compute and 55% less compute at equal accuracy compared to the largest model, 3) adaptive test-time scaling, which increases accuracy by 11% in high compute regime, and 7% in low compute regime. Re-FORC allows dynamic reasoning with length control via cost-per-token thresholds while estimating computation time upfront.
Authors: Mansi Choudhary, Karthik Sangaiah, Sonali Singh, Muhammad Osama, Lisa Wu Wills, Ganesh Dasika
Abstract: The rise of disaggregated AI GPUs has exposed a critical bottleneck in large-scale attention workloads: non-uniform memory access (NUMA). As multi-chiplet designs become the norm for scaling compute capabilities, memory latency and bandwidth vary sharply across compute regions, undermining the performance of traditional GPU kernel scheduling strategies that assume uniform memory access. We identify how these NUMA effects distort locality in multi-head attention (MHA) and present Swizzled Head-first Mapping, a spatially-aware scheduling strategy that aligns attention heads with GPU NUMA domains to exploit intra-chiplet cache reuse. On AMD's MI300X architecture, our method achieves up to 50% higher performance over state-of-the-art attention algorithms using conventional scheduling techniques and sustains consistently high L2 cache hit rates of 80-97%. These results demonstrate that NUMA-aware scheduling is now fundamental to achieving full efficiency on next-generation disaggregated GPUs, offering a path forward for scalable AI training and inference.
Authors: Dongze Wu, Feng Qiu, Yao Xie
Abstract: Time-series forecasting increasingly demands not only accurate observational predictions but also causal forecasting under interventional and counterfactual queries in multivariate systems. We present DoFlow, a flow based generative model defined over a causal DAG that delivers coherent observational and interventional predictions, as well as counterfactuals through the natural encoding and decoding mechanism of continuous normalizing flows (CNFs). We also provide a supporting counterfactual recovery result under certain assumptions. Beyond forecasting, DoFlow provides explicit likelihoods of future trajectories, enabling principled anomaly detection. Experiments on synthetic datasets with various causal DAG and real world hydropower and cancer treatment time series show that DoFlow achieves accurate system-wide observational forecasting, enables causal forecasting over interventional and counterfactual queries, and effectively detects anomalies. This work contributes to the broader goal of unifying causal reasoning and generative modeling for complex dynamical systems.
Authors: Asrin Efe Yorulmaz, Tamer Ba\c{s}ar
Abstract: No-regret learning dynamics play a central role in game theory, enabling decentralized convergence to equilibrium for concepts such as Coarse Correlated Equilibrium (CCE) or Correlated Equilibrium (CE). In this work, we improve the convergence rate to CCE in general-sum Markov games, reducing it from the previously best-known rate of $\mathcal{O}(\log^5 T / T)$ to a sharper $\mathcal{O}(\log T / T)$. This matches the best known convergence rate for CE in terms of $T$, number of iterations, while also improving the dependence on the action set size from polynomial to polylogarithmic-yielding exponential gains in high-dimensional settings. Our approach builds on recent advances in adaptive step-size techniques for no-regret algorithms in normal-form games, and extends them to the Markovian setting via a stage-wise scheme that adjusts learning rates based on real-time feedback. We frame policy updates as an instance of Optimistic Follow-the-Regularized-Leader (OFTRL), customized for value-iteration-based learning. The resulting self-play algorithm achieves, to our knowledge, the fastest known convergence rate to CCE in Markov games.
Authors: Eric Vin, Kyle A. Miller, Inigo Incer, Sanjit A. Seshia, Daniel J. Fremont
Abstract: Full verification of learning-enabled cyber-physical systems (CPS) has long been intractable due to challenges including black-box components and complex real-world environments. Existing tools either provide formal guarantees for limited types of systems or test the system as a monolith, but no general framework exists for compositional analysis of learning-enabled CPS using varied verification techniques over complex real-world environments. This paper introduces ScenicProver, a verification framework that aims to fill this gap. Built upon the Scenic probabilistic programming language, the framework supports: (1) compositional system description with clear component interfaces, ranging from interpretable code to black boxes; (2) assume-guarantee contracts over those components using an extension of Linear Temporal Logic containing arbitrary Scenic expressions; (3) evidence generation through testing, formal proofs via Lean 4 integration, and importing external assumptions; (4) systematic combination of generated evidence using contract operators; and (5) automatic generation of assurance cases tracking the provenance of system-level guarantees. We demonstrate the framework's effectiveness through a case study on an autonomous vehicle's automatic emergency braking system with sensor fusion. By leveraging manufacturer guarantees for radar and laser sensors and focusing testing efforts on uncertain conditions, our approach enables stronger probabilistic guarantees than monolithic testing with the same computational budget.
Authors: Octavian Alexandru Trifan, Karthik Sangaiah, Muhammad Awad, Muhammad Osama, Sumanth Gudaparthi, Alexandru Nicolau, Alexander Veidenbaum, Ganesh Dasika
Abstract: As large language models (LLMs) continue to scale, their workloads increasingly rely on distributed execution across multiple GPUs. However, the conventional bulk synchronous parallel~(BSP) model used in such settings introduces significant performance inefficiencies. To characterize these bottlenecks, we introduce the ''Three Taxes'' (Bulk Synchronous, Inter-Kernel Data Locality, and Kernel Launch Overhead) as an analytical framework. We propose moving beyond the rigid BSP model to address key inefficiencies in distributed GPU execution. By exploiting libraries like Iris for Triton, we gain access to in-kernel communication primitives that enable the design of novel fine-grained programming patterns, offering greater flexibility and performance than traditional BSP-based approaches. These patterns systematically eliminate the three taxes by creating direct, tile-level producer-consumer pipelines and replacing global barriers with fine-grained dataflow synchronization. Applying this methodology to critical kernels, from the foundational All-Gather + general matrix multiplication operation to the complex Flash Decode algorithm, we observe a 10-20% speedup in end-to-end latency over BSP-based approaches, establishing a more programmable and efficient paradigm for distributed LLM workloads.
Authors: Fuyi Wang, Zekai Chen, Mingyuan Fan, Jianying Zhou, Lei Pan, Leo Yu Zhang
Abstract: Graph neural networks (GNNs) are powerful tools for analyzing and learning from graph-structured (GS) data, facilitating a wide range of services. Deploying such services in privacy-critical cloud environments necessitates the development of secure inference (SI) protocols that safeguard sensitive GS data. However, existing SI solutions largely focus on convolutional models for image and text data, leaving the challenge of securing GNNs and GS data relatively underexplored. In this work, we design, implement, and evaluate $\sysname$, a lightweight cryptographic scheme for graph-centric inference in the cloud. By hybridizing additive and function secret sharings within secure two-party computation (2PC), $\sysname$ is carefully designed based on a series of novel 2PC interactive protocols that achieve $1.5\times \sim 1.7\times$ speedups for linear layers and $2\times \sim 15\times$ for non-linear layers over state-of-the-art (SotA) solutions. A thorough theoretical analysis is provided to prove $\sysname$'s correctness, security, and lightweight nature. Extensive experiments across four datasets demonstrate $\sysname$'s superior efficiency with $1.3\times \sim 4.7\times$ faster secure predictions while maintaining accuracy comparable to plaintext graph property inference.
Authors: Yibo Zhao, Yang Zhao, Hongru Du, Hao Frank Yang
Abstract: Decision-making models for individuals, particularly in high-stakes scenarios like vaccine uptake, often diverge from population optimal predictions. This gap arises from the uniqueness of the individual decision-making process, shaped by numerical attributes (e.g., cost, time) and linguistic influences (e.g., personal preferences and constraints). Developing upon Utility Theory and leveraging the textual-reasoning capabilities of Large Language Models (LLMs), this paper proposes an Adaptive Textual-symbolic Human-centric Reasoning framework (ATHENA) to address the optimal information integration. ATHENA uniquely integrates two stages: First, it discovers robust, group-level symbolic utility functions via LLM-augmented symbolic discovery; Second, it implements individual-level semantic adaptation, creating personalized semantic templates guided by the optimal utility to model personalized choices. Validated on real-world travel mode and vaccine choice tasks, ATHENA consistently outperforms utility-based, machine learning, and other LLM-based models, lifting F1 score by at least 6.5% over the strongest cutting-edge models. Further, ablation studies confirm that both stages of ATHENA are critical and complementary, as removing either clearly degrades overall predictive performance. By organically integrating symbolic utility modeling and semantic adaptation, ATHENA provides a new scheme for modeling human-centric decisions. The project page can be found at https://yibozh.github.io/Athena.
Authors: Weiwei Sun, Xuhui Zhou, Weihua Du, Xingyao Wang, Sean Welleck, Graham Neubig, Maarten Sap, Yiming Yang
Abstract: While existing work focuses primarily on task success, we argue that effective real-world agents require optimizing three dimensions: productivity (task completion), proactivity (asking essential questions), and personalization (adapting to diverse user preferences). We introduce UserVille, an interactive environment with LLM-based user simulators enabling diverse, configurable user preferences. Leveraging UserVille, we introduce PPP, a multi-objective reinforcement learning approach that jointly optimizes all three dimensions: Productivity, Proactivity, and Personalization. Experiments on software engineering and deep research tasks show that agents trained with PPP achieve substantial improvements over strong baselines such as GPT-5 (+21.6 on average), demonstrating the ability to ask strategic clarifying questions, adapt to unseen user preferences, and improve task success through better interaction. This work demonstrates that explicitly optimizing for user-centered interaction is critical for building practical and effective AI agents.
Authors: Manonmani Sekar, Nasim Nezamoddini
Abstract: One of the main challenges in managing traffic at multilane intersections is ensuring smooth coordination between human-driven vehicles (HDVs) and connected autonomous vehicles (CAVs). This paper presents a novel traffic signal control framework that combines Graph Attention Networks (GAT) with Soft Actor-Critic (SAC) reinforcement learning to address this challenge. GATs are used to model the dynamic graph- structured nature of traffic flow to capture spatial and temporal dependencies between lanes and signal phases. The proposed SAC is a robust off-policy reinforcement learning algorithm that enables adaptive signal control through entropy-optimized decision making. This design allows the system to coordinate the signal timing and vehicle movement simultaneously with objectives focused on minimizing travel time, enhancing performance, ensuring safety, and improving fairness between HDVs and CAVs. The model is evaluated using a SUMO-based simulation of a four-way intersection and incorporating different traffic densities and CAV penetration rates. The experimental results demonstrate the effectiveness of the GAT-SAC approach by achieving a 24.1% reduction in average delay and up to 29.2% fewer traffic violations compared to traditional methods. Additionally, the fairness ratio between HDVs and CAVs improved to 1.59, indicating more equitable treatment across vehicle types. These findings suggest that the GAT-SAC framework holds significant promise for real-world deployment in mixed-autonomy traffic systems.
Authors: Brennen A. Hill
Abstract: Traditional neural networks, while powerful, rely on biologically implausible learning mechanisms such as global backpropagation. This paper introduces the Structurally Adaptive Predictive Inference Network (SAPIN), a novel computational model inspired by the principles of active inference and the morphological plasticity observed in biological neural cultures. SAPIN operates on a 2D grid where processing units, or cells, learn by minimizing local prediction errors. The model features two primary, concurrent learning mechanisms: a local, Hebbian-like synaptic plasticity rule based on the temporal difference between a cell's actual activation and its learned expectation, and a structural plasticity mechanism where cells physically migrate across the grid to optimize their information-receptive fields. This dual approach allows the network to learn both how to process information (synaptic weights) and also where to position its computational resources (network topology). We validated the SAPIN model on the classic Cart Pole reinforcement learning benchmark. Our results demonstrate that the architecture can successfully solve the CartPole task, achieving robust performance. The network's intrinsic drive to minimize prediction error and maintain homeostasis was sufficient to discover a stable balancing policy. We also found that while continual learning led to instability, locking the network's parameters after achieving success resulted in a stable policy. When evaluated for 100 episodes post-locking (repeated over 100 successful agents), the locked networks maintained an average 82% success rate.
Authors: Jonathan Liu, Haoling Qiu, Jonathan Lasko, Damianos Karakos, Mahsa Yarmohammadi, Mark Dredze
Abstract: Recent research has shown that hallucinations, omissions, and biases are prevalent in everyday use-cases of LLMs. However, chatbots used in medical contexts must provide consistent advice in situations where non-medical factors are involved, such as when demographic information is present. In order to understand the conditions under which medical chatbots fail to perform as expected, we develop an infrastructure that 1) automatically generates queries to probe LLMs and 2) evaluates answers to these queries using multiple LLM-as-a-judge setups and prompts. For 1), our prompt creation pipeline samples the space of patient demographics, histories, disorders, and writing styles to create realistic questions that we subsequently use to prompt LLMs. In 2), our evaluation pipeline provides hallucination and omission detection using LLM-as-a-judge as well as agentic workflows, in addition to LLM-as-a-judge treatment category detectors. As a baseline study, we perform two case studies on inter-LLM agreement and the impact of varying the answering and evaluation LLMs. We find that LLM annotators exhibit low agreement scores (average Cohen's Kappa $\kappa=0.118$), and only specific (answering, evaluation) LLM pairs yield statistically significant differences across writing styles, genders, and races. We recommend that studies using LLM evaluation use multiple LLMs as evaluators in order to avoid arriving at statistically significant but non-generalizable results, particularly in the absence of ground-truth data. We also suggest publishing inter-LLM agreement metrics for transparency. Our code and dataset are available here: https://github.com/BBN-E/medic-neurips-2025-demo.
Authors: Xingqi Cui, Chieh-Jan Mike Liang, Jiarong Xing, Haoran Qiu
Abstract: Serving large generative models such as LLMs and multi- modal transformers requires balancing user-facing SLOs (e.g., time-to-first-token, time-between-tokens) with provider goals of efficiency and cost reduction. Existing solutions rely on static provisioning or model-level autoscaling, both of which treat the model as a monolith. This coarse-grained resource management leads to degraded performance or significant resource underutilization due to poor adaptability to dynamic inference traffic that is common online. The root cause of this inefficiency lies in the internal structure of generative models: they are executed as graphs of interconnected operators. Through detailed characterization and systematic analysis, we find that operators are heterogeneous in their compute and memory footprints and exhibit diverse sensitivity to workload and resource factors such as batch size, sequence length, and traffic rate. This heterogeneity suggests that the operator, rather than the entire model, is the right granularity for scaling decisions. We propose an operator-level autoscaling framework, which allocates resources at finer (operator)-granularity, optimizing the scaling, batching, and placement based on individual operator profiles. Evaluated on production-scale traces, our approach preserves SLOs with up to 40% fewer GPUs and 35% less energy, or under fixed resources achieves 1.6x higher throughput with 5% less energy. These results show that the operator, rather than the model, is fundamentally a more effective unit for scaling large generative workloads.
Authors: Parsa Rangriz
Abstract: This paper studies the high-dimensional scaling limits of online stochastic gradient descent (SGD) for single-layer networks. Building on the seminal work of Saad and Solla, which analyzed the deterministic (ballistic) scaling limits of SGD corresponding to the gradient flow of the population loss, we focus on the critical scaling regime of the step size. Below this critical scale, the effective dynamics are governed by ballistic (ODE) limits, but at the critical scale, new correction term appears that changes the phase diagram. In this regime, near the fixed points, the corresponding diffusive (SDE) limits of the effective dynamics reduces to an Ornstein-Uhlenbeck process under certain conditions. These results highlight how the information exponent controls sample complexity and illustrates the limitations of deterministic scaling limit in capturing the stochastic fluctuations of high-dimensional learning dynamics.
Authors: Beyazit Yalcinkaya, Marcell Vazquez-Chanlatte, Ameesh Shah, Hanna Krasowski, Sanjit A. Seshia
Abstract: We study the problem of learning multi-task, multi-agent policies for cooperative, temporal objectives, under centralized training, decentralized execution. In this setting, using automata to represent tasks enables the decomposition of complex tasks into simpler sub-tasks that can be assigned to agents. However, existing approaches remain sample-inefficient and are limited to the single-task case. In this work, we present Automata-Conditioned Cooperative Multi-Agent Reinforcement Learning (ACC-MARL), a framework for learning task-conditioned, decentralized team policies. We identify the main challenges to ACC-MARL's feasibility in practice, propose solutions, and prove the correctness of our approach. We further show that the value functions of learned policies can be used to assign tasks optimally at test time. Experiments show emergent task-aware, multi-step coordination among agents, e.g., pressing a button to unlock a door, holding the door, and short-circuiting tasks.
Authors: Xu Liu, Yan Chen, Kan Ling, Yichi Zhu, Hengrun Zhang, Guisheng Fan, Huiqun Yu
Abstract: The widespread deployment of Large Language Models (LLMs) as public-facing web services and APIs has made their security a core concern for the web ecosystem. Jailbreak attacks, as one of the significant threats to LLMs, have recently attracted extensive research. In this paper, we reveal a jailbreak strategy which can effectively evade current defense strategies. It can extract valuable information from failed or partially successful attack attempts and contains self-evolution from attack interactions, resulting in sufficient strategy diversity and adaptability. Inspired by continuous learning and modular design principles, we propose ASTRA, a jailbreak framework that autonomously discovers, retrieves, and evolves attack strategies to achieve more efficient and adaptive attacks. To enable this autonomous evolution, we design a closed-loop "attack-evaluate-distill-reuse" core mechanism that not only generates attack prompts but also automatically distills and generalizes reusable attack strategies from every interaction. To systematically accumulate and apply this attack knowledge, we introduce a three-tier strategy library that categorizes strategies into Effective, Promising, and Ineffective based on their performance scores. The strategy library not only provides precise guidance for attack generation but also possesses exceptional extensibility and transferability. We conduct extensive experiments under a black-box setting, and the results show that ASTRA achieves an average Attack Success Rate (ASR) of 82.7%, significantly outperforming baselines.
Authors: Wongyu Kim, Hochang Lee, Sanghak Lee, Yoonsung Kim, Jaehyun Park
Abstract: Query augmentation makes queries more meaningful by appending further information to the queries to find relevant documents. Current studies have proposed Large Language Model (LLM)-based embedders, which learn representation for embedding and generation for query augmentation in a multi-task manner by leveraging the generative capabilities of LLM. During inference, these jointly trained embedders have conducted query augmentation followed by embedding, showing effective results. However, augmenting every query leads to substantial embedding latency and query augmentation can be detrimental to performance for some queries. Also, previous methods have not been explored in multimodal environments. To tackle these problems, we propose M-Solomon, a universal multimodal embedder that can adaptively determine when to augment queries. Our approach first divides the queries of the training datasets into two groups at the dataset level. One includes queries that require augmentation and the other includes queries that do not. Then, we introduces a synthesis process that generates appropriate augmentations for queries that require them by leveraging a powerful Multimodal LLM (MLLM). Next, we present adaptive query augmentation. Through this step, M-Solomon can conduct query augmentation only when necessary by learning to generate synthetic augmentations with the prefix /augment for queries that demand them and to generate the simple string /embed for others. Experimental results showed that M-Solomon not only surpassed the baseline without augmentation by a large margin but also outperformed the baseline that always used augmentation, providing much faster embedding latency.
Authors: Jean-Baptiste Courbot, Hugo Gangloff, Bruno Colicchio
Abstract: This work addresses the problem of efficient sampling of Markov random fields (MRF). The sampling of Potts or Ising MRF is most often based on Gibbs sampling, and is thus computationally expensive. We consider in this work how to circumvent this bottleneck through a link with Gaussian Markov Random fields. The latter can be sampled in several cost-effective ways, and we introduce a mapping from real-valued GMRF to discrete-valued MRF. The resulting new class of MRF benefits from a few theoretical properties that validate the new model. Numerical results show the drastic performance gain in terms of computational efficiency, as we sample at least 35x faster than Gibbs sampling using at least 37x less energy, all the while exhibiting empirical properties close to classical MRFs.
Authors: Aashray Reddy, Andrew Zagula, Nicholas Saban
Abstract: Large Language Models (LLMs) remain vulnerable to jailbreaking attacks where adversarial prompts elicit harmful outputs, yet most evaluations focus on single-turn interactions while real-world attacks unfold through adaptive multi-turn conversations. We present AutoAdv, a training-free framework for automated multi-turn jailbreaking that achieves up to 95% attack success rate on Llama-3.1-8B within six turns a 24 percent improvement over single turn baselines. AutoAdv uniquely combines three adaptive mechanisms: a pattern manager that learns from successful attacks to enhance future prompts, a temperature manager that dynamically adjusts sampling parameters based on failure modes, and a two-phase rewriting strategy that disguises harmful requests then iteratively refines them. Extensive evaluation across commercial and open-source models (GPT-4o-mini, Qwen3-235B, Mistral-7B) reveals persistent vulnerabilities in current safety mechanisms, with multi-turn attacks consistently outperforming single-turn approaches. These findings demonstrate that alignment strategies optimized for single-turn interactions fail to maintain robustness across extended conversations, highlighting an urgent need for multi-turn-aware defenses.
Authors: Leon Schwarzer, Matthias Zeller, Daniel Casado Herraez, Simon Dierl, Michael Heidingsfeld, Cyrill Stachniss
Abstract: Moving object segmentation is a crucial task for safe and reliable autonomous mobile systems like self-driving cars, improving the reliability and robustness of subsequent tasks like SLAM or path planning. While the segmentation of camera or LiDAR data is widely researched and achieves great results, it often introduces an increased latency by requiring the accumulation of temporal sequences to gain the necessary temporal context. Radar sensors overcome this problem with their ability to provide a direct measurement of a point's Doppler velocity, which can be exploited for single-scan moving object segmentation. However, radar point clouds are often sparse and noisy, making data annotation for use in supervised learning very tedious, time-consuming, and cost-intensive. To overcome this problem, we address the task of self-supervised moving object segmentation of sparse and noisy radar point clouds. We follow a two-step approach of contrastive self-supervised representation learning with subsequent supervised fine-tuning using limited amounts of annotated data. We propose a novel clustering-based contrastive loss function with cluster refinement based on dynamic points removal to pretrain the network to produce motion-aware representations of the radar data. Our method improves label efficiency after fine-tuning, effectively boosting state-of-the-art performance by self-supervised pretraining.
Authors: Yalda Zafari, Hongyi Pan, Gorkem Durak, Ulas Bagci, Essam A. Rashed, Mohamed Mabrok
Abstract: The development of clinically reliable artificial intelligence (AI) systems for mammography is hindered by profound heterogeneity in data quality, metadata standards, and population distributions across public datasets. This heterogeneity introduces dataset-specific biases that severely compromise the generalizability of the model, a fundamental barrier to clinical deployment. We present MammoClean, a public framework for standardization and bias quantification in mammography datasets. MammoClean standardizes case selection, image processing (including laterality and intensity correction), and unifies metadata into a consistent multi-view structure. We provide a comprehensive review of breast anatomy, imaging characteristics, and public mammography datasets to systematically identify key sources of bias. Applying MammoClean to three heterogeneous datasets (CBIS-DDSM, TOMPEI-CMMD, VinDr-Mammo), we quantify substantial distributional shifts in breast density and abnormality prevalence. Critically, we demonstrate the direct impact of data corruption: AI models trained on corrupted datasets exhibit significant performance degradation compared to their curated counterparts. By using MammoClean to identify and mitigate bias sources, researchers can construct unified multi-dataset training corpora that enable development of robust models with superior cross-domain generalization. MammoClean provides an essential, reproducible pipeline for bias-aware AI development in mammography, facilitating fairer comparisons and advancing the creation of safe, effective systems that perform equitably across diverse patient populations and clinical settings. The open-source code is publicly available from: https://github.com/Minds-R-Lab/MammoClean.
Authors: Christoph Hertrich, Stefan Kober, Georg Loho
Abstract: We prove that there exist uniform $(+,\times,/)$-circuits of size $O(n^3)$ to compute the basis generating polynomial of regular matroids on $n$ elements. By tropicalization, this implies that there exist uniform $(\max,+,-)$-circuits and ReLU neural networks of the same size for weighted basis maximization of regular matroids. As a consequence in linear programming theory, we obtain a first example where taking the difference of two extended formulations can be more efficient than the best known individual extended formulation of size $O(n^6)$ by Aprile and Fiorini. Such differences have recently been introduced as virtual extended formulations. The proof of our main result relies on a fine-tuned version of Seymour's decomposition of regular matroids which allows us to identify and maintain graphic substructures to which we can apply a local version of the star-mesh transformation.
Authors: Junghee Pyeon, Davide Cacciarelli, Kamran Paynabar
Abstract: Concept drift and label scarcity are two critical challenges limiting the robustness of predictive models in dynamic industrial environments. Existing drift detection methods often assume global shifts and rely on dense supervision, making them ill-suited for regression tasks with local drifts and limited labels. This paper proposes an adaptive sampling framework that combines residual-based exploration and exploitation with EWMA monitoring to efficiently detect local concept drift under labeling budget constraints. Empirical results on synthetic benchmarks and a case study on electricity market demonstrate superior performance in label efficiency and drift detection accuracy.
Authors: Weiming Feng, Xiongxin Yang, Yixiao Yu, Yiyao Zhang
Abstract: We study the problem of learning a $n$-variables $k$-CNF formula $\Phi$ from its i.i.d. uniform random solutions, which is equivalent to learning a Boolean Markov random field (MRF) with $k$-wise hard constraints. Revisiting Valiant's algorithm (Commun. ACM'84), we show that it can exactly learn (1) $k$-CNFs with bounded clause intersection size under Lov\'asz local lemma type conditions, from $O(\log n)$ samples; and (2) random $k$-CNFs near the satisfiability threshold, from $\widetilde{O}(n^{\exp(-\sqrt{k})})$ samples. These results significantly improve the previous $O(n^k)$ sample complexity. We further establish new information-theoretic lower bounds on sample complexity for both exact and approximate learning from i.i.d. uniform random solutions.
Authors: Marc Schneider, Walter Fichter
Abstract: This paper presents a novel approach to many-vs-many missile guidance using virtual targets (VTs) generated by a Normalizing Flows-based trajectory predictor. Rather than assigning n interceptors directly to m physical targets through conventional weapon target assignment algorithms, we propose a centralized strategy that constructs n VT trajectories representing probabilistic predictions of maneuvering target behavior. Each interceptor is guided toward its assigned VT using Zero-Effort-Miss guidance during midcourse flight, transitioning to Proportional Navigation guidance for terminal interception. This approach treats many-vs-many engagements as many-vs-distribution scenarios, exploiting numerical superiority (n > m) by distributing interceptors across diverse trajectory hypotheses rather than pursuing identical deterministic predictions. Monte Carlo simulations across various target-interceptor configurations (1-6 targets, 1-8 interceptors) demonstrate that the VT method matches or exceeds baseline straight-line prediction performance by 0-4.1% when n = m, with improvements increasing to 5.8-14.4% when n > m. The results confirm that probabilistic VTs enable effective exploitation of numerical superiority, significantly increasing interception probability in many-vs-many scenarios.
Authors: Jorge Pellejero, Luis A. Hern\'andez G\'omez, Luis Mendo Tom\'as, Zoraida Frias Barroso
Abstract: Agentic AI represents a new paradigm for automating complex systems by using Large AI Models (LAMs) to provide human-level cognitive abilities with multimodal perception, planning, memory, and reasoning capabilities. This will lead to a new generation of AI systems that autonomously decompose goals, retain context over time, learn continuously, operate across tools and environments, and adapt dynamically. The complexity of 5G and upcoming 6G networks renders manual optimization ineffective, pointing to Agentic AI as a method for automating decisions in dynamic RAN environments. However, despite its rapid advances, there is no established framework outlining the foundational components and operational principles of Agentic AI systems nor a universally accepted definition. This paper contributes to ongoing research on Agentic AI in 5G and 6G networks by outlining its core concepts and then proposing a practical use case that applies Agentic principles to RAN optimization. We first introduce Agentic AI, tracing its evolution from classical agents and discussing the progress from workflows and simple AI agents to Agentic AI. Core design patterns-reflection, planning, tool use, and multi-agent collaboration-are then described to illustrate how intelligent behaviors are orchestrated. These theorical concepts are grounded in the context of mobile networks, with a focus on RAN management and optimization. A practical 5G RAN case study shows how time-series analytics and LAM-driven agents collaborate for KPI-based autonomous decision-making.
Authors: Ali Farki, Elaheh Moradi, Deepika Koundal, Jussi Tohka
Abstract: Predicting future brain state from a baseline magnetic resonance image (MRI) is a central challenge in neuroimaging and has important implications for studying neurodegenerative diseases such as Alzheimer's disease (AD). Most existing approaches predict future cognitive scores or clinical outcomes, such as conversion from mild cognitive impairment to dementia. Instead, here we investigate longitudinal MRI image-to-image prediction that forecasts a participant's entire brain MRI several years into the future, intrinsically modeling complex, spatially distributed neurodegenerative patterns. We implement and evaluate five deep learning architectures (UNet, U2-Net, UNETR, Time-Embedding UNet, and ODE-UNet) on two longitudinal cohorts (ADNI and AIBL). Predicted follow-up MRIs are directly compared with the actual follow-up scans using metrics that capture global similarity and local differences. The best performing models achieve high-fidelity predictions, and all models generalize well to an independent external dataset, demonstrating robust cross-cohort performance. Our results indicate that deep learning can reliably predict participant-specific brain MRI at the voxel level, offering new opportunities for individualized prognosis.
Authors: Anastasios T. Sotiropoulos, Stavros Tsimpoukis, Dimitrios Tyrovolas, Sotiris Ioannidis, George K. Karagiannidis, Christos K. Liaskos
Abstract: The pursuit of immersive and structurally aware multimedia experiences has intensified interest in sensing modalities that reconstruct objects beyond the limits of visible light. Conventional optical pipelines degrade under occlusion or low illumination, motivating the use of radio-frequency (RF) sensing, whose electromagnetic waves penetrate materials and encode both geometric and compositional information. Yet, uncontrolled multipath propagation restricts reconstruction accuracy. Recent advances in Programmable Wireless Environments (PWEs) mitigate this limitation by enabling software-defined manipulation of propagation through Reconfigurable Intelligent Surfaces (RISs), thereby providing controllable illumination diversity. Building on this capability, this work introduces a PWE-driven RF framework for three-dimensional object reconstruction using material-aware spherical primitives. The proposed approach combines RIS-enabled field synthesis with a Detection Transformer (DETR) that infers spatial and material parameters directly from extracted RF features. Simulation results confirm the framework's ability to approximate object geometries and classify material composition with an overall accuracy of 79.35%, marking an initial step toward programmable and physically grounded RF-based 3D object composition visualization.
Authors: Daichi Nagai, Ryugo Morita, Shunsuke Kitada, Hitoshi Iyatomi
Abstract: Despite the remarkable success of text-to-image diffusion models, their output of a single, flattened image remains a critical bottleneck for professional applications requiring layer-wise control. Existing solutions either rely on fine-tuning with large, inaccessible datasets or are training-free yet limited to generating isolated foreground elements, failing to produce a complete and coherent scene. To address this, we introduce the Training-free Noise Transplantation and Cultivation Diffusion Model (TAUE), a novel framework for zero-shot, layer-wise image generation. Our core technique, Noise Transplantation and Cultivation (NTC), extracts intermediate latent representations from both foreground and composite generation processes, transplanting them into the initial noise for subsequent layers. This ensures semantic and structural coherence across foreground, background, and composite layers, enabling consistent, multi-layered outputs without requiring fine-tuning or auxiliary datasets. Extensive experiments show that our training-free method achieves performance comparable to fine-tuned methods, enhancing layer-wise consistency while maintaining high image quality and fidelity. TAUE not only eliminates costly training and dataset requirements but also unlocks novel downstream applications, such as complex compositional editing, paving the way for more accessible and controllable generative workflows.
Authors: Mark Bl\"umel, Andreas C. Schneider, Valentin Neuhaus, David A. Ehrlich, Marcel Graetz, Michael Wibral, Abdullah Makkeh, Viola Priesemann
Abstract: Associative memory, traditionally modeled by Hopfield networks, enables the retrieval of previously stored patterns from partial or noisy cues. Yet, the local computational principles which are required to enable this function remain incompletely understood. To formally characterize the local information processing in such systems, we employ a recent extension of information theory - Partial Information Decomposition (PID). PID decomposes the contribution of different inputs to an output into unique information from each input, redundant information across inputs, and synergistic information that emerges from combining different inputs. Applying this framework to individual neurons in classical Hopfield networks we find that below the memory capacity, the information in a neuron's activity is characterized by high redundancy between the external pattern input and the internal recurrent input, while synergy and unique information are close to zero until the memory capacity is surpassed and performance drops steeply. Inspired by this observation, we use redundancy as an information-theoretic learning goal, which is directly optimized for each neuron, dramatically increasing the network's memory capacity to 1.59, a more than tenfold improvement over the 0.14 capacity of classical Hopfield networks and even outperforming recent state-of-the-art implementations of Hopfield networks. Ultimately, this work establishes redundancy maximization as a new design principle for associative memories and opens pathways for new associative memory models based on information-theoretic goals.
Authors: Roy Rinberg, Adam Karvonen, Alex Hoover, Daniel Reuter, Keri Warr
Abstract: As large AI models become increasingly valuable assets, the risk of model weight exfiltration from inference servers grows accordingly. An attacker controlling an inference server may exfiltrate model weights by hiding them within ordinary model outputs, a strategy known as steganography. This work investigates how to verify model responses to defend against such attacks and, more broadly, to detect anomalous or buggy behavior during inference. We formalize model exfiltration as a security game, propose a verification framework that can provably mitigate steganographic exfiltration, and specify the trust assumptions associated with our scheme. To enable verification, we characterize valid sources of non-determinism in large language model inference and introduce two practical estimators for them. We evaluate our detection framework on several open-weight models ranging from 3B to 30B parameters. On MOE-Qwen-30B, our detector reduces exfiltratable information to <0.5% with false-positive rate of 0.01%, corresponding to a >200x slowdown for adversaries. Overall, this work further establishes a foundation for defending against model weight exfiltration and demonstrates that strong protection can be achieved with minimal additional cost to inference providers.
Authors: Xinliang Liu, Tong Mao, Jinchao Xu
Abstract: We present an estimation of the condition numbers of the \emph{mass} and \emph{stiffness} matrices arising from shallow ReLU$^k$ neural networks defined on the unit sphere~$\mathbb{S}^d$. In particular, when $\{\theta_j^*\}_{j=1}^n \subset \mathbb{S}^d$ is \emph{antipodally quasi-uniform}, the condition number is sharp. Indeed, in this case, we obtain sharp asymptotic estimates for the full spectrum of eigenvalues and characterize the structure of the corresponding eigenspaces, showing that the smallest eigenvalues are associated with an eigenbasis of low-degree polynomials while the largest eigenvalues are linked to high-degree polynomials. This spectral analysis establishes a precise correspondence between the approximation power of the network and its numerical stability.
Authors: Xiumei Deng, Zehui Xiong, Binbin Chen, Dong In Kim, Merouane Debbah, H. Vincent Poor
Abstract: Large language models (LLMs) are proliferating rapidly at the edge, delivering intelligent capabilities across diverse application scenarios. However, their practical deployment in collaborative scenarios confronts fundamental challenges: privacy vulnerabilities, communication overhead, and computational bottlenecks. To address these, we propose Federated Attention (FedAttn), which integrates the federated paradigm into the self-attention mechanism, creating a new distributed LLM inference framework that simultaneously achieves privacy protection, communication efficiency, and computational efficiency. FedAttn enables participants to perform local self-attention over their own token representations while periodically exchanging and aggregating Key-Value (KV) matrices across multiple Transformer blocks, collaboratively generating LLM responses without exposing private prompts. Further, we identify a structural duality between contextual representation refinement in FedAttn and parameter optimization in FL across private data, local computation, and global aggregation. This key insight provides a principled foundation for systematically porting federated optimization techniques to collaborative LLM inference. Building on this framework, we theoretically analyze how local self-attention computation within participants and heterogeneous token relevance among participants shape error propagation dynamics across Transformer blocks. Moreover, we characterize the fundamental trade-off between response quality and communication/computation efficiency, which is governed by the synchronization interval and the number of participants. Experimental results validate our theoretical analysis, and reveal significant optimization opportunities through sparse attention and adaptive KV aggregation, highlighting FedAttn's potential to deliver scalability and efficiency in real-world edge deployments.
Authors: Adam Umra, Aya M. Ahmed, Aydin Sezgin
Abstract: This paper proposes a reinforcement learning (RL)-aided cognitive framework for massive MIMO-based integrated sensing and communication (ISAC) systems employing a uniform planar array (UPA). The focus is on enhancing radar sensing performance in environments with unknown and dynamic disturbance characteristics. A Wald-type detector is employed for robust target detection under non-Gaussian clutter, while a SARSA-based RL algorithm enables adaptive estimation of target positions without prior environmental knowledge. Based on the RL-derived sensing information, a joint waveform optimization strategy is formulated to balance radar sensing accuracy and downlink communication throughput. The resulting design provides an adaptive trade-off between detection performance and achievable sum rate through an analytically derived closed-form solution. Monte Carlo simulations demonstrate that the proposed cognitive ISAC framework achieves significantly improved detection probability compared to orthogonal and non-learning adaptive baselines, while maintaining competitive communication performance. These results underline the potential of RL-assisted sensing for robust and spectrum-efficient ISAC in next-generation wireless networks.
Authors: Mohammadsajad Alipour, Mohammad Mohammadi Amiri
Abstract: Large language models (LLMs) are increasingly prevalent across diverse applications. However, their enormous size limits storage and processing capabilities to a few well-resourced stakeholders. As a result, most applications rely on pre-trained LLMs, fine-tuned for specific tasks. However, even storing the fine-tuned versions of these models remains a significant challenge due to the wide range of tasks they address. Recently, studies show that fine-tuning these models primarily affects a small fraction of parameters, highlighting the need for more efficient storage of fine-tuned models. This paper focuses on efficient storage of parameter updates in pre-trained models after fine-tuning. To address this challenge, we leverage the observation that fine-tuning updates are both low-rank and sparse, which can be utilized for storage efficiency. However, using only low-rank approximation or sparsification may discard critical singular components that enhance model expressivity. We first observe that given the same memory budget, sparsified low-rank approximations with larger ranks outperform standard low-rank approximations with smaller ranks. Building on this, we propose our method, optimal singular damage, that selectively sparsifies low-rank approximated updates by leveraging the interleaved importance of singular vectors, ensuring that the most impactful components are retained. We demonstrate through extensive experiments that our proposed methods lead to significant storage efficiency and superior accuracy within the same memory budget compared to employing the low-rank approximation or sparsification individually.
Authors: Deyao Chen, Fran\c{c}ois Cl\'ement, Carola Doerr, Nathan Kirk
Abstract: Kernel discrepancies are a powerful tool for analyzing worst-case errors in quasi-Monte Carlo (QMC) methods. Building on recent advances in optimizing such discrepancy measures, we extend the subset selection problem to the setting of kernel discrepancies, selecting an m-element subset from a large population of size $n \gg m$. We introduce a novel subset selection algorithm applicable to general kernel discrepancies to efficiently generate low-discrepancy samples from both the uniform distribution on the unit hypercube, the traditional setting of classical QMC, and from more general distributions $F$ with known density functions by employing the kernel Stein discrepancy. We also explore the relationship between the classical $L_2$ star discrepancy and its $L_\infty$ counterpart.
Authors: Farhad Rezazadeh, Hatim Chergui, Merouane Debbah, Houbing Song, Dusit Niyato, Lingjia Liu
Abstract: We argue that sixth-generation (6G) intelligence is not fluent token prediction but the capacity to imagine and choose -- to simulate future scenarios, weigh trade-offs, and act with calibrated uncertainty. We reframe open radio access network (O-RAN) near-real-time (Near-RT) control via counterfactual dynamics and a world modeling (WM) paradigm that learns an action-conditioned generative state space. This enables quantitative "what-if" forecasting beyond large language models (LLMs) as the primary modeling primitive. Actions such as physical resource blocks (PRBs) are treated as first-class control inputs in a causal world model, and both aleatoric and epistemic uncertainty are modeled for prediction and what-if analysis. An agentic, model predictive control (MPC)-based cross-entropy method (CEM) planner operates over short horizons, using prior-mean rollouts within data-driven PRB bounds to maximize a deterministic reward. The model couples multi-scale structured state-space mixtures (MS3M) with a compact stochastic latent to form WM-MS3M, summarizing key performance indicators (KPIs) histories and predicting next-step KPIs under hypothetical PRB sequences. On realistic O-RAN traces, WM-MS3M cuts mean absolute error (MAE) by 1.69% versus MS3M with 32% fewer parameters and similar latency, and achieves 35-80% lower root mean squared error (RMSE) than attention/hybrid baselines with 2.3-4.1x faster inference, enabling rare-event simulation and offline policy screening.
Authors: Zebin Wang, Ziming Gan, Weijing Tang, Zongqi Xia, Tianrun Cai, Tianxi Cai, Junwei Lu
Abstract: Classical probabilistic graphical models face fundamental challenges in modern data environments, which are characterized by high dimensionality, source heterogeneity, and stringent data-sharing constraints. In this work, we revisit the Ising model, a well-established member of the Markov Random Field (MRF) family, and develop a distributed framework that enables scalable and privacy-preserving representation learning from large-scale binary data with inherent low-rank structure. Our approach optimizes a non-convex surrogate loss function via bi-factored gradient descent, offering substantial computational and communication advantages over conventional convex approaches. We evaluate our algorithm on multi-institutional electronic health record (EHR) datasets from 58,248 patients across the University of Pittsburgh Medical Center (UPMC) and Mass General Brigham (MGB), demonstrating superior performance in global representation learning and downstream clinical tasks, including relationship detection, patient phenotyping, and patient clustering. These results highlight a broader potential for statistical inference in federated, high-dimensional settings while addressing the practical challenges of data complexity and multi-institutional integration.
Authors: Mohamed Bouadi, Pratinav Seth, Aditya Tanna, Vinay Kumar Sankarapu
Abstract: Tabular data remain the predominant format for real-world applications. Yet, developing effective neural models for tabular data remains challenging due to heterogeneous feature types and complex interactions occurring at multiple scales. Recent advances in tabular in-context learning (ICL), such as TabPFN and TabICL, have achieved state-of-the-art performance comparable to gradient-boosted trees (GBTs) without task-specific fine-tuning. However, current architectures exhibit key limitations: (1) single-scale feature processing that overlooks hierarchical dependencies, (2) dense attention with quadratic scaling in table width, and (3) strictly sequential component processing that prevents iterative representation refinement and cross-component communication. To address these challenges, we introduce Orion-MSP, a tabular ICL architecture featuring three key innovations: (1) multi-scale processing to capture hierarchical feature interactions; (2) block-sparse attention combining windowed, global, and random patterns for scalable efficiency and long-range connectivity; and (3) a Perceiver-style memory enabling safe bidirectional information flow across components. Across diverse benchmarks, Orion-MSP matches or surpasses state-of-the-art performance while scaling effectively to high-dimensional tables, establishing a new standard for efficient tabular in-context learning. The model is publicly available at https://github.com/Lexsi-Labs/Orion-MSP .
Authors: Dan Garber
Abstract: We develop new accelerated first-order algorithms in the Frank-Wolfe (FW) family for minimizing smooth convex functions over compact convex sets, with a focus on two prominent constraint classes: (1) polytopes and (2) matrix domains given by the spectrahedron and the unit nuclear-norm ball. A key technical ingredient is a complementarity condition that captures solution sparsity -- face dimension for polytopes and rank for matrices. We present two algorithms: (1) a purely linear optimization oracle (LOO) method for polytopes that has optimal worst-case first-order (FO) oracle complexity and, aside of a finite \emph{burn-in} phase and up to a logarithmic factor, has LOO complexity that scales with $r/\sqrt{\epsilon}$, where $\epsilon$ is the target accuracy and $r$ is the solution sparsity $r$ (independently of the ambient dimension), and (2) a hybrid scheme that combines FW with a sparse projection oracle (e.g., low-rank SVDs for matrix domains with low-rank solutions), which also has optimal FO oracle complexity, and after a finite burn-in phase, only requires $O(1/\sqrt{\epsilon})$ sparse projections and LOO calls (independently of both the ambient dimension and the rank of optimal solutions). Our results close a gap on how to accelerate recent advancements in linearly-converging FW algorithms for strongly convex optimization, without paying the price of the dimension.
Authors: Yanjie Ze, Siheng Zhao, Weizhuo Wang, Angjoo Kanazawa, Rocky Duan, Pieter Abbeel, Guanya Shi, Jiajun Wu, C. Karen Liu
Abstract: Large-scale data has driven breakthroughs in robotics, from language models to vision-language-action models in bimanual manipulation. However, humanoid robotics lacks equally effective data collection frameworks. Existing humanoid teleoperation systems either use decoupled control or depend on expensive motion capture setups. We introduce TWIST2, a portable, mocap-free humanoid teleoperation and data collection system that preserves full whole-body control while advancing scalability. Our system leverages PICO4U VR for obtaining real-time whole-body human motions, with a custom 2-DoF robot neck (cost around $250) for egocentric vision, enabling holistic human-to-humanoid control. We demonstrate long-horizon dexterous and mobile humanoid skills and we can collect 100 demonstrations in 15 minutes with an almost 100% success rate. Building on this pipeline, we propose a hierarchical visuomotor policy framework that autonomously controls the full humanoid body based on egocentric vision. Our visuomotor policy successfully demonstrates whole-body dexterous manipulation and dynamic kicking tasks. The entire system is fully reproducible and open-sourced at https://yanjieze.com/TWIST2 . Our collected dataset is also open-sourced at https://twist-data.github.io .
URLs: https://yanjieze.com/TWIST2, https://twist-data.github.io
Authors: Huawei Lin, Yunzhi Shi, Tong Geng, Weijie Zhao, Wei Wang, Ravender Pal Singh
Abstract: Multimodal large language models (MLLMs) have shown strong capabilities but remain limited to fixed modality pairs and require costly fine-tuning with large aligned datasets. Building fully omni-capable models that can integrate text, images, audio, and video remains impractical and lacks robust reasoning support. In this paper, we propose an Agent-Omni framework that coordinates existing foundation models through a master-agent system, enabling flexible multimodal reasoning without retraining. The master agent interprets user intent, delegates subtasks to modality-specific agents, and integrates their outputs into coherent responses. Extensive experiments across text, image, audio, video, and omni benchmarks show that Agent-Omni consistently achieves state-of-the-art performance, particularly on tasks requiring complex cross-modal reasoning. Its agent-based design enables seamless integration of specialized foundation models, ensuring adaptability to diverse inputs while maintaining transparency and interpretability. In addition, the framework is modular and easily extensible, allowing future improvements as stronger models become available. %We release an open-source implementation to support continued research on scalable and reliable omni-modal reasoning.
Authors: Yuya Sasaki
Abstract: The number of graph neural network (GNN) architectures has increased rapidly due to the growing adoption of graph analysis. Although we use GNNs in wide application scenarios, it is a laborious task to design/select optimal GNN architectures in diverse graphs. To reduce human efforts, graph neural architecture search (Graph NAS) has been used to search for a sub-optimal GNN architecture that combines existing components. However, existing Graph NAS methods lack explainability to understand the reasons why the model architecture is selected because they use complex search space and neural models to select architecture. Therefore, we propose an explainable Graph NAS method, called ExGNAS, which consists of (i) a simple search space that can adapt to various graphs and (ii) a search algorithm with Monte-Carlo tree that makes the decision process explainable. The combination of our search space and algorithm achieves finding accurate GNN models and the important functions within the search space. We comprehensively evaluate ExGNAS compared with four state-of-the-art Graph NAS methods in twelve graphs. Our experimental results show that ExGNAS achieves high average accuracy and efficiency; improving accuracy up to 26.1% and reducing run time up to 88%. Furthermore, we show the effectiveness of explainability by questionnaire-based user study and architecture analysis.
Authors: Mingyang Yi, Bohan Wang
Abstract: Recently, optimization on the Riemannian manifold have provided valuable insights to the optimization community. In this regard, extending these methods to to the Wasserstein space is of particular interest, since optimization on Wasserstein space is closely connected to practical sampling processes. Generally, the standard (continuous) optimization method on Wasserstein space is Riemannian gradient flow (i.e., Langevin dynamics when minimizing KL divergence). In this paper, we aim to enrich the family of continuous optimization methods in the Wasserstein space, by extending the gradient flow on it into the stochastic gradient descent (SGD) flow and stochastic variance reduction gradient (SVRG) flow. By leveraging the property of Wasserstein space, we construct stochastic differential equations (SDEs) to approximate the corresponding discrete Euclidean dynamics of the desired Riemannian stochastic methods. Then, we obtain the flows in Wasserstein space by Fokker-Planck equation. Finally, we establish convergence rates of the proposed stochastic flows, which align with those known in the Euclidean setting.
Authors: Hongjoon Ahn, Jinu Hyeon, Youngmin Oh, Bosun Hwang, Taesup Moon
Abstract: We argue that the negative transfer problem occurring when the new task to learn arrives is an important problem that needs not be overlooked when developing effective Continual Reinforcement Learning (CRL) algorithms. Through comprehensive experimental validation, we demonstrate that such issue frequently exists in CRL and cannot be effectively addressed by several recent work on either mitigating plasticity loss of RL agents or enhancing the positive transfer in CRL scenario. To that end, we develop Reset & Distill (R&D), a simple yet highly effective baseline method, to overcome the negative transfer problem in CRL. R&D combines a strategy of resetting the agent's online actor and critic networks to learn a new task and an offline learning step for distilling the knowledge from the online actor and previous expert's action probabilities. We carried out extensive experiments on long sequence of Meta World tasks and show that our simple baseline method consistently outperforms recent approaches, achieving significantly higher success rates across a range of tasks. Our findings highlight the importance of considering negative transfer in CRL and emphasize the need for robust strategies like R&D to mitigate its detrimental effects.
Authors: Cheng Lu, Jiusun Zeng, Yu Xia, Jinhui Cai, Shihua Luo
Abstract: Shapley value is a widely used tool in explainable artificial intelligence (XAI), as it provides a principled way to attribute contributions of input features to model outputs. However, estimation of Shapley value requires capturing conditional dependencies among all feature combinations, which poses significant challenges in complex data environments. In this article, EmSHAP (Energy-based model for Shapley value estimation), an accurate Shapley value estimation method, is proposed to estimate the expectation of Shapley contribution function under the arbitrary subset of features given the rest. By utilizing the ability of energy-based model (EBM) to model complex distributions, EmSHAP provides an effective solution for estimating the required conditional probabilities. To further improve estimation accuracy, a GRU (Gated Recurrent Unit)-coupled partition function estimation method is introduced. The GRU network captures long-term dependencies with a lightweight parameterization and maps input features into a latent space to mitigate the influence of feature ordering. Additionally, a dynamic masking mechanism is incorporated to further enhance the robustness and accuracy by progressively increasing the masking rate. Theoretical analysis on the error bound as well as application to four case studies verified the higher accuracy and better scalability of EmSHAP in contrast to competitive methods.
Authors: Lukas Aichberger, Kajetan Schweighofer, Mykyta Ielanskyi, Sepp Hochreiter
Abstract: Large language models (LLMs) can suffer from hallucinations when generating text. These hallucinations impede various applications in society and industry by making LLMs untrustworthy. Current LLMs generate text in an autoregressive fashion by predicting and appending text tokens. When an LLM is uncertain about the semantic meaning of the next tokens to generate, it is likely to start hallucinating. Thus, it has been suggested that predictive uncertainty is one of the main causes of hallucinations. We introduce Semantically Diverse Language Generation (SDLG) to quantify predictive uncertainty in LLMs. SDLG steers the LLM to generate semantically diverse yet likely alternatives for an initially generated text. This approach provides a precise measure of aleatoric semantic uncertainty, detecting whether the initial text is likely to be hallucinated. Experiments on question-answering tasks demonstrate that SDLG consistently outperforms existing methods while being the most computationally efficient, setting a new standard for uncertainty estimation in LLMs.
Authors: Shyam Venkatasubramanian, Bosung Kang, Ali Pezeshki, Muralidhar Rangaswamy, Vahid Tarokh
Abstract: We present a large-scale dataset called RASPNet for radar adaptive signal processing (RASP) applications to support the development of data-driven models within the adaptive radar community. RASPNet exceeds 16 TB in size and comprises 100 realistic scenarios compiled over a variety of topographies and land types across the contiguous United States. For each scenario, RASPNet comprises 10,000 clutter realizations from an airborne radar setting, which can be used to benchmark radar and complex-valued learning algorithms. RASPNet intends to fill a prominent gap in the availability of a large-scale, realistic dataset that standardizes the evaluation of RASP techniques and complex-valued neural networks. We outline its construction, organization, and several applications, including a transfer learning example to demonstrate how RASPNet can be used for real-world adaptive radar scenarios.
Authors: Jingchao Gao, Ziqing Lu, Raghu Mudumbai, Xiaodong Wu, Jirong Yi, Myung Cho, Catherine Xu, Hui Xie, Weiyu Xu
Abstract: In this paper, we uniquely study the adversarial robustness of deep neural networks (NN) for classification tasks against that of optimal classifiers. We look at the smallest magnitude of possible additive perturbations that can change a classifier's output. We provide a matrix-theoretic explanation of the adversarial fragility of deep neural networks for classification. In particular, our theoretical results show that a neural network's adversarial robustness can degrade as the input dimension $d$ increases. Analytically, we show that neural networks' adversarial robustness can be only $1/\sqrt{d}$ of the best possible adversarial robustness of optimal classifiers. Our theories match remarkably well with numerical experiments of practically trained NN, including NN for ImageNet images. The matrix-theoretic explanation is consistent with an earlier information-theoretic feature-compression-based explanation for the adversarial fragility of neural networks.
Authors: Lisi Qarkaxhija, Anatol E. Wegner, Ingo Scholtes
Abstract: Message passing neural networks (MPNNs) operate on graphs by exchanging information between neigbouring nodes. MPNNs have been successfully applied to various node-, edge-, and graph-level tasks in areas like molecular science, computer vision, natural language processing, and combinatorial optimization. However, most MPNNs require training on large amounts of labeled data, which can be costly and time-consuming. In this work, we explore the use of various untrained message passing layers in graph neural networks, i.e. variants of popular message passing architecture where we remove all trainable parameters that are used to transform node features in the message passing step. Focusing on link prediction, we find that untrained message passing layers can lead to competitive and even superior performance compared to fully trained MPNNs, especially in the presence of high-dimensional features. We provide a theoretical analysis of untrained message passing by relating the inner products of features implicitly produced by untrained message passing layers to path-based topological node similarity measures. As such, untrained message passing architectures can be viewed as a highly efficient and interpretable approach to link prediction.
Authors: Song Hao, Wentao Fu, Xuanze Chen, Chengxiang Jin, Jiajun Zhou, Shanqing Yu, Qi Xuan
Abstract: Traditional anomalous traffic detection methods are based on single-view analysis, which has obvious limitations in dealing with complex attacks and encrypted communications. In this regard, we propose a Multi-view Feature Fusion (MuFF) method for network anomaly traffic detection. MuFF models the temporal and interactive relationships of packets in network traffic based on the temporal and interactive viewpoints respectively. It learns temporal and interactive features. These features are then fused from different perspectives for anomaly traffic detection. Extensive experiments on six real traffic datasets show that MuFF has excellent performance in network anomalous traffic detection, which makes up for the shortcomings of detection under a single perspective.
Authors: Muhammad Umar, Andras Lakatos, Muhammad Asif, Arif Mahmood
Abstract: Single-cell RNA sequencing (scRNA-seq) enables the study of cellular diversity at single cell level. It provides a global view of cell-type specification during the onset of biological mechanisms such as developmental processes and human organogenesis. Various statistical, machine and deep learning-based methods have been proposed for cell-type classification. Most of the methods utilizes unsupervised lower dimensional projections obtained from for a large reference data. In this work, we proposed a reference-based method for cell type classification, called EnProCell. The EnProCell, first, computes lower dimensional projections that capture both the high variance and class separability through an ensemble of principle component analysis and multiple discriminant analysis. In the second phase, EnProCell trains a deep neural network on the lower dimensional representation of data to classify cell types. The proposed method outperformed the existing state-of-the-art methods when tested on four different data sets produced from different single-cell sequencing technologies. The EnProCell showed higher accuracy (98.91) and F1 score (98.64) than other methods for predicting reference from reference datasets. Similarly, EnProCell also showed better performance than existing methods in predicting cell types for data with unknown cell types (query) from reference datasets (accuracy:99.52; F1 score: 99.07). In addition to improved performance, the proposed methodology is simple and does not require more computational resources and time. the EnProCell is available at https://github.com/umar1196/EnProCell.
Authors: Shuchang Yan, Haoran Sun
Abstract: In our prior work, we investigated the minimum fuel consumption of a hybrid electric vehicle (HEV) under a state-of-charge (SOC) balance constraint, assuming perfect SOC measurements and accurate reference speed profiles. The constrained optimal fuel consumption (COFC) problem was addressed using a constrained reinforcement learning (CRL) framework. However, in real-world scenarios, SOC readings are often corrupted by sensor noise, and reference speeds may deviate from actual driving conditions. To account for these imperfections, this study reformulates the COFC problem by explicitly incorporating observational noise in both SOC and reference speed. We adopt a robust CRL approach, where the noise is modeled as a uniform distribution, and employ a structured training procedure to ensure stability. The proposed method is evaluated through simulations on the Toyota Prius hybrid system (THS), using both the New European Driving Cycle (NEDC) and the Worldwide Harmonized Light Vehicles Test Cycle (WLTC). Results show that fuel consumption and SOC constraint satisfaction remain robust across varying noise levels. Furthermore, the analysis reveals that observational noise in SOC and speed can impact fuel consumption to different extents. To the best of our knowledge, this is the first study to explicitly examine how observational noise -- commonly encountered in dynamometer testing and predictive energy control (PEC) applications -- affects constrained optimal fuel consumption in HEVs.
Authors: Flavio Corradini, Flavio Gerosa, Marco Gori, Carlo Lucheroni, Marco Piangerelli, Martina Zannotti
Abstract: In recent years, spatio-temporal graph neural networks (GNNs) have attracted considerable interest in the field of time series analysis, due to their ability to capture, at once, dependencies among variables and across time points. The objective of this systematic literature review is hence to provide a comprehensive overview of the various modeling approaches and application domains of GNNs for time series classification and forecasting. A database search was conducted, and 366 papers were selected for a detailed examination of the current state-of-the-art in the field. This examination is intended to offer to the reader a comprehensive review of proposed models, links to related source code, available datasets, benchmark models, and fitting results. All this information is hoped to assist researchers in their studies. To the best of our knowledge, this is the first and broadest systematic literature review presenting a detailed comparison of results from current spatio-temporal GNN models applied to different domains. In its final part, this review discusses current limitations and challenges in the application of spatio-temporal GNNs, such as comparability, reproducibility, explainability, poor information capacity, and scalability. This paper is complemented by a GitHub repository at https://github.com/FlaGer99/SLR-Spatio-Temporal-GNN.git providing additional interactive tools to further explore the presented findings.
URLs: https://github.com/FlaGer99/SLR-Spatio-Temporal-GNN.git
Authors: Yichen Wang, Qing Yu, Yancun Song
Abstract: Bike-sharing is an environmentally friendly shared mobility mode, but its self-loop phenomenon, where bikes are returned to the same station after several time usage, significantly impacts equity in accessing its services. Therefore, this study conducts a multiscale analysis with a spatial autoregressive model and double machine learning framework to assess socioeconomic features and geospatial location's impact on the self-loop phenomenon at metro stations and street scales. The results reveal that bike-sharing self-loop intensity exhibits significant spatial lag effect at street scale and is positively associated with residential land use. Marginal treatment effects of residential land use is higher on streets with middle-aged residents, high fixed employment, and low car ownership. The multimodal public transit condition reveals significant positive marginal treatment effects at both scales. To enhance bike-sharing cooperation, we advocate augmenting bicycle availability in areas with high metro usage and low bus coverage, alongside implementing adaptable redistribution strategies.
Authors: Yinsicheng Jiang, Yao Fu, Yeqi Huang, Ping Nie, Zhan Lu, Leyang Xue, Congjie He, Man-Kit Sit, Jilong Xue, Li Dong, Ziming Miao, Dayou Du, Tairan Xu, Kai Zou, Edoardo Ponti, Luo Mai
Abstract: The sparse Mixture-of-Experts (MoE) architecture is increasingly favored for scaling Large Language Models (LLMs) efficiently, but it depends on heterogeneous compute and memory resources. These factors jointly affect system Cost, Accuracy, and Performance (CAP), making trade-offs inevitable. Existing benchmarks often fail to capture these trade-offs accurately, complicating practical deployment decisions. To address this, we introduce MoE-CAP, a benchmark specifically designed for MoE systems. Our analysis reveals that achieving an optimal balance across CAP is difficult with current hardware; MoE systems typically optimize two of the three dimensions at the expense of the third-a dynamic we term the MoE-CAP trade-off. To visualize this, we propose the CAP Radar Diagram. We further introduce sparsity-aware performance metrics-Sparse Memory Bandwidth Utilization (S-MBU) and Sparse Model FLOPS Utilization (S-MFU)-to enable accurate performance benchmarking of MoE systems across diverse hardware platforms and deployment scenarios.
Authors: Jierui Zhang, Jianhao Huang, Kaibin Huang
Abstract: Federated learning (FL) has emerged as a widely adopted paradigm for enabling edge learning with distributed data while ensuring data privacy. However, the traditional FL with deep neural networks trained via backpropagation can hardly meet the low-latency learning requirements in the sixth generation (6G) mobile networks. This challenge mainly arises from the high-dimensional model parameters to be transmitted and the numerous rounds of communication required for convergence due to the inherent randomness of the training process. To address this issue, we adopt the state-of-the-art principle of maximal coding rate reduction to learn linear discriminative features and extend the resultant white-box neural network into FL, yielding the novel framework of Low-Latency Federated Learning (LoLaFL) via forward-only propagation. LoLaFL enables layer-wise transmissions and aggregation with significantly fewer communication rounds, thereby considerably reducing latency. Additionally, we propose two \emph{nonlinear} aggregation schemes for LoLaFL. The first scheme is based on the proof that the optimal NN parameter aggregation in LoLaFL should be harmonic-mean-like. The second scheme further exploits the low-rank structures of the features and transmits the low-rank-approximated covariance matrices of features to achieve additional latency reduction. Theoretic analysis and experiments are conducted to evaluate the performance of LoLaFL. In comparison with traditional FL, the two nonlinear aggregation schemes for LoLaFL can achieve reductions in latency of over 87\% and 97\%, respectively, while maintaining comparable accuracies.
Authors: Xiao-Yin Liu, Guotao Li, Xiao-Hu Zhou, Zeng-Guang Hou
Abstract: Offline preference-based reinforcement learning (PbRL) provides an effective way to overcome the challenges of designing reward and the high costs of online interaction. However, since labeling preference needs real-time human feedback, acquiring sufficient preference labels is challenging. To solve this, this paper proposes a offLine prEference-bAsed RL with high Sample Efficiency (LEASE) algorithm, where a learned transition model is leveraged to generate unlabeled preference data. Considering the pretrained reward model may generate incorrect labels for unlabeled data, we design an uncertainty-aware mechanism to ensure the performance of reward model, where only high confidence and low variance data are selected. Moreover, we provide the generalization bound of reward model to analyze the factors influencing reward accuracy, and demonstrate that the policy learned by LEASE has theoretical improvement guarantee. The developed theory is based on state-action pair, which can be easily combined with other offline algorithms. The experimental results show that LEASE can achieve comparable performance to baseline under fewer preference data without online interaction.
Authors: Xudong Wang, Qingbo Hao, Xu Cheng, Yingyuan Xiao
Abstract: Federated learning offers a privacy-preserving framework for recommendation systems by enabling local data processing; however, data localization introduces substantial obstacles. Traditional federated recommendation approaches treat each user as an isolated entity, failing to construct global user relationship graphs that capture collaborative signals, which limits the accuracy of recommendations. To address this limitation, we derive insight from the insight that semantic similarity reflects preference. similarity, which can be used to improve the construction of user relationship graphs. This paper proposes UFGraphFR, a novel framework with three key components: 1) On the client side, private structured data is first transformed into text descriptions. These descriptions are then encoded into semantic vectors using pre-trained models; 2) On the server side, user relationship graphs are securely reconstructed using aggregated model weights without accessing raw data, followed by information propagation through lightweight graph neural networks; 3) On the client side, user behavior sequences are personalized using Transformer architectures. Extensive experiments conducted on four benchmark datasets demonstrate that UFGraphFR significantly outperforms state-of-the-art baselines in both recommendation accuracy and personalization. The framework also maintains robustness across different pre-trained models, as evidenced by the consistent performance metrics obtained. This work provides a practical method for efficient federated recommendations with strict privacy by using semantic vectors, secure user relationship graphs, and personalized behavior sequences. The code is available at: https://github.com/trueWangSyutung/UFGraphFR
Authors: Gongxu Luo, Haoyue Dai, Loka Li, Chengqian Gao, Boyang Sun, Kun Zhang
Abstract: Gene regulatory network inference (GRNI) aims to discover how genes causally regulate each other from gene expression data. It is well-known that statistical dependencies in observed data do not necessarily imply causation, as spurious dependencies may arise from latent confounders, such as non-coding RNAs. Numerous GRNI methods have thus been proposed to address this confounding issue. However, dependencies may also result from selection--only cells satisfying certain survival or inclusion criteria are observed--while these selection-induced spurious dependencies are frequently overlooked in gene expression data analyses. In this work, we show that such selection is ubiquitous and, when ignored or conflated with true regulations, can lead to flawed causal interpretation and misguided intervention recommendations. To address this challenge, a fundamental question arises: can we distinguish dependencies due to regulation, confounding, and crucially, selection? We show that gene perturbations offer a simple yet effective answer: selection-induced dependencies are symmetric under perturbation, while those from regulation or confounding are not. Building on this motivation, we propose GISL (Gene regulatory network Inference in the presence of Selection bias and Latent confounders), a principled algorithm that leverages perturbation data to uncover both true gene regulatory relations and non-regulatory mechanisms of selection and confounding up to the equivalence class. Experiments on synthetic and real-world gene expression data demonstrate the effectiveness of our method.
Authors: Xin Zhang, Weiliang Li, Rui Li, Zihang Fu, Tongyi Tang, Zhengyu Zhang, Wen-Yen Chen, Nima Noorshams, Nirav Jasapara, Xiaowen Ding, Ellie Wen, Xue Feng
Abstract: Optimizing conversions is crucial in modern online advertising systems, enabling advertisers to deliver relevant products to users and drive business outcomes. However, accurately predicting conversion events remains challenging due to variable time delays between user interactions (e.g., impressions or clicks) and the actual conversions. These delays vary substantially across advertisers and products, necessitating flexible optimization windows tailored to specific conversion behaviors. To address this, we propose a novel \textit{Personalized Interpolation} method that extends existing models based on fixed conversion windows to support flexible advertiser-specific optimization windows. Our method enables accurate conversion estimation across diverse delay distributions without increasing system complexity. We evaluate the effectiveness of the proposed approach through extensive experiments using a real-world ads conversion model. Our results show that this method achieves both high prediction accuracy and improved efficiency compared to existing solutions. This study demonstrates the potential of our Personalized Interpolation method to improve conversion optimization and support a wider range of advertising strategies in large-scale online advertising systems.
Authors: Arya Honarpisheh, Mustafa Bozdag, Octavia Camps, Mario Sznaier
Abstract: State-space models (SSMs) have recently emerged as a compelling alternative to Transformers for sequence modeling tasks. This paper presents a theoretical generalization analysis of selective SSMs, the core architectural component behind the Mamba model. We derive a novel covering number-based generalization bound for selective SSMs, building upon recent theoretical advances in the analysis of Transformer models. Using this result, we analyze how the spectral abscissa of the continuous-time state matrix influences the model's stability during training and its ability to generalize across sequence lengths. We empirically validate our findings on a synthetic majority task, the IMDb sentiment classification benchmark, and the ListOps task, demonstrating how our theoretical insights translate into practical model behavior.
Authors: Chaofan Lin, Jiaming Tang, Shuo Yang, Hanshuo Wang, Tian Tang, Boyu Tian, Ion Stoica, Song Han, Mingyu Gao
Abstract: Leveraging attention sparsity to accelerate long-context large language models (LLMs) has been a hot research topic. However, current algorithms such as sparse attention or key-value (KV) cache compression tend to use a fixed budget, which presents a significant challenge during deployment because it fails to account for the dynamic nature of real-world scenarios, where the optimal balance between accuracy and efficiency can vary greatly. In this paper, we find that borrowing top-$p$ sampling (nucleus sampling) to sparse attention can surprisingly achieve adaptive budgeting. Based on this, we propose Twilight, a framework to bring adaptive sparsity to any existing sparse attention algorithm without sacrificing their accuracy. Empirical results show that Twilight can adaptively prune at most 98% of redundant tokens, leading to $15.4\times$ acceleration in self-attention operations and $3.9\times$ acceleration in end-to-end per token latency in long context LLM decoding.
Authors: Daman Arora, Andrea Zanette
Abstract: Scaling model size and training data has led to great advances in the performance of Large Language Models (LLMs). However, the diminishing returns of this approach necessitate alternative methods to improve model capabilities, particularly in tasks requiring advanced reasoning. Large reasoning models, which leverage long chain-of-thoughts, bring unprecedented breakthroughs in problem-solving capabilities but at a substantial deployment cost associated to longer generations. Reducing inference costs is crucial for the economic feasibility, user experience, and environmental sustainability of these models. In this work, we propose to train large reasoning models to reason efficiently. More precisely, we use reinforcement learning (RL) to train reasoning models to dynamically allocate inference-time compute based on task complexity. Our method incentivizes models to minimize unnecessary computational overhead while maintaining accuracy, thereby achieving substantial efficiency gains. It enables the derivation of a family of reasoning models with varying efficiency levels, controlled via a single hyperparameter. Experiments on two open-weight large reasoning models demonstrate significant reductions in inference cost while preserving most of the accuracy.
Authors: Annie Marsden, Elad Hazan
Abstract: We study the problem of preconditioning in sequential prediction. From the theoretical lens of linear dynamical systems, we show that convolving the target sequence corresponds to applying a polynomial to the hidden transition matrix. Building on this insight, we propose a universal preconditioning method that convolves the target with coefficients from orthogonal polynomials such as Chebyshev or Legendre. We prove that this approach reduces regret for two distinct prediction algorithms and yields the first ever sublinear and hidden-dimension-independent regret bounds (up to logarithmic factors) that hold for systems with marginally table and asymmetric transition matrices. Finally, extensive synthetic and real-world experiments show that this simple preconditioning strategy improves the performance of a diverse range of algorithms, including recurrent neural networks, and generalizes to signals beyond linear dynamical systems.
Authors: Bill Marino, Meghdad Kurmanji, Nicholas D. Lane
Abstract: The ''right to be forgotten'' and the data privacy laws that encode it have motivated machine unlearning since its earliest days. Now, some argue that an inbound wave of artificial intelligence regulations -- like the European Union's Artificial Intelligence Act (AIA) -- may offer important new use cases for machine unlearning. However, this position paper argues, this opportunity will only be realized if researchers proactively bridge the (sometimes sizable) gaps between machine unlearning's state of the art and its potential applications to AI regulation. To demonstrate this point, we use the AIA as our primary case study. Specifically, we deliver a ``state of the union'' as regards machine unlearning's current potential (or, in many cases, lack thereof) for aiding compliance with various provisions of the AIA. This starts with a precise cataloging of the potential applications of machine unlearning to AIA compliance. For each, we flag the technical gaps that exist between the potential application and the state of the art of machine unlearning. Finally, we end with a call to action: for machine learning researchers to solve the open technical questions that could unlock machine unlearning's potential to assist compliance with the AIA -- and other AI regulations like it.
Authors: Ratun Rahman, Pablo Moriano, Samee U. Khan, Dinh C. Nguyen
Abstract: Electric load forecasting is essential for power management and stability in smart grids. This is mainly achieved via advanced metering infrastructure, where smart meters (SMs) record household energy data. Traditional machine learning (ML) methods are often employed for load forecasting, but require data sharing, which raises data privacy concerns. Federated learning (FL) can address this issue by running distributed ML models at local SMs without data exchange. However, current FL-based approaches struggle to achieve efficient load forecasting due to imbalanced data distribution across heterogeneous SMs. This paper presents a novel personalized federated learning (PFL) method for high-quality load forecasting in metering networks. A meta-learning-based strategy is developed to address data heterogeneity at local SMs in the collaborative training of local load forecasting models. Moreover, to minimize the load forecasting delays in our PFL model, we study a new latency optimization problem based on optimal resource allocation at SMs. A theoretical convergence analysis is also conducted to provide insights into FL design for federated load forecasting. Extensive simulations from real-world datasets show that our method outperforms existing approaches regarding better load forecasting and reduced operational latency costs.
Authors: Guanghan Wang, Yair Schiff, Subham Sekhar Sahoo, Volodymyr Kuleshov
Abstract: Part of the success of diffusion models stems from their ability to perform iterative refinement, i.e., repeatedly correcting outputs during generation. However, modern masked discrete diffusion lacks this capability: when a token is generated, it cannot be updated again, even when it introduces an error. Here, we address this limitation by introducing the remasking diffusion model (ReMDM) sampler, a method that can be applied to pretrained masked diffusion models in a principled way and that is derived from a discrete diffusion model with a custom remasking backward process. Most interestingly, ReMDM endows discrete diffusion with a form of inference-time compute scaling. By increasing the number of sampling steps, ReMDM generates natural language outputs that approach the quality of autoregressive models, whereas when the computation budget is limited, ReMDM better maintains quality. ReMDM also improves sample quality of masked diffusion models for discretized images, and in scientific domains such as molecule design, ReMDM facilitates diffusion guidance and pushes the Pareto frontier of controllability relative to classical masking and uniform noise diffusion. We provide the code along with a blog post on the project page: https://remdm.github.io
URLs: https://remdm.github.io
Authors: Abdullah Akg\"ul, Gulcin Baykal, Manuel Hau{\ss}mann, Melih Kandemir
Abstract: Continuous control of non-stationary environments is a major challenge for deep reinforcement learning algorithms. The time-dependency of the state transition dynamics aggravates the notorious stability problems of model-free deep actor-critic architectures. We posit that two properties will play a key role in overcoming non-stationarity in transition dynamics: (i)~preserving the plasticity of the critic network and (ii) directed exploration for rapid adaptation to changing dynamics. We show that performing on-policy reinforcement learning with an evidential critic provides both. The evidential design ensures a fast and accurate approximation of the uncertainty around the state value, which maintains the plasticity of the critic network by detecting the distributional shifts caused by changes in dynamics. The probabilistic critic also makes the actor training objective a random variable, enabling the use of directed exploration approaches as a by-product. We name the resulting algorithm \emph{Evidential Proximal Policy Optimization (EPPO)} due to the integral role of evidential uncertainty quantification in both policy evaluation and policy improvement stages. Through experiments on non-stationary continuous control tasks, where the environment dynamics change at regular intervals, we demonstrate that our algorithm outperforms state-of-the-art on-policy reinforcement learning variants in both task-specific and overall return.
Authors: Bassel El Mabsout, Abdelrahman Abdelgawad, Renato Mancuso
Abstract: Practitioners designing reinforcement learning policies face a fundamental challenge: translating intended behavioral objectives into representative reward functions. This challenge stems from behavioral intent requiring simultaneous achievement of multiple competing objectives, typically addressed through labor-intensive linear reward composition that yields brittle results. Consider the ubiquitous robotics scenario where performance maximization directly conflicts with energy conservation. Such competitive dynamics are resistant to simple linear reward combinations. In this paper, we present the concept of objective fulfillment upon which we build Fulfillment Priority Logic (FPL). FPL allows practitioners to define logical formula representing their intentions and priorities within multi-objective reinforcement learning. Our novel Balanced Policy Gradient algorithm leverages FPL specifications to achieve up to 500\% better sample efficiency compared to Soft Actor Critic. Notably, this work constitutes the first implementation of non-linear utility scalarization design, specifically for continuous control problems.
Authors: Aditya Shankar, Lydia Y. Chen, Arie van Deursen, Rihan Hai
Abstract: Generating temporal data under conditions is crucial for forecasting, imputation, and generative tasks. Such data often has metadata and partially observed signals that jointly influence the generated values. However, existing methods face three key limitations: (1) they condition on either the metadata or observed values, but rarely both together; (2) they adopt either training-time approaches that fail to generalize to unseen scenarios, or inference-time approaches that ignore metadata; and (3) they suffer from trade-offs between generation speed and temporal coherence across time windows--choosing either slow but coherent autoregressive methods or fast but incoherent parallel ones. We propose WaveStitch, a novel diffusion-based method to overcome these hurdles through: (1) dual-sourced conditioning on both metadata and partially observed signals; (2) a hybrid training-inference architecture, incorporating metadata during training and observations at inference via gradient-based guidance; and (3) a novel pipeline-style paradigm that generates time windows in parallel while preserving coherence through an inference-time conditional loss and a stitching mechanism. Across diverse datasets, WaveStitch demonstrates adaptability to arbitrary patterns of observed signals, achieving 1.81x lower mean-squared-error compared to the state-of-the-art, and generates data up to 166.48x faster than autoregressive methods while maintaining coherence. Our code is available at: https://github.com/adis98/WaveStitch
Authors: Jes\'us Garc\'ia Fern\'andez, Nasir Ahmad, Marcel van Gerven
Abstract: The pursuit of energy-efficient and adaptive artificial intelligence (AI) has positioned neuromorphic computing as a promising alternative to conventional computing. However, achieving learning on these platforms requires techniques that prioritize local information while enabling effective credit assignment. Here, we propose noise-based reward-modulated learning (NRL), a novel synaptic plasticity rule that mathematically unifies reinforcement learning and gradient-based optimization with biologically-inspired local updates. NRL addresses the computational bottleneck of exact gradients by approximating them through stochastic neural activity, transforming the inherent noise of biological and neuromorphic substrates into a functional resource. Drawing inspiration from biological learning, our method uses reward prediction errors as its optimization target to generate increasingly advantageous behavior, and eligibility traces to facilitate retrospective credit assignment. Experimental validation on reinforcement tasks, featuring immediate and delayed rewards, shows that NRL achieves performance comparable to baselines optimized using backpropagation, although with slower convergence, while showing significantly superior performance and scalability in multi-layer networks compared to reward-modulated Hebbian learning (RMHL), the most prominent similar approach. While tested on simple architectures, the results highlight the potential of noise-driven, brain-inspired learning for low-power adaptive systems, particularly in computing substrates with locality constraints. NRL offers a theoretically grounded paradigm well-suited for the event-driven characteristics of next-generation neuromorphic AI.
Authors: Claudio Spiess, Mandana Vaziri, Louis Mandel, Martin Hirzel
Abstract: The performance of large language models (LLMs) depends on how they are prompted, with choices spanning both the high-level prompting pattern (e.g., Zero-Shot, CoT, ReAct, ReWOO) and the specific prompt content (instructions and few-shot demonstrations). Manually tuning this combination is tedious, error-prone, and specific to a given LLM and task. Therefore, this paper proposes AutoPDL, an automated approach to discovering good LLM agent configurations. Our approach frames this as a structured AutoML problem over a combinatorial space of agentic and non-agentic prompting patterns and demonstrations, using successive halving to efficiently navigate this space. We introduce a library implementing common prompting patterns using the PDL prompt programming language. AutoPDL solutions are human-readable, editable, and executable PDL programs that use this library. This approach also enables source-to-source optimization, allowing human-in-the-loop refinement and reuse. Evaluations across three tasks and seven LLMs (ranging from 3B to 70B parameters) show consistent accuracy gains ($9.21\pm15.46$ percentage points), up to 67.5pp, and reveal that selected prompting strategies vary across models and tasks.
Authors: Ashwinee Panda, Vatsal Baherwani, Zain Sarwar, Benjamin Therien, Sambit Sahu, Tom Goldstein, Supriyo Chakraborty
Abstract: Mixture of Experts (MoE) pretraining is more scalable than dense Transformer pretraining, because MoEs learn to route inputs to a sparse set of their feedforward parameters. However, this means that MoEs only receive a sparse backward update, leading to training instability and suboptimal performance. We present a lightweight approximation method that gives the MoE router a dense gradient update while continuing to sparsely activate its parameters. Our method, which we refer to as Default MoE, substitutes missing expert activations with default outputs consisting of an exponential moving average of expert outputs previously seen over the course of training. This allows the router to receive signals from every expert for each token, leading to significant improvements in training performance. Our Default MoE outperforms standard TopK routing in a variety of settings without requiring significant computational overhead. Code: https://github.com/vatsal0/default-moe.
Authors: Yunwei Ren, Eshaan Nichani, Denny Wu, Jason D. Lee
Abstract: We study the complexity of online stochastic gradient descent (SGD) for learning a two-layer neural network with $P$ neurons on isotropic Gaussian data: $f_*(\boldsymbol{x}) = \sum_{p=1}^P a_p\cdot \sigma(\langle\boldsymbol{x},\boldsymbol{v}_p^*\rangle)$, $\boldsymbol{x} \sim \mathcal{N}(0,\boldsymbol{I}_d)$, where the activation $\sigma:\mathbb{R}\to\mathbb{R}$ is an even function with information exponent $k_*>2$ (defined as the lowest degree in the Hermite expansion), $\{\boldsymbol{v}^*_p\}_{p\in[P]}\subset \mathbb{R}^d$ are orthonormal signal directions, and the non-negative second-layer coefficients satisfy $\sum_{p} a_p^2=1$. We focus on the challenging ``extensive-width'' regime $P\gg 1$ and permit diverging condition number in the second-layer, covering as a special case the power-law scaling $a_p\asymp p^{-\beta}$ where $\beta\in\mathbb{R}_{\ge 0}$. We provide a precise analysis of SGD dynamics for the training of a student two-layer network to minimize the mean squared error (MSE) objective, and explicitly identify sharp transition times to recover each signal direction. In the power-law setting, we characterize scaling law exponents for the MSE loss with respect to the number of training samples and SGD steps, as well as the number of parameters in the student neural network. Our analysis entails that while the learning of individual teacher neurons exhibits abrupt transitions, the juxtaposition of $P\gg 1$ emergent learning curves at different timescales leads to a smooth scaling law in the cumulative objective.
Authors: Tien Comlekoglu, J. Quetzalc\'oatl Toledo-Mar\'in, Tina Comlekoglu, Douglas W. DeSimone, Shayn M. Peirce, Geoffrey Fox, James A. Glazier
Abstract: The Cellular-Potts model is a powerful and ubiquitous framework for developing computational models for simulating complex multicellular biological systems. Cellular-Potts models (CPMs) are often computationally expensive due to the explicit modeling of interactions among large numbers of individual model agents and diffusive fields described by partial differential equations (PDEs). In this work, we develop a convolutional neural network (CNN) surrogate model using a U-Net architecture that accounts for periodic boundary conditions. We use this model to accelerate the evaluation of a mechanistic CPM previously used to investigate in vitro vasculogenesis. The surrogate model was trained to predict 100 computational steps ahead (Monte-Carlo steps, MCS), accelerating simulation evaluations by a factor of 590 times compared to CPM code execution. Over multiple recursive evaluations, our model effectively captures the emergent behaviors demonstrated by the original Cellular-Potts model of such as vessel sprouting, extension and anastomosis, and contraction of vascular lacunae. This approach demonstrates the potential for deep learning to serve as efficient surrogate models for CPM simulations, enabling faster evaluation of computationally expensive CPM of biological processes at greater spatial and temporal scales.
Authors: Antonio Candito, Matthew D Blackledge, Richard Holbrey, Nuria Porta, Ana Ribeiro, Fabio Zugni, Luca D'Erme, Francesca Castagnoli, Alina Dragan, Ricardo Donners, Christina Messiou, Nina Tunariu, Dow-Mu Koh
Abstract: Quantitative assessment of treatment response in Advanced Prostate Cancer (APC) with bone metastases remains an unmet clinical need. Whole-Body Diffusion-Weighted MRI (WB-DWI) provides two response biomarkers: Total Diffusion Volume (TDV) and global Apparent Diffusion Coefficient (gADC). However, tracking post-treatment changes of TDV and gADC from manually delineated lesions is cumbersome and increases inter-reader variability. We developed a software to automate this process. Core technologies include: (i) a weakly-supervised Residual U-Net model generating a skeleton probability map to isolate bone; (ii) a statistical framework for WB-DWI intensity normalisation, obtaining a signal-normalised b=900s/mm^2 (b900) image; and (iii) a shallow convolutional neural network that processes outputs from (i) and (ii) to generate a mask of suspected bone lesions, characterised by higher b900 signal intensity due to restricted water diffusion. This mask is applied to the gADC map to extract TDV and gADC statistics. We tested the tool using expert-defined metastatic bone disease delineations on 66 datasets, assessed repeatability of imaging biomarkers (N=10), and compared software-based response assessment with a construct reference standard (N=118). Average dice score between manual and automated delineations was 0.6 for lesions within pelvis and spine, with an average surface distance of 2mm. Relative differences for log-transformed TDV (log-TDV) and median gADC were 8.8% and 5%, respectively. Repeatability analysis showed coefficients of variation of 4.6% for log-TDV and 3.5% for median gADC, with intraclass correlation coefficients of 0.94 or higher. The software achieved 80.5% accuracy, 84.3% sensitivity, and 85.7% specificity in assessing response to treatment. Average computation time was 90s per scan.
Authors: Hongjoon Ahn, Heewoong Choi, Jisu Han, Taesup Moon
Abstract: Offline goal-conditioned reinforcement learning (GCRL) offers a practical learning paradigm in which goal-reaching policies are trained from abundant state-action trajectory datasets without additional environment interaction. However, offline GCRL still struggles with long-horizon tasks, even with recent advances that employ hierarchical policy structures, such as HIQL. Identifying the root cause of this challenge, we observe the following insight. Firstly, performance bottlenecks mainly stem from the high-level policy's inability to generate appropriate subgoals. Secondly, when learning the high-level policy in the long-horizon regime, the sign of the advantage estimate frequently becomes incorrect. Thus, we argue that improving the value function to produce a clear advantage estimate for learning the high-level policy is essential. In this paper, we propose a simple yet effective solution: Option-aware Temporally Abstracted value learning, dubbed OTA, which incorporates temporal abstraction into the temporal-difference learning process. By modifying the value update to be option-aware, our approach contracts the effective horizon length, enabling better advantage estimates even in long-horizon regimes. We experimentally show that the high-level policy learned using the OTA value function achieves strong performance on complex tasks from OGBench, a recently proposed offline GCRL benchmark, including maze navigation and visual robotic manipulation environments.
Authors: Viet Anh Khoa Tran, Emre Neftci, Willem A. M. Wybo
Abstract: Biological brains learn continually from a stream of unlabeled data, while integrating specialized information from sparsely labeled examples without compromising their ability to generalize. Meanwhile, machine learning methods are susceptible to catastrophic forgetting in this natural learning setting, as supervised specialist fine-tuning degrades performance on the original task. We introduce task-modulated contrastive learning (TMCL), which takes inspiration from the biophysical machinery in the neocortex, using predictive coding principles to integrate top-down information continually and without supervision. We follow the idea that these principles build a view-invariant representation space, and that this can be implemented using a contrastive loss. Then, whenever labeled samples of a new class occur, new affine modulations are learned that improve separation of the new class from all others, without affecting feedforward weights. By co-opting the view-invariance learning mechanism, we then train feedforward weights to match the unmodulated representation of a data sample to its modulated counterparts. This introduces modulation invariance into the representation space, and, by also using past modulations, stabilizes it. Our experiments show improvements in both class-incremental and transfer learning over state-of-the-art unsupervised approaches, as well as over comparable supervised approaches, using as few as 1% of available labels. Taken together, our work suggests that top-down modulations play a crucial role in balancing stability and plasticity.
Authors: Sho Sonoda, Yuka Hashimoto, Isao Ishikawa, Masahiro Ikeda
Abstract: Why and when is deep better than shallow? We answer this question in a framework that is agnostic to network implementation. We formulate a deep model as an abstract state-transition semigroup acting on a general metric space, and separate the implementation (e.g., ReLU nets, transformers, and chain-of-thought) from the abstract state transition. We prove a bias-variance decomposition in which the variance depends only on the abstract depth-$k$ network and not on the implementation (Theorem 1). We further split the bounds into output and hidden parts to tie the depth dependence of the variance to the metric entropy of the state-transition semigroup (Theorem 2). We then investigate implementation-free conditions under which the variance grow polynomially or logarithmically with depth (Section 4). Combining these with exponential or polynomial bias decay identifies four canonical bias-variance trade-off regimes (EL/EP/PL/PP) and produces explicit optimal depths $k^\ast$. Across regimes, $k^\ast>1$ typically holds, giving a rigorous form of depth supremacy. The lowest generalization error bound is achieved under the EL regime (exp-decay bias + log-growth variance), explaining why and when deep is better, especially for iterative or hierarchical concept classes such as neural ODEs, diffusion/score-matching models, and chain-of-thought reasoning.
Authors: Nicolas Castanet, Olivier Sigaud, Sylvain Lamprier
Abstract: Goal-Conditioned Reinforcement Learning (GCRL) enables agents to autonomously acquire diverse behaviors, but faces major challenges in visual environments due to high-dimensional, semantically sparse observations. In the online setting, where agents learn representations while exploring, the latent space evolves with the agent's policy, to capture newly discovered areas of the environment. However, without incentivization to maximize state coverage in the representation, classical approaches based on auto-encoders may converge to latent spaces that over-represent a restricted set of states frequently visited by the agent. This is exacerbated in an intrinsic motivation setting, where the agent uses the distribution encoded in the latent space to sample the goals it learns to master. To address this issue, we propose to progressively enforce distributional shifts towards a uniform distribution over the full state space, to ensure a full coverage of skills that can be learned in the environment. We introduce DRAG (Distributionally Robust Auto-Encoding for GCRL), a method that combines the $\beta$-VAE framework with Distributionally Robust Optimization. DRAG leverages an adversarial neural weighter of training states of the VAE, to account for the mismatch between the current data distribution and unseen parts of the environment. This allows the agent to construct semantically meaningful latent spaces beyond its immediate experience. Our approach improves state space coverage and downstream control performance on hard exploration environments such as mazes and robotic control involving walls to bypass, without pre-training nor prior environment knowledge.
Authors: Qi Cao, Ruiyi Wang, Ruiyi Zhang, Sai Ashish Somayajula, Pengtao Xie
Abstract: Reasoning has substantially improved the performance of large language models (LLMs) on complicated tasks. Central to the current reasoning studies, Process Reward Models (PRMs) offer a fine-grained evaluation of intermediate reasoning steps and guide the reasoning process. However, extending PRMs to multimodal large language models (MLLMs) introduces challenges. Since multimodal reasoning covers a wider range of tasks compared to text-only scenarios, the resulting distribution shift from the training to testing sets is more severe, leading to greater generalization difficulty. Training a reliable multimodal PRM, therefore, demands large and diverse datasets to ensure sufficient coverage. However, current multimodal reasoning datasets suffer from a marked quality imbalance, which degrades PRM performance and highlights the need for an effective data selection strategy. To address the issues, we introduce DreamPRM, a domain-reweighted training framework for multimodal PRMs which employs bi-level optimization. In the lower-level optimization, DreamPRM performs fine-tuning on multiple datasets with domain weights, allowing the PRM to prioritize high-quality reasoning signals and alleviating the impact of dataset quality imbalance. In the upper-level optimization, the PRM is evaluated on a separate meta-learning dataset; this feedback updates the domain weights through an aggregation loss function, thereby improving the generalization capability of trained PRM. Extensive experiments on multiple multimodal reasoning benchmarks covering both mathematical and general reasoning show that test-time scaling with DreamPRM consistently improves the performance of state-of-the-art MLLMs. Further comparisons reveal that DreamPRM's domain-reweighting strategy surpasses other data selection methods and yields higher accuracy gains than existing test-time scaling approaches.
Authors: Benyamin Trachtenberg, Nir Rosenfeld
Abstract: In strategic classification, the standard supervised learning setting is extended to support the notion of strategic user behavior in the form of costly feature manipulations made in response to a classifier. While standard learning supports a broad range of model classes, the study of strategic classification has, so far, been dedicated mostly to linear classifiers. This work aims to expand the horizon by exploring how strategic behavior manifests under non-linear classifiers and what this implies for learning. We take a bottom-up approach showing how non-linearity affects decision boundary points, classifier expressivity, and model class complexity. Our results show how, unlike the linear case, strategic behavior may either increase or decrease effective class complexity, and that the complexity decrease may be arbitrarily large. Another key finding is that universal approximators (e.g., neural nets) are no longer universal once the environment is strategic. We demonstrate empirically how this can create performance gaps even on an unrestricted model class.
Authors: Yuepeng Zheng, Fu Luo, Zhenkun Wang, Yaoxin Wu, Yu Zhou
Abstract: Multi-Task Learning (MTL) in Neural Combinatorial Optimization (NCO) is a promising approach to train a unified model capable of solving multiple Vehicle Routing Problem (VRP) variants. However, existing Reinforcement Learning (RL)-based multi-task methods can only train light decoder models on small-scale problems, exhibiting limited generalization ability when solving large-scale problems. To overcome this limitation, this work introduces a novel multi-task learning method driven by knowledge distillation (MTL-KD), which enables the efficient training of heavy decoder models with strong generalization ability. The proposed MTL-KD method transfers policy knowledge from multiple distinct RL-based single-task models to a single heavy decoder model, facilitating label-free training and effectively improving the model's generalization ability across diverse tasks. In addition, we introduce a flexible inference strategy termed Random Reordering Re-Construction (R3C), which is specifically adapted for diverse VRP tasks and further boosts the performance of the multi-task model. Experimental results on 6 seen and 10 unseen VRP variants with up to 1000 nodes indicate that our proposed method consistently achieves superior performance on both uniform and real-world benchmarks, demonstrating robust generalization abilities.
Authors: Val\'erie Costa, Thomas Fel, Ekdeep Singh Lubana, Bahareh Tolooshams, Demba Ba
Abstract: Motivated by the hypothesis that neural network representations encode abstract, interpretable features as linearly accessible, approximately orthogonal directions, sparse autoencoders (SAEs) have become a popular tool in interpretability. However, recent work has demonstrated phenomenology of model representations that lies outside the scope of this hypothesis, showing signatures of hierarchical, nonlinear, and multi-dimensional features. This raises the question: do SAEs represent features that possess structure at odds with their motivating hypothesis? If not, does avoiding this mismatch help identify said features and gain further insights into neural network representations? To answer these questions, we take a construction-based approach and re-contextualize the popular matching pursuits (MP) algorithm from sparse coding to design MP-SAE -- an SAE that unrolls its encoder into a sequence of residual-guided steps, allowing it to capture hierarchical and nonlinearly accessible features. Comparing this architecture with existing SAEs on a mixture of synthetic and natural data settings, we show: (i) hierarchical concepts induce conditionally orthogonal features, which existing SAEs are unable to faithfully capture, and (ii) the nonlinear encoding step of MP-SAE recovers highly meaningful features, helping us unravel shared structure in the seemingly dichotomous representation spaces of different modalities in a vision-language model, hence demonstrating the assumption that useful features are solely linearly accessible is insufficient. We also show that the sequential encoder principle of MP-SAE affords an additional benefit of adaptive sparsity at inference time, which may be of independent interest. Overall, we argue our results provide credence to the idea that interpretability should begin with the phenomenology of representations, with methods emerging from assumptions that fit it.
Authors: Viet Nguyen, Changjian Shui, Vijay Giri, Siddharth Arya, Amol Verma, Fahad Razak, Rahul G. Krishnan
Abstract: The distribution of data changes over time; models operating in dynamic environments need retraining. But knowing when to retrain, without access to labels, is an open challenge since some, but not all shifts degrade model performance. This paper formalizes and addresses the problem of post-deployment deterioration (PDD) monitoring. We propose D3M, a practical and efficient monitoring algorithm based on the disagreement of predictive models, achieving low false positive rates under non-deteriorating shifts and provides sample complexity bounds for high true positive rates under deteriorating shifts. Empirical results on both standard benchmark and a real-world large-scale internal medicine dataset demonstrate the effectiveness of the framework and highlight its viability as an alert mechanism for high-stakes machine learning pipelines.
Authors: Val\'erie Costa, Thomas Fel, Ekdeep Singh Lubana, Bahareh Tolooshams, Demba Ba
Abstract: Sparse autoencoders (SAEs) have recently become central tools for interpretability, leveraging dictionary learning principles to extract sparse, interpretable features from neural representations whose underlying structure is typically unknown. This paper evaluates SAEs in a controlled setting using MNIST, which reveals that current shallow architectures implicitly rely on a quasi-orthogonality assumption that limits the ability to extract correlated features. To move beyond this, we compare them with an iterative SAE that unrolls Matching Pursuit (MP-SAE), enabling the residual-guided extraction of correlated features that arise in hierarchical settings such as handwritten digit generation while guaranteeing monotonic improvement of the reconstruction as more atoms are selected.
Authors: Chuang Ma, Tomoyuki Obuchi, Toshiyuki Tanaka
Abstract: A phenomenon known as ''Neural Collapse (NC)'' in deep classification tasks, in which the penultimate-layer features and the final classifiers exhibit an extremely simple geometric structure, has recently attracted considerable attention, with the expectation that it can deepen our understanding of how deep neural networks behave. The Unconstrained Feature Model (UFM) has been proposed to explain NC theoretically, and there emerges a growing body of work that extends NC to tasks other than classification and leverages it for practical applications. In this study, we investigate whether a similar phenomenon arises in deep Ordinal Regression (OR) tasks, via combining the cumulative link model for OR and UFM. We show that a phenomenon we call Ordinal Neural Collapse (ONC) indeed emerges and is characterized by the following three properties: (ONC1) all optimal features in the same class collapse to their within-class mean when regularization is applied; (ONC2) these class means align with the classifier, meaning that they collapse onto a one-dimensional subspace; (ONC3) the optimal latent variables (corresponding to logits or preactivations in classification tasks) are aligned according to the class order, and in particular, in the zero-regularization limit, a highly local and simple geometric relationship emerges between the latent variables and the threshold values. We prove these properties analytically within the UFM framework with fixed threshold values and corroborate them empirically across a variety of datasets. We also discuss how these insights can be leveraged in OR, highlighting the use of fixed thresholds.
Authors: Utkarsh Utkarsh, Danielle C. Maddix, Ruijun Ma, Michael W. Mahoney, Yuyang Wang
Abstract: We present ProbHardE2E, a probabilistic forecasting framework that incorporates hard operational/physical constraints, and provides uncertainty quantification. Our methodology uses a novel differentiable probabilistic projection layer (DPPL) that can be combined with a wide range of neural network architectures. DPPL allows the model to learn the system in an end-to-end manner, compared to other approaches where constraints are satisfied either through a post-processing step or at inference. ProbHardE2E optimizes a strictly proper scoring rule, without making any distributional assumptions on the target, which enables it to obtain robust distributional estimates (in contrast to existing approaches that generally optimize likelihood-based objectives, which are heavily biased by their distributional assumptions and model choices); and it can incorporate a range of non-linear constraints (increasing the power of modeling and flexibility). We apply ProbHardE2E in learning partial differential equations with uncertainty estimates and to probabilistic time-series forecasting, showcasing it as a broadly applicable general framework that connects these seemingly disparate domains.
Authors: Vincenzo Collura, Karim Tit, Laura Bussi, Eleonora Giunchiglia, Maxime Cordy
Abstract: Sequence generation and prediction form a cornerstone of modern machine learning, with applications spanning natural language processing, program synthesis, and time-series forecasting. These tasks are typically modeled in an autoregressive fashion, where each token is generated conditional on the preceding ones, and beam search is commonly used to balance exploration and fluency during decoding. While deep learning models and Large Language Models (LLMs) excel at capturing statistical patterns in this setting, they remain ill-equipped to guarantee compliance with formal constraints. In this paper, we introduce ABS: a general and model-agnostic inference-time algorithm that guarantees compliance with any constraint that can be compiled into a Deterministic Finite Automaton (DFA), without requiring retraining. ABS leverages the DFA to guide a constrained variant of beam search: at each decoding step, transitions leading to violations are masked, while remaining paths are dynamically re-ranked according to both the model's probabilities and the automaton's acceptance structure. We formally prove that the resulting sequences are guaranteed to satisfy the given constraints, and we empirically demonstrate that ABS also improves output quality. We validate our approach on three distinct tasks: constrained image-stream classification, controlled text generation, and text infilling. In all settings, ABS achieves perfect constraint satisfaction, while outperforming or matching state-of-the-art baselines on standard quality metrics and efficiency.
Authors: Minhyuk Seo, Taeheon Kim, Hankook Lee, Jonghyun Choi, Tinne Tuytelaars
Abstract: As AI becomes more personal, e.g., Agentic AI, there is an increasing need for personalizing models for various use cases. Personalized federated learning (PFL) enables each client to collaboratively leverage other clients' knowledge for better adaptation to the task of interest, without privacy risks. Despite its potential, existing PFL methods remain confined to rather simplified scenarios where data and models are the same across clients. To move towards realistic scenarios, we propose FedMosaic, a method that jointly addresses data and model heterogeneity with a task-relevance-aware model aggregation strategy to reduce parameter interference, and a dimension-invariant module that enables knowledge sharing across heterogeneous architectures without huge computational cost. To mimic the real-world task diversity, we propose a multi-modal PFL benchmark spanning 40 distinct tasks with distribution shifts over time. The empirical study shows that FedMosaic outperforms the state-of-the-art PFL methods, excelling in both personalization and generalization capabilities under challenging, realistic scenarios.
Authors: Elif Y{\i}lmaz, Christos Dimitrakakis
Abstract: We study a two-player zero-sum game in which the row player aims to maximize their payoff against an adversarial column player, under an unknown payoff matrix estimated through bandit feedback. We propose three algorithms based on the Explore-Then-Commit framework. The first adapts it to zero-sum games, the second incorporates adaptive elimination that leverages the $\varepsilon$-Nash Equilibrium property to efficiently select the optimal action pair, and the third extends the elimination algorithm by employing non-uniform exploration. Our objective is to demonstrate the applicability of ETC in a zero-sum game setting by focusing on learning pure strategy Nash Equilibria. A key contribution of our work is a derivation of instance-dependent upper bounds on the expected regret of our proposed algorithms, which has received limited attention in the literature on zero-sum games. Particularly, after $T$ rounds, we achieve an instance-dependent regret upper bounds of $O(\Delta + \sqrt{T})$ for ETC in zero-sum game setting and $O(\log (T \Delta^2) / \Delta)$ for the adaptive elimination algorithm and its variant with non-uniform exploration, where $\Delta$ denotes the suboptimality gap. Therefore, our results indicate that ETC-based algorithms perform effectively in adversarial game settings, achieving regret bounds comparable to existing methods while providing insight through instance-dependent analysis.
Authors: Michael Plainer, Hao Wu, Leon Klein, Stephan G\"unnemann, Frank No\'e
Abstract: In recent years, diffusion models trained on equilibrium molecular distributions have proven effective for sampling biomolecules. Beyond direct sampling, the score of such a model can also be used to derive the forces that act on molecular systems. However, while classical diffusion sampling usually recovers the training distribution, the corresponding energy-based interpretation of the learned score is often inconsistent with this distribution, even for low-dimensional toy systems. We trace this inconsistency to inaccuracies of the learned score at very small diffusion timesteps, where the model must capture the correct evolution of the data distribution. In this regime, diffusion models fail to satisfy the Fokker--Planck equation, which governs the evolution of the score. We interpret this deviation as one source of the observed inconsistencies and propose an energy-based diffusion model with a Fokker--Planck-derived regularization term to enforce consistency. We demonstrate our approach by sampling and simulating multiple biomolecular systems, including fast-folding proteins, and by introducing a state-of-the-art transferable Boltzmann emulator for dipeptides that supports simulation and achieves improved consistency and efficient sampling. Our code, model weights, and self-contained JAX and PyTorch notebooks are available at https://github.com/noegroup/ScoreMD.
Authors: Paul Saves, Edward Hall\'e-Hannan, Jasper Bussemaker, Youssef Diouane, Nathalie Bartoli
Abstract: Simulation-based problems involving mixed-variable inputs frequently feature domains that are hierarchical, conditional, heterogeneous, or tree-structured. These characteristics pose challenges for data representation, modeling, and optimization. This paper reviews extensive literature on these structured input spaces and proposes a unified framework that generalizes existing approaches. In this framework, input variables may be continuous, integer, or categorical. A variable is described as meta if its value governs the presence of other decreed variables, enabling the modeling of conditional and hierarchical structures. We further introduce the concept of partially-decreed variables, whose activation depends on contextual conditions. To capture these inter-variable hierarchical relationships, we introduce design space graphs, combining principles from feature modeling and graph theory. This allows the definition of general hierarchical domains suitable for describing complex system architectures. Our framework defines hierarchical distances and kernels to enable surrogate modeling and optimization on hierarchical domains. We demonstrate its effectiveness on complex system design problems, including a neural network and a green-aircraft case study. Our methods are available in the open-source Surrogate Modeling Toolbox (SMT 2.0).
Authors: Reece Bourisaw, Reid McCants, Jean-Marie Le Corre, Anna Iskhakova, Arsen S. Iskhakov
Abstract: Critical heat flux (CHF) marks the onset of boiling crisis in light-water reactors, defining safe thermal-hydraulic operating limits. To support Phase II of the OECD/NEA AI/ML CHF benchmark, which introduces spatially varying power profiles, this work compiles and digitizes a broad CHF dataset covering both uniform and non-uniform axial heating conditions. Heating profiles were extracted from technical reports, interpolated onto a consistent axial mesh, validated via energy-balance checks, and encoded in machine-readable formats for benchmark compatibility. Classical CHF correlations exhibit substantial errors under uniform heating and degrade markedly when applied to non-uniform profiles, while modern tabular methods offer improved but still imperfect predictions. A neural network trained solely on uniform data performs well in that regime but fails to generalize to spatially varying scenarios, underscoring the need for models that explicitly incorporate axial power distributions. By providing these curated datasets and baseline modeling results, this study lays the groundwork for advanced transfer-learning strategies, rigorous uncertainty quantification, and design-optimization efforts in the next phase of the CHF benchmark.
Authors: Matteo Negro, Andrea Piras, Ragib Ahsan, David Arbour, Elena Zheleva
Abstract: Estimating causal effects from real-world relational data can be challenging when the underlying causal model and potential confounders are unknown. While several causal discovery algorithms exist for learning causal models with latent confounders from data, they assume that the data is independent and identically distributed (i.i.d.) and are not well-suited for learning from relational data. Similarly, existing relational causal discovery algorithms assume causal sufficiency, which is unrealistic for many real-world datasets. To address this gap, we propose RelFCI, a sound and complete causal discovery algorithm for relational data with latent confounders. Our work builds upon the Fast Causal Inference (FCI) and Relational Causal Discovery (RCD) algorithms and it defines new graphical models, necessary to support causal discovery in relational domains. We also establish soundness and completeness guarantees for relational d-separation with latent confounders. We present experimental results demonstrating the effectiveness of RelFCI in identifying the correct causal structure in relational causal models with latent confounders.
Authors: A. Bochkov
Abstract: The prevailing paradigm for scaling large language models (LLMs) involves monolithic, end-to-end training, a resource-intensive process that lacks flexibility. This paper explores an alternative, constructive scaling paradigm, enabled by the principle of emergent semantics in Transformers with frozen, non-semantic input embeddings. We posit that because high-level meaning is a compositional property of a Transformer's deep layers, not its input vectors, the embedding layer and trained lower layers can serve as a fixed foundation. This liberates backpropagation to focus solely on newly added components, making incremental growth viable. We operationalize this with a layer-wise constructive methodology that combines strict layer freezing in early stages with efficient, holistic fine-tuning of the entire model stack via low-rank adaptation (LoRA) as complexity increases. This method not only demonstrates stable convergence but also reveals a direct correlation between model depth and the emergence of complex reasoning abilities, such as those required for SQuAD, which are absent in shallower models. In a controlled study, our constructively grown model rivals the performance of a monolithically trained baseline of the same size, validating the efficiency and efficacy of the approach. Our findings suggest a path towards a paradigm shift from monolithic optimization towards a more biological or constructive model of AI development. This opens a path for more resource-efficient scaling, continual learning, and a more modular approach to building powerful AI systems. We release all code and models to facilitate further research.
Authors: Thomas T. Zhang, Daniel Pfrommer, Chaoyi Pan, Nikolai Matni, Max Simchowitz
Abstract: This paper presents a theoretical analysis of two of the most impactful interventions in modern learning from demonstration in robotics and continuous control: the practice of action-chunking (predicting sequences of actions in open-loop) and exploratory augmentation of expert demonstrations. Though recent results show that learning from demonstration, also known as imitation learning (IL), can suffer errors that compound exponentially with task horizon in continuous settings, we demonstrate that action chunking and exploratory data collection circumvent exponential compounding errors in different regimes. Our results identify control-theoretic stability as the key mechanism underlying the benefits of these interventions. On the empirical side, we validate our predictions and the role of control-theoretic stability through experimentation on popular robot learning benchmarks. On the theoretical side, we demonstrate that the control-theoretic lens provides fine-grained insights into how compounding error arises, leading to tighter statistical guarantees on imitation learning error when these interventions are applied than previous techniques based on information-theoretic considerations alone.
Authors: Ziniu Zhang, Zhenshuo Zhang, Dongyue Li, Lu Wang, Jennifer Dy, Hongyang R. Zhang
Abstract: This paper introduces an algorithm to select demonstration examples for in-context learning of a query set. Given a set of $n$ examples, how can we quickly select $k$ out of $n$ to best serve as the conditioning for downstream inference? This problem has broad applications in prompt tuning and chain-of-thought reasoning. Since model weights remain fixed during in-context learning, previous work has sought to design methods based on the similarity of token embeddings. This work proposes a new approach based on gradients of the output taken in the input embedding space. Our approach estimates model outputs through a first-order approximation using the gradients. Then, we apply this estimation to multiple randomly sampled subsets. Finally, we aggregate the sampled subset outcomes to form an influence score for each demonstration, and select $k$ most relevant examples. This procedure only requires pre-computing model outputs and gradients once, resulting in a linear-time algorithm relative to model and training set sizes. Extensive experiments across various models and datasets validate the efficiency of our approach. We show that the gradient estimation procedure yields approximations of full inference with less than ${1}\%$ error across six datasets. This allows us to scale up subset selection that would otherwise run full inference by up to ${37.7}\times$ on models with up to $34$ billion parameters, and outperform existing selection methods based on input embeddings by ${11}\%$ on average.
Authors: Jingyuan Zhou, Hao Qian, Shikui Tu, Lei Xu
Abstract: Structure-based drug design (SBDD), aiming to generate 3D molecules with high binding affinity toward target proteins, is a vital approach in novel drug discovery. Although recent generative models have shown great potential, they suffer from unstable probability dynamics and mismatch between generated molecule size and the protein pockets geometry, resulting in inconsistent quality and off-target effects. We propose PAFlow, a novel target-aware molecular generation model featuring prior interaction guidance and a learnable atom number predictor. PAFlow adopts the efficient flow matching framework to model the generation process and constructs a new form of conditional flow matching for discrete atom types. A protein-ligand interaction predictor is incorporated to guide the vector field toward higher-affinity regions during generation, while an atom number predictor based on protein pocket information is designed to better align generated molecule size with target geometry. Extensive experiments on the CrossDocked2020 benchmark show that PAFlow achieves a new state-of-the-art in binding affinity (up to -8.31 Avg. Vina Score), simultaneously maintains favorable molecular properties.
Authors: Brennen Hill
Abstract: The advancement of general-purpose intelligent agents is intrinsically linked to the environments in which they are trained. While scaling models and datasets has yielded remarkable capabilities, scaling the complexity, diversity, and interactivity of environments remains a crucial bottleneck. Hand-crafted environments are finite and often contain implicit biases, limiting the potential for agents to develop truly generalizable and robust skills. In this work, we propose a paradigm for generating a boundless and adaptive curriculum of challenges by framing the environment generation process as an adversarial game. We introduce a system where a team of cooperative multi-agent defenders learns to survive against a procedurally generative attacker. The attacker agent learns to produce increasingly challenging configurations of enemy units, dynamically creating novel worlds tailored to exploit the defenders' current weaknesses. Concurrently, the defender team learns cooperative strategies to overcome these generated threats. This co-evolutionary dynamic creates a self-scaling environment where complexity arises organically from the adversarial interaction, providing an effectively infinite stream of novel and relevant training data. We demonstrate that with minimal training, this approach leads to the emergence of complex, intelligent behaviors, such as flanking and shielding by the attacker, and focus-fire and spreading by the defenders. Our findings suggest that adversarial co-evolution is a powerful mechanism for automatically scaling environmental complexity, driving agents towards greater robustness and strategic depth.
Authors: Feng Ruan, Keli Liu, Michael Jordan
Abstract: We study a compositional variant of kernel ridge regression in which the predictor is applied to a coordinate-wise reweighting of the inputs. Formulated as a variational problem, this model provides a simple testbed for feature learning in compositional architectures. From the perspective of variable selection, we show how relevant variables are recovered while noise variables are eliminated. We establish guarantees showing that both global minimizers and stationary points discard noise coordinates when the noise variables are Gaussian distributed. A central finding is that $\ell_1$-type kernels, such as the Laplace kernel, succeed in recovering features contributing to nonlinear effects at stationary points, whereas Gaussian kernels recover only linear ones.
Authors: Xuekai Zhu, Daixuan Cheng, Dinghuai Zhang, Hengli Li, Kaiyan Zhang, Che Jiang, Youbang Sun, Ermo Hua, Yuxin Zuo, Xingtai Lv, Qizheng Zhang, Lin Chen, Fanghao Shao, Bo Xue, Yunchong Song, Zhenjie Yang, Ganqu Cui, Ning Ding, Jianfeng Gao, Xiaodong Liu, Bowen Zhou, Hongyuan Mei, Zhouhan Lin
Abstract: We propose FlowRL: matching the full reward distribution via flow balancing instead of maximizing rewards in large language model (LLM) reinforcement learning (RL). Recent advanced reasoning models adopt reward-maximizing methods (\eg, PPO and GRPO), which tend to over-optimize dominant reward signals while neglecting less frequent but valid reasoning paths, thus reducing diversity. In contrast, we transform scalar rewards into a normalized target distribution using a learnable partition function, and then minimize the reverse KL divergence between the policy and the target distribution. We implement this idea as a flow-balanced optimization method that promotes diverse exploration and generalizable reasoning trajectories. We conduct experiments on math and code reasoning tasks: FlowRL achieves a significant average improvement of $10.0\%$ over GRPO and $5.1\%$ over PPO on math benchmarks, and performs consistently better on code reasoning tasks. These results highlight reward distribution-matching as a key step toward efficient exploration and diverse reasoning in LLM reinforcement learning.
Authors: Zinan Lin, Enshu Liu, Xuefei Ning, Junyi Zhu, Wenyu Wang, Sergey Yekhanin
Abstract: Generative modeling, representation learning, and classification are three core problems in machine learning (ML), yet their state-of-the-art (SoTA) solutions remain largely disjoint. In this paper, we ask: Can a unified principle address all three? Such unification could simplify ML pipelines and foster greater synergy across tasks. We introduce Latent Zoning Network (LZN) as a step toward this goal. At its core, LZN creates a shared Gaussian latent space that encodes information across all tasks. Each data type (e.g., images, text, labels) is equipped with an encoder that maps samples to disjoint latent zones, and a decoder that maps latents back to data. ML tasks are expressed as compositions of these encoders and decoders: for example, label-conditional image generation uses a label encoder and image decoder; image embedding uses an image encoder; classification uses an image encoder and label decoder. We demonstrate the promise of LZN in three increasingly complex scenarios: (1) LZN can enhance existing models (image generation): When combined with the SoTA Rectified Flow model, LZN improves FID on CIFAR10 from 2.76 to 2.59-without modifying the training objective. (2) LZN can solve tasks independently (representation learning): LZN can implement unsupervised representation learning without auxiliary loss functions, outperforming the seminal MoCo and SimCLR methods by 9.3% and 0.2%, respectively, on downstream linear classification on ImageNet. (3) LZN can solve multiple tasks simultaneously (joint generation and classification): With image and label encoders/decoders, LZN performs both tasks jointly by design, improving FID and achieving SoTA classification accuracy on CIFAR10. The code and trained models are available at https://github.com/microsoft/latent-zoning-networks. The project website is at https://zinanlin.me/blogs/latent_zoning_networks.html.
URLs: https://github.com/microsoft/latent-zoning-networks., https://zinanlin.me/blogs/latent_zoning_networks.html.
Authors: Feng Jiang, Amina Mollaysa, Hehuan Ma, Tommaso Mansi, Junzhou Huang, Mangal Prakash, Rui Liao
Abstract: Drug target interaction (DTI) prediction is a cornerstone of computational drug discovery, enabling rational design, repurposing, and mechanistic insights. While deep learning has advanced DTI modeling, existing approaches primarily rely on SMILES protein pairs and fail to exploit the rich multimodal information available for small molecules and proteins. We introduce GRAMDTI, a pretraining framework that integrates multimodal molecular and protein inputs into unified representations. GRAMDTI extends volume based contrastive learning to four modalities, capturing higher-order semantic alignment beyond conventional pairwise approaches. To handle modality informativeness, we propose adaptive modality dropout, dynamically regulating each modality's contribution during pre-training. Additionally, IC50 activity measurements, when available, are incorporated as weak supervision to ground representations in biologically meaningful interaction strengths. Experiments on four publicly available datasets demonstrate that GRAMDTI consistently outperforms state of the art baselines. Our results highlight the benefits of higher order multimodal alignment, adaptive modality utilization, and auxiliary supervision for robust and generalizable DTI prediction.
Authors: Jie Yang, Yifan Hu, Kexin Zhang, Luyang Niu, Philip S. Yu, Kaize Ding
Abstract: Missing values are common in real-world time series, and multivariate time series forecasting with missing values (MTSF-M) has become a crucial area of research for ensuring reliable predictions. To address the challenge of missing data, current approaches have developed an imputation-then-prediction framework that uses imputation modules to fill in missing values, followed by forecasting on the imputed data. However, this framework overlooks a critical issue: there is no ground truth for the missing values, making the imputation process susceptible to errors that can degrade prediction accuracy. In this paper, we conduct a systematic empirical study and reveal that imputation without direct supervision can corrupt the underlying data distribution and actively degrade prediction accuracy. To address this, we propose a paradigm shift that moves away from imputation and directly predicts from the partially observed time series. We introduce Consistency-Regularized Information Bottleneck (CRIB), a novel framework built on the Information Bottleneck principle. CRIB combines a unified-variate attention mechanism with a consistency regularization scheme to learn robust representations that filter out noise introduced by missing values while preserving essential predictive signals. Comprehensive experiments on four real-world datasets demonstrate the effectiveness of CRIB, which predicts accurately even under high missing rates. Our code is available in https://github.com/Muyiiiii/CRIB.
Authors: Max van Spengler, Artem Moskalev, Tommaso Mansi, Mangal Prakash, Rui Liao
Abstract: Language models are increasingly applied to biological sequences like proteins and mRNA, yet their default Euclidean geometry may mismatch the hierarchical structures inherent to biological data. While hyperbolic geometry provides a better alternative for accommodating hierarchical data, it has yet to find a way into language modeling for mRNA sequences. In this work, we introduce HyperHELM, a framework that implements masked language model pre-training in hyperbolic space for mRNA sequences. Using a hybrid design with hyperbolic layers atop Euclidean backbone, HyperHELM aligns learned representations with the biological hierarchy defined by the relationship between mRNA and amino acids. Across multiple multi-species datasets, it outperforms Euclidean baselines on 9 out of 10 tasks involving property prediction, with 10% improvement on average, and excels in out-of-distribution generalization to long and low-GC content sequences; for antibody region annotation, it surpasses hierarchy-aware Euclidean models by 3% in annotation accuracy. Our results highlight hyperbolic geometry as an effective inductive bias for hierarchical language modeling of mRNA sequences.
Authors: Yifang Zhang, Shengwu Xiong, Henan Wang, Wenjie Yin, Jiawang Peng, Yuqiang Zhang, Chen Zhou, Hua Chen, Qile Zhao, Pengfei Duan
Abstract: Precipitation Nowcasting, which aims to predict precipitation within the next 0 to 6 hours, is critical for disaster mitigation and real-time response planning. However, most time series forecasting benchmarks in meteorology are evaluated on variables with strong periodicity, such as temperature and humidity, which fail to reflect model capabilities in more complex and practically meteorology scenarios like precipitation nowcasting. To address this gap, we propose RainfallBench, a benchmark designed for precipitation nowcasting, a highly challenging and practically relevant task characterized by zero inflation, temporal decay, and non-stationarity, focusing on predicting precipitation within the next 0 to 6 hours. The dataset is derived from five years of meteorological observations, recorded at hourly intervals across six essential variables, and collected from more than 140 Global Navigation Satellite System (GNSS) stations globally. In particular, it incorporates precipitable water vapor (PWV), a crucial indicator of rainfall that is absent in other datasets. We further design specialized evaluation protocols to assess model performance on key meteorological challenges, including multi-scale prediction, multi-resolution forecasting, and extreme rainfall events, benchmarking 17 state-of-the-art models across six major architectures on RainfallBench. Additionally, to address the zero-inflation and temporal decay issues overlooked by existing models, we introduce Bi-Focus Precipitation Forecaster (BFPF), a plug-and-play module that incorporates domain-specific priors to enhance rainfall time series forecasting. Statistical analysis and ablation studies validate the comprehensiveness of our dataset as well as the superiority of our methodology.
Authors: Yihang Lu, Xianwei Meng, Enhong Chen
Abstract: Neural Forecasters (NFs) are a cornerstone of Long-term Time Series Forecasting (LTSF). However, progress has been hampered by an overemphasis on architectural complexity at the expense of fundamental forecasting principles. In this work, we return to first principles to redesign the LTSF paradigm. We begin by introducing a Multiple Neural Forecasting Theorem that provides a theoretical basis for our approach. We propose Boosted Direct Output (BDO), a novel forecasting strategy that synergistically combines the advantages of both Auto-Regressive (AR) and Direct Output (DO). In addition, we stabilize the learning process by smoothly tracking the model's parameters. Extensive experiments show that these principled improvements enable a simple MLP to achieve state-of-the-art performance, outperforming recent, complex models in nearly all cases, without any specific considerations in the area. Finally, we empirically verify our theorem, establishing a dynamic performance bound and identifying promising directions for future research. The code for review is available at: .
Authors: Joseph Ramsey, Bryan Andrews, Peter Spirtes
Abstract: Learning causal structure from observational data is especially challenging when latent variables or selection bias are present. The Fast Causal Inference (FCI) algorithm addresses this setting but often performs exhaustive conditional independence tests across many subsets, leading to spurious independence claims, extra or missing edges, and unreliable orientations. We present a family of score-guided mixed-strategy causal search algorithms that build on this tradition. First, we introduce BOSS-FCI and GRaSP-FCI, straightforward variants of GFCI that substitute BOSS or GRaSP for FGES, thereby retaining correctness while incurring different scalability tradeoffs. Second, we develop FCI Targeted-testing (FCIT), a novel mixed-strategy method that improves upon these variants by replacing exhaustive all-subsets testing with targeted tests guided by BOSS, yielding well-formed PAGs with higher precision and efficiency. Finally, we propose a simple heuristic, LV-Dumb (also known as BOSS-POD), which bypasses latent-variable-specific reasoning and directly returns the PAG of the BOSS DAG. Although not strictly correct in the FCI sense, it scales better and often achieves superior accuracy in practice. Simulations and real-data analyses demonstrate that BOSS-FCI and GRaSP-FCI provide sound baselines, FCIT improves both efficiency and reliability, and LV-Dumb offers a practical heuristic with strong empirical performance. Together, these method highlight the value of score-guided and targeted strategies for scalable latent-variable causal discovery.
Authors: Yanjie Zhu, Zhen Zhang, Yunli Wang, Zhiqiang Wang, Yu Li, Rufan Zhou, Shiyang Wen, Peng Jiang, Chenhao Lin, Jian Yang
Abstract: Cascade ranking is a widely adopted paradigm in large-scale information retrieval systems for Top-K item selection. However, the Top-K operator is non-differentiable, hindering end-to-end training. Existing methods include Learning-to-Rank approaches (e.g., LambdaLoss), which optimize ranking metrics like NDCG and suffer from objective misalignment, and differentiable sorting-based methods (e.g., ARF, LCRON), which relax permutation matrices for direct Top-K optimization but introduce gradient conflicts through matrix aggregation. A promising alternative is to directly construct a differentiable approximation of the Top-K selection operator, bypassing the use of soft permutation matrices. However, even state-of-the-art differentiable Top-K operator (e.g., LapSum) require $O(n \log n)$ complexity due to their dependence on sorting for solving the threshold. Thus, we propose DFTopK, a novel differentiable Top-K operator achieving optimal $O(n)$ time complexity. By relaxing normalization constraints, DFTopK admits a closed-form solution and avoids sorting. DFTopK also avoids the gradient conflicts inherent in differentiable sorting-based methods. We evaluate DFTopK on both the public benchmark RecFLow and an industrial system. Experimental results show that DFTopK significantly improves training efficiency while achieving superior performance, which enables us to scale up training samples more efficiently. In the online A/B test, DFTopK yielded a +1.77% revenue lift with the same computational budget compared to the baseline. To the best of our knowledge, this work is the first to introduce differentiable Top-K operators into recommendation systems and the first to achieve theoretically optimal linear-time complexity for Top-K selection. We have open-sourced our implementation to facilitate future research in both academia and industry.
Authors: Emily Alsentzer, Marie-Laure Charpignon, Bill Chen, Niharika D'Souza, Jason Fries, Yixing Jiang, Aparajita Kashyap, Chanwoo Kim, Simon Lee, Aishwarya Mandyam, Ashery Mbilinyi, Nikita Mehandru, Nitish Nagesh, Brighton Nuwagira, Emma Pierson, Arvind Pillai, Akane Sano, Tanveer Syeda-Mahmood, Shashank Yadav, Elias Adhanom, Muhammad Umar Afza, Amelia Archer, Suhana Bedi, Vasiliki Bikia, Trenton Chang, George H. Chen, Winston Chen, Erica Chiang, Edward Choi, Octavia Ciora, Paz Dozie-Nnamah, Shaza Elsharief, Matthew Engelhard, Ali Eshragh, Jean Feng, Josh Fessel, Scott Fleming, Kei Sen Fong, Thomas Frost, Soham Gadgil, Judy Gichoya, Leeor Hershkovich, Sujeong Im, Bhavya Jain, Vincent Jeanselme, Furong Jia, Qixuan Jin, Yuxuan Jin, Daniel Kapash, Geetika Kapoor, Behdokht Kiafar, Matthias Kleiner, Stefan Kraft, Annika Kumar, Daeun Kyung, Zhongyuan Liang, Joanna Lin, Qianchu Liu, Chang Liu, Hongzhou Luan, Chris Lunt, Leopoldo Jul\'ian Lechuga L\'opez, Matthew B. A. McDermott, Shahriar Noroozizadeh, Connor O'Brien, YongKyung Oh, Mixail Ota, Stephen Pfohl, Meagan Pi, Tanmoy Sarkar Pias, Emma Rocheteau, Avishaan Sethi, Toru Shirakawa, Anita Silver, Neha Simha, Kamile Stankeviciute, Max Sunog, Peter Szolovits, Shengpu Tang, Jialu Tang, Aaron Tierney, John Valdovinos, Byron Wallace, Will Ke Wang, Peter Washington, Jeremy Weiss, Daniel Wolfe, Emily Wong, Hye Sun Yun, Xiaoman Zhang, Xiao Yu Cindy Zhang, Hayoung Jeong, Kaveri A. Thakoor
Abstract: The 6th Annual Conference on Health, Inference, and Learning (CHIL 2025), hosted by the Association for Health Learning and Inference (AHLI), was held in person on June 25-27, 2025, at the University of California, Berkeley, in Berkeley, California, USA. As part of this year's program, we hosted Research Roundtables to catalyze collaborative, small-group dialogue around critical, timely topics at the intersection of machine learning and healthcare. Each roundtable was moderated by a team of senior and junior chairs who fostered open exchange, intellectual curiosity, and inclusive engagement. The sessions emphasized rigorous discussion of key challenges, exploration of emerging opportunities, and collective ideation toward actionable directions in the field. In total, eight roundtables were held by 19 roundtable chairs on topics of "Explainability, Interpretability, and Transparency," "Uncertainty, Bias, and Fairness," "Causality," "Domain Adaptation," "Foundation Models," "Learning from Small Medical Data," "Multimodal Methods," and "Scalable, Translational Healthcare Solutions."
Authors: Yang Zhang, Rui Zhang, Jiaming Guo, Lei Huang, Di Huang, Yunpu Zhao, Shuyao Cheng, Pengwei Jin, Chongxiao Li, Zidong Du, Xing Hu, Qi Guo, Yunji Chen
Abstract: The remarkable progress of Large Language Models (LLMs) presents promising opportunities for Verilog code generation which is significantly important for automated circuit design. The lacking of meaningful functional rewards hinders the preference optimization based on Reinforcement Learning (RL) for producing functionally correct Verilog code. In this paper, we propose Signal-Aware Learning for Verilog code generation (QiMeng-SALV) by leveraging code segments of functionally correct output signal to optimize RL training. Considering Verilog code specifies the structural interconnection of hardware gates and wires so that different output signals are independent, the key insight of QiMeng-SALV is to extract verified signal-aware implementations in partially incorrect modules, so as to enhance the extraction of meaningful functional rewards. Roughly, we verify the functional correctness of signals in generated module by comparing with that of reference module in the training data. Then abstract syntax tree (AST) is employed to identify signal-aware code segments which can provide meaningful functional rewards from erroneous modules. Finally, we introduce signal-aware DPO which is optimized on the correct signal-level code segments, thereby preventing noise and interference from incorrect signals. The proposed QiMeng-SALV underscores the paradigm shift from conventional module-level to fine-grained signal-level optimization in Verilog code generation, addressing the issue of insufficient functional rewards. Experiments demonstrate that our method achieves state-of-the-art performance on VerilogEval and RTLLM, with a 7B parameter model matching the performance of the DeepSeek v3 671B model and significantly outperforming the leading open-source model CodeV trained on the same dataset. Our code is available at https://github.com/zy1xxx/SALV.
Authors: Judah Goldfeder, Matthew So, Hod Lipson
Abstract: Auxiliary Learning (AL) is a form of multi-task learning in which a model trains on auxiliary tasks to boost performance on a primary objective. While AL has improved generalization across domains such as navigation, image classification, and NLP, it often depends on human-labeled auxiliary tasks that are costly to design and require domain expertise. Meta-learning approaches mitigate this by learning to generate auxiliary tasks, but typically rely on gradient based bi-level optimization, adding substantial computational and implementation overhead. We propose RL-AUX, a reinforcement-learning (RL) framework that dynamically creates auxiliary tasks by assigning auxiliary labels to each training example, rewarding the agent whenever its selections improve the performance on the primary task. We also explore learning per-example weights for the auxiliary loss. On CIFAR-100 grouped into 20 superclasses, our RL method outperforms human-labeled auxiliary tasks and matches the performance of a prominent bi-level optimization baseline. We present similarly strong results on other classification datasets. These results suggest RL is a viable path to generating effective auxiliary tasks.
Authors: Woojin Cho, Kookjin Lee, Noseong Park, Donsub Rim, Gerrit Welper
Abstract: We present a data-driven dimensionality reduction method that is well-suited for physics-based data representing hyperbolic wave propagation. The method utilizes a specialized neural network architecture called low rank neural representation (LRNR) inside a hypernetwork framework. The architecture is motivated by theoretical results that rigorously prove the existence of efficient representations for this wave class. We illustrate through archetypal examples that such an efficient low-dimensional representation of propagating waves can be learned directly from data through a combination of deep learning techniques. We observe that a low rank tensor representation arises naturally in the trained LRNRs, and that this reveals a new decomposition of wave propagation where each decomposed mode corresponds to interpretable physical features. Furthermore, we demonstrate that the LRNR architecture enables efficient inference via a compression scheme, which is a potentially important feature when deploying LRNRs in demanding performance regimes.
Authors: Wenduan Xu
Abstract: We present a hybrid quantum-classical recurrent neural network (QRNN) architecture in which the recurrent core is realized as a parametrized quantum circuit (PQC) controlled by a classical feedforward network. The hidden state is the quantum state of an $n$-qubit PQC in an exponentially large Hilbert space $\mathbb{C}^{2^n}$, which serves as a coherent recurrent quantum memory. The PQC is unitary by construction, making the hidden-state evolution norm-preserving without external constraints. At each timestep, mid-circuit Pauli expectation-value readouts are combined with the input embedding and processed by the feedforward network, which provides explicit classical nonlinearity. The outputs parametrize the PQC, which updates the hidden state via unitary dynamics. The QRNN is compact and physically consistent, and it unifies (i) unitary recurrence as a high-capacity memory, (ii) partial observation via mid-circuit readouts, and (iii) nonlinear classical control for input-conditioned parametrization. We evaluate the model in simulation with up to 14 qubits on sentiment analysis, MNIST, permuted MNIST, copying memory, and language modeling. For sequence-to-sequence learning, we further devise a soft attention mechanism over the mid-circuit readouts and show its effectiveness for machine translation. To our knowledge, this is the first model (RNN or otherwise) grounded in quantum operations to achieve competitive performance against strong classical baselines across a broad class of sequence-learning tasks.
Authors: Semyon Lomasov, Judah Goldfeder, Mehmet Hamza Erol, Matthew So, Yao Yan, Addison Howard, Nathan Kutz, Ravid Shwartz Ziv
Abstract: Do AI systems truly understand human concepts or merely mimic surface patterns? We investigate this through chess, where human creativity meets precise strategic concepts. Analyzing a 270M-parameter transformer that achieves grandmaster-level play, we uncover a striking paradox: while early layers encode human concepts like center control and knight outposts with up to 85\% accuracy, deeper layers, despite driving superior performance, drift toward alien representations, dropping to 50-65\% accuracy. To test conceptual robustness beyond memorization, we introduce the first Chess960 dataset: 240 expert-annotated positions across 6 strategic concepts. When opening theory is eliminated through randomized starting positions, concept recognition drops 10-20\% across all methods, revealing the model's reliance on memorized patterns rather than abstract understanding. Our layer-wise analysis exposes a fundamental tension in current architectures: the representations that win games diverge from those that align with human thinking. These findings suggest that as AI systems optimize for performance, they develop increasingly alien intelligence, a critical challenge for creative AI applications requiring genuine human-AI collaboration. Dataset and code are available at: https://github.com/slomasov/ChessConceptsLLM.
Authors: Chuyan Chen, Chenyang Ma, Zhangxin Li, Yutong He, Yanjie Dong, Kun Yuan
Abstract: Communication remains a central bottleneck in large-scale distributed machine learning, and gradient sparsification has emerged as a promising strategy to alleviate this challenge. However, existing gradient compressors face notable limitations: Rand-$K$ discards structural information and performs poorly in practice, while Top-$K$ preserves informative entries but loses the contraction property and requires costly All-Gather operations. In this paper, we propose ARC-Top-$K$, an {All-Reduce}-Compatible Top-$K$ compressor that aligns sparsity patterns across nodes using a lightweight sketch of the gradient, enabling index-free All-Reduce while preserving globally significant information. ARC-Top-$K$ is provably contractive and, when combined with momentum error feedback (EF21M), achieves linear speedup and sharper convergence rates than the original EF21M under standard assumptions. Empirically, ARC-Top-$K$ matches the accuracy of Top-$K$ while reducing wall-clock training time by up to 60.7\%, offering an efficient and scalable solution that combines the robustness of Rand-$K$ with the strong performance of Top-$K$.
Authors: Zheng Nie, Peijie Sun
Abstract: Recent advances in large language models (LLMs) promise more effective information extraction for review-based recommender systems, yet current methods still (i) mine free-form reviews without scope control, producing redundant and noisy representations, (ii) lack principled metrics that link LLM hallucination to downstream effectiveness, and (iii) leave the cost-quality trade-off across model scales largely unexplored. We address these gaps with the Hyper-Adaptive Dual-Stage Semantic Framework (HADSF), a two-stage approach that first induces a compact, corpus-level aspect vocabulary via adaptive selection and then performs vocabulary-guided, explicitly constrained extraction of structured aspect-opinion triples. To assess the fidelity of the resulting representations, we introduce Aspect Drift Rate (ADR) and Opinion Fidelity Rate (OFR) and empirically uncover a nonmonotonic relationship between hallucination severity and rating prediction error. Experiments on approximately 3 million reviews across LLMs spanning 1.5B-70B parameters show that, when integrated into standard rating predictors, HADSF yields consistent reductions in prediction error and enables smaller models to achieve competitive performance in representative deployment scenarios. We release code, data pipelines, and metric implementations to support reproducible research on hallucination-aware, LLM-enhanced explainable recommendation. Code is available at https://github.com/niez233/HADSF
Authors: Lei Liu, Zhongyi Yu, Hong Wang, Huanshuo Dong, Haiyang Xin, Hongwei Zhao, Bin Li
Abstract: In recent years, Neural Operators(NO) have gradually emerged as a popular approach for solving Partial Differential Equations (PDEs). However, their application to large-scale engineering tasks suffers from significant computational overhead. And the fact that current models impose a uniform computational cost while physical fields exhibit vastly different complexities constitutes a fundamental mismatch, which is the root of this inefficiency. For instance, in turbulence flows, intricate vortex regions require deeper network processing compared to stable flows. To address this, we introduce a framework: Skip-Block Routing (SBR), a general framework designed for Transformer-based neural operators, capable of being integrated into their multi-layer architectures. First, SBR uses a routing mechanism to learn the complexity and ranking of tokens, which is then applied during inference. Then, in later layers, it decides how many tokens are passed forward based on this ranking. This way, the model focuses more processing capacity on the tokens that are more complex. Experiments demonstrate that SBR is a general framework that seamlessly integrates into various neural operators. Our method reduces computational cost by approximately 50% in terms of Floating Point Operations (FLOPs), while still delivering up to 2x faster inference without sacrificing accuracy.
Authors: Leonhard Duda, Khadijeh Alibabaei, Elena Vollmer, Leon Klug, Valentin Kozlov, Lisana Berberi, Mishal Benz, Rebekka Volk, Juan Pedro Guti\'errez Hermosillo Muriedas, Markus G\"otz, Judith S\'a\'inz-Pardo D\'iaz, \'Alvaro L\'opez Garc\'ia, Frank Schultmann, Achim Streit
Abstract: Federated Learning (FL) is an approach for training a shared Machine Learning (ML) model with distributed training data and multiple participants. FL allows bypassing limitations of the traditional Centralized Machine Learning CL if data cannot be shared or stored centrally due to privacy or technical restrictions -- the participants train the model locally with their training data and do not need to share it among the other participants. This paper investigates the practical implementation and effectiveness of FL in a real-world scenario, specifically focusing on unmanned aerial vehicle (UAV)-based thermal images for common thermal feature detection in urban environments. The distributed nature of the data arises naturally and makes it suitable for FL applications, as images captured in two German cities are available. This application presents unique challenges due to non-identical distribution and feature characteristics of data captured at both locations. The study makes several key contributions by evaluating FL algorithms in real deployment scenarios rather than simulation. We compare several FL approaches with a centralized learning baseline across key performance metrics such as model accuracy, training time, communication overhead, and energy usage. This paper also explores various FL workflows, comparing client-controlled workflows and server-controlled workflows. The findings of this work serve as a valuable reference for understanding the practical application and limitations of the FL methods in segmentation tasks in UAV-based imaging.
Authors: Bernd Bohnet, Rumen Dangovski, Kevin Swersky, Sherry Moore, Arslan Chaudhry, Kathleen Kenealy, Noah Fiedel
Abstract: The remarkable capabilities of Large Language Models (LLMs) often need to be tailored for specific applications, requiring the integration of new knowledge or the acquisition of new skills. While full fine-tuning is a powerful adaptation method, it is computationally expensive and can lead to a degradation of general reasoning abilities, a phenomenon known as catastrophic forgetting. A range of alternative techniques exists, each with its own trade-offs. In-Context Learning (ICL) is fast but limited by context length, while Parameter-Efficient Fine-Tuning (PEFT) methods like Low-Rank Adaptation (LoRA) offer a middle ground by minimizing parameter changes. However, the challenge of catastrophic forgetting persists, raising questions about the best adaptation strategy for a given task. This paper presents a comparative analysis of Supervised Finetuning (SFT), LoRA, and ICL in data-scarce scenarios. We find that LoRA provides the most effective balance, successfully instilling new skills with minimal impact on the base model's general knowledge. In contrast, while SFT excels at skill acquisition, it is highly susceptible to catastrophic forgetting. ICL is effective for incorporating factual knowledge but struggles with complex skills. Our findings offer a practical framework for selecting an LLM adaptation strategy. We highlight the critical distinction between skill acquisition and knowledge integration, clarify the trade-offs between task-specific performance and the preservation of general capabilities.
Authors: Sophie Li, Nicholas Huang, Nayan Saxena, Nina Luo, Vincent Lin, Kevin Zhu, Sunishchal Dev
Abstract: Large language models (LLMs) improve reasoning accuracy when generating multiple candidate solutions at test time, but standard methods like Best-of-N (BoN) incur high computational cost by fully generating all branches. Self-Truncation Best-of-N (ST-BoN) mitigates this by truncating unpromising paths early, but its reliance on consistency-based heuristics is a limitation as it does not directly evaluate branch quality. We present KL-Adjusted Pruned Path Algorithm (KAPPA), an inference-time method that combines Kullback-Leibler divergence, confidence, and entropy into a principled scoring function to guide progressive pruning. By promoting diversity during exploration and selectively eliminating low-scoring branches, KAPPA maintains accuracy while substantially reducing memory and token usage. Experiments on GSM8K and MATH500 with DeepSeek-R1-Distill-Qwen-1.5B and Qwen2.5-7B-Instruct demonstrate that KAPPA stabilizes performance in smaller models and achieves up to ~60% reduction in peak memory and ~90% reduction in total token generation relative to BoN, with minimal impact on accuracy.
Authors: Morgan Lee, Artem Frenk, Eamon Worden, Karish Gupta, Thinh Pham, Ethan Croteau, Neil Heffernan
Abstract: Knowledge Tracing (KT) has been an established problem in the educational data mining field for decades, and it is commonly assumed that the underlying learning process being modeled remains static. Given the ever-changing landscape of online learning platforms (OLPs), we investigate how concept drift and changing student populations can impact student behavior within an OLP through testing model performance both within a single academic year and across multiple academic years. Four well-studied KT models were applied to five academic years of data to assess how susceptible KT models are to concept drift. Through our analysis, we find that all four families of KT models can exhibit degraded performance, Bayesian Knowledge Tracing (BKT) remains the most stable KT model when applied to newer data, while more complex, attention based models lose predictive power significantly faster.
Authors: Jungyeon Koh, Hyun Jong Yang
Abstract: The growing demand for on-device large language model (LLM) inference highlights the need for efficient mobile edge computing (MEC) solutions, especially in resource-constrained settings. Speculative decoding offers a promising solution by partitioning token generation between a lightweight draft model on mobile devices and a powerful target model on edge servers, but suffers from communication overhead and asynchronous delays. This paper is the first to propose a unified framework that jointly optimizes user association and resource allocation (UARA) to support efficient parallel speculative decoding. We solve the UARA problem using a multi-agent deep reinforcement learning algorithm. To evaluate our approach under realistic conditions, we conduct experiments using the Sionna simulator. Results show that our method achieves up to 28.0% and an average of 23.7% reduction in end-to-end latency without compromising inference accuracy, enabling scalable and low-latency LLM services in MEC systems.
Authors: Jay Mohta, Kenan Emir Ak, Dimitrios Dimitriadis, Yan Xu, Mingwei Shen
Abstract: Vision-Language Models (VLMs) suffer from catastrophic forgetting when sequentially fine-tuned on new tasks, degrading performance on previously learned foundational and task-specific capabilities. While multi-task learning can mitigate forgetting, it requires simultaneous access to all datasets and imposes computational overhead that scales linearly with the number of tasks. In this work, we introduce a routing-based approach that enables the integration of new tasks while preserving the foundational knowledge acquired during pretraining. We evaluate our method using InternVL-2 models (2B and 8B parameters) and demonstrate that routing preserves the model's foundational capabilities by maintaining performance on general-purpose benchmarks such as ChartQA, MMBench, and DocVQA, while simultaneously improving accuracy on specialized tasks. Importantly, our approach achieves this without requiring concurrent access to data from all tasks, avoiding the significant computational and data overhead associated with traditional multi-task learning. We further conduct extensive ablation studies to evaluate the scalability and robustness of routing-based learning, showing that the approach is resilient to a growing number of tasks and performs particularly well when new tasks are semantically related. Finally, we show that the routing mechanism enables superior cross-modal transfer between language and vision capabilities, allowing knowledge learned in one modality to enhance performance in another capability not achieved by existing continual learning methods.
Authors: Omid Askarisichani, Elizabeth Y. Huang, Abed K. Musaffar, Noah E. Friedkin, Francesco Bullo, Ambuj K. Singh
Abstract: Discovering the antecedents of individuals' influence in collaborative environments is an important, practical, and challenging problem. In this paper, we study interpersonal influence in small groups of individuals who collectively execute a sequence of intellective tasks. We observe that along an issue sequence with feedback, individuals with higher expertise and social confidence are accorded higher interpersonal influence. We also observe that low-performing individuals tend to underestimate their high-performing teammate's expertise. Based on these observations, we introduce three hypotheses and present empirical and theoretical support for their validity. We report empirical evidence on longstanding theories of transactive memory systems, social comparison, and confidence heuristics on the origins of social influence. We propose a cognitive dynamical model inspired by these theories to describe the process by which individuals adjust interpersonal influences over time. We demonstrate the model's accuracy in predicting individuals' influence and provide analytical results on its asymptotic behavior for the case with identically performing individuals. Lastly, we propose a novel approach using deep neural networks on a pre-trained text embedding model for predicting the influence of individuals. Using message contents, message times, and individual correctness collected during tasks, we are able to accurately predict individuals' self-reported influence over time. Extensive experiments verify the accuracy of the proposed models compared to baselines such as structural balance and reflected appraisal model. While the neural networks model is the most accurate, the dynamical model is the most interpretable for influence prediction.
Authors: Sabee Grewal, Vishnu Iyer, William Kretschmer, Daniel Liang
Abstract: We give a pair of algorithms that efficiently learn a quantum state prepared by Clifford gates and $O(\log n)$ non-Clifford gates. Specifically, for an $n$-qubit state $|\psi\rangle$ prepared with at most $t$ non-Clifford gates, our algorithms use $\mathsf{poly}(n,2^t,1/\varepsilon)$ time and copies of $|\psi\rangle$ to learn $|\psi\rangle$ to trace distance at most $\varepsilon$. The first algorithm for this task is more efficient, but requires entangled measurements across two copies of $|\psi\rangle$. The second algorithm uses only single-copy measurements at the cost of polynomial factors in runtime and sample complexity. Our algorithms more generally learn any state with sufficiently large stabilizer dimension, where a quantum state has stabilizer dimension $k$ if it is stabilized by an abelian group of $2^k$ Pauli operators. We also develop an efficient property testing algorithm for stabilizer dimension, which may be of independent interest.
Authors: Shivam Garg, Chirag Pabbaraju, Kirankumar Shiragur, Gregory Valiant
Abstract: We examine the extent to which sublinear-sample property testing and estimation apply to settings where samples are independently but not identically distributed. Specifically, we consider the following distributional property testing framework: Suppose there is a set of distributions over a discrete support of size $k$, $p_1, p_2,\ldots,p_T$, and we obtain $c$ independent draws from each distribution. Suppose the goal is to learn or test a property of the average distribution, $p_{avg}$. This setup models a number of important practical settings where the individual distributions correspond to heterogeneous entities -- either individuals, chronologically distinct time periods, spatially separated data sources, etc. From a learning standpoint, even with $c=1$ samples from each distribution, $\Theta(k/\varepsilon^2)$ samples are necessary and sufficient to learn $p_{avg}$ to within error $\varepsilon$ in $\ell_1$ distance. To test uniformity or identity -- distinguishing the case that $p_{avg}$ is equal to some reference distribution, versus has $\ell_1$ distance at least $\varepsilon$ from the reference distribution, we show that a linear number of samples in $k$ is necessary given $c=1$ samples from each distribution. In contrast, for $c \ge 2$, we recover the usual sublinear sample testing guarantees of the i.i.d.\ setting: we show that $O(\sqrt{k}/\varepsilon^2 + 1/\varepsilon^4)$ total samples are sufficient, matching the optimal sample complexity in the i.i.d.\ case in the regime where $\varepsilon \ge k^{-1/4}$. Additionally, we show that in the $c=2$ case, there is a constant $\rho > 0$ such that even in the linear regime with $\rho k$ samples, no tester that considers the multiset of samples (ignoring which samples were drawn from the same $p_i$) can perform uniformity testing. We also extend our techniques to the problem of testing "closeness" of two distributions.
Authors: Xiaofei Wang, Xingxu Huang, Stephen J. Price, Chao Li
Abstract: The recent advancement of spatial transcriptomics (ST) allows to characterize spatial gene expression within tissue for discovery research. However, current ST platforms suffer from low resolution, hindering in-depth understanding of spatial gene expression. Super-resolution approaches promise to enhance ST maps by integrating histology images with gene expressions of profiled tissue spots. However, current super-resolution methods are limited by restoration uncertainty and mode collapse. Although diffusion models have shown promise in capturing complex interactions between multi-modal conditions, it remains a challenge to integrate histology images and gene expression for super-resolved ST maps. This paper proposes a cross-modal conditional diffusion model for super-resolving ST maps with the guidance of histology images. Specifically, we design a multi-modal disentangling network with cross-modal adaptive modulation to utilize complementary information from histology images and spatial gene expression. Moreover, we propose a dynamic cross-attention modelling strategy to extract hierarchical cell-to-tissue information from histology images. Lastly, we propose a co-expression-based gene-correlation graph network to model the co-expression relationship of multiple genes. Experiments show that our method outperforms other state-of-the-art methods in ST super-resolution on three public datasets.
Authors: H\'edi Hadiji (UvA), Sarah Sachs (UvA), Crist\'obal Guzm\'an (PUC)
Abstract: Tracking the solution of time-varying variational inequalities is an important problem with applications in game theory, optimization, and machine learning. Existing work considers time-varying games or time-varying optimization problems. For strongly convex optimization problems or strongly monotone games, these results provide tracking guarantees under the assumption that the variation of the time-varying problem is restrained, that is, problems with a sublinear solution path. In this work we extend existing results in two ways: In our first result, we provide tracking bounds for (1) variational inequalities with a sublinear solution path but not necessarily monotone functions, and (2) for periodic time-varying variational inequalities that do not necessarily have a sublinear solution path-length. Our second main contribution is an extensive study of the convergence behavior and trajectory of discrete dynamical systems of periodic time-varying VI. We show that these systems can exhibit provably chaotic behavior or can converge to the solution. Finally, we illustrate our theoretical results with experiments.
Authors: Yahya Saleh, Armin Iske
Abstract: Let $C_h$ be a composition operator mapping $L^2(\Omega_1)$ into $L^2(\Omega_2)$ for some open sets $\Omega_1, \Omega_2 \subseteq \mathbb{R}^n$. We characterize the mappings $h$ that transform Riesz bases of $L^2(\Omega_1)$ into Riesz bases of $L^2(\Omega_2)$. Restricting our analysis to differentiable mappings, we demonstrate that mappings $h$ that preserve Riesz bases have Jacobian determinants that are bounded away from zero and infinity. We discuss implications of these results for approximation theory, highlighting the potential of using bijective neural networks to construct Riesz bases with favorable approximation properties.
Authors: Matthieu Lin, Jenny Sheng, Andrew Zhao, Shenzhi Wang, Yang Yue, Victor Shea Jay Huang, Huan Liu, Jun Liu, Gao Huang, Yong-Jin Liu
Abstract: This survey organizes the intricate literature on the design and optimization of emerging structures around post-trained LMs. We refer to this overarching structure as scaffolded LMs and focus on LMs that are integrated into multi-step processes with tools. We view scaffolded LMs as semi-parametric models wherein we train non-parametric variables, including the prompt, tools, and scaffold's code. In particular, they interpret instructions, use tools, and receive feedback all in language. Recent works use an LM as an optimizer to interpret language supervision and update non-parametric variables according to intricate objectives. In this survey, we refer to this paradigm as training of scaffolded LMs with language supervision. A key feature of non-parametric training is the ability to learn from language. Parametric training excels in learning from demonstration (supervised learning), exploration (reinforcement learning), or observations (unsupervised learning), using well-defined loss functions. Language-based optimization enables rich, interpretable, and expressive objectives, while mitigating issues like catastrophic forgetting and supporting compatibility with closed-source models. Furthermore, agents are increasingly deployed as co-workers in real-world applications such as Copilot in Office tools or software development. In these mixed-autonomy settings, where control and decision-making are shared between human and AI, users point out errors or suggest corrections. Accordingly, we discuss agents that continuously improve by learning from this real-time, language-based feedback and refer to this setting as streaming learning from language supervision.
Authors: Jacopo D'Ignazi, Andreas Kaltenbrunner, Yelena Mejova, Michele Tizzani, Kyriaki Kalimeri, Mariano Beir\'o, Pablo Arag\'on
Abstract: Over the last few years, verifying the credibility of information sources has become a fundamental need to combat disinformation. Here, we present a language-agnostic model designed to assess the reliability of web domains as sources in references across multiple language editions of Wikipedia. Utilizing editing activity data, the model evaluates domain reliability within different articles of varying controversiality, such as Climate Change, COVID-19, History, Media, and Biology topics. Crafting features that express domain usage across articles, the model effectively predicts domain reliability, achieving an F1 Macro score of approximately 0.80 for English and other high-resource languages. For mid-resource languages, we achieve 0.65, while the performance of low-resource languages varies. In all cases, the time the domain remains present in the articles (which we dub as permanence) is one of the most predictive features. We highlight the challenge of maintaining consistent model performance across languages of varying resource levels and demonstrate that adapting models from higher-resource languages can improve performance. We believe these findings can assist Wikipedia editors in their ongoing efforts to verify citations and may offer useful insights for other user-generated content communities.
Authors: Yu-Han Huang, Argyrios Gerogiannis, Subhonmesh Bose, Venugopal V. Veeravalli
Abstract: Conventional Multi-Armed Bandit (MAB) algorithms are designed for stationary environments, where the reward distributions associated with the arms do not change with time. In many applications, however, the environment is more accurately modeled as being non-stationary. In this work, piecewise stationary MAB (PS-MAB) environments are investigated, in which the reward distributions associated with a subset of the arms change at some change-points and remain stationary between change-points. Our focus is on the asymptotic analysis of PS-MABs, for which practical algorithms based on change detection have been previously proposed. Our goal is to modularize the design and analysis of such Detection Augmented Bandit (DAB) procedures. To this end, we first provide novel, improved performance lower bounds for PS-MABs. Then, we identify the requirements for stationary bandit algorithms and change detectors in a DAB procedure that are needed for the modularization. We assume that the rewards are sub-Gaussian. Under this assumption and a condition on the separation of the change-points, we show that the analysis of DAB procedures can indeed be modularized, so that the regret bounds can be obtained in a unified manner for various combinations of change detectors and bandit algorithms. Through this analysis, we develop new modular DAB procedures that are order-optimal. Finally, we showcase the practical effectiveness of our modular DAB approach in our experiments, studying its regret performance compared to other methods and investigating its detection capabilities.
Authors: Sengim Karayal\c{c}in, Marina Kr\v{c}ek, Stjepan Picek
Abstract: Side-channel analysis (SCA) poses a real-world threat by exploiting unintentional physical signals to extract secret information from secure devices. Evaluation labs also use the same techniques to certify device security. In recent years, deep learning has emerged as a prominent method for SCA, achieving state-of-the-art attack performance at the cost of interpretability. Understanding how neural networks extract secrets is crucial for security evaluators aiming to defend against such attacks, as only by understanding the attack can one propose better countermeasures. In this work, we apply mechanistic interpretability to neural networks trained for SCA, revealing \textit{how} models exploit \textit{what} leakage in side-channel traces. We focus on sudden jumps in performance to reverse engineer learned representations, ultimately recovering secret masks and moving the evaluation process from black-box to white-box. Our results show that mechanistic interpretability can scale to realistic SCA settings, even when relevant inputs are sparse, model accuracies are low, and side-channel protections prevent standard input interventions.
Authors: Cristobal Donoso-Oliva, Ignacio Becker, Pavlos Protopapas, Guillermo Cabrera-Vives, Martina C\'adiz-Leyton, Daniel Moreno-Cartagena
Abstract: Foundational models have emerged as a powerful paradigm in deep learning field, leveraging their capacity to learn robust representations from large-scale datasets and effectively to diverse downstream applications such as classification. In this paper, we present Astromer 2 a foundational model specifically designed for extracting light curve embeddings. We introduce Astromer 2 as an enhanced iteration of our self-supervised model for light curve analysis. This paper highlights the advantages of its pre-trained embeddings, compares its performance with that of its predecessor, Astromer 1, and provides a detailed empirical analysis of its capabilities, offering deeper insights into the model's representations. Astromer 2 is pretrained on 1.5 million single-band light curves from the MACHO survey using a self-supervised learning task that predicts randomly masked observations within sequences. Fine-tuning on a smaller labeled dataset allows us to assess its performance in classification tasks. The quality of the embeddings is measured by the F1 score of an MLP classifier trained on Astromer-generated embeddings. Our results demonstrate that Astromer 2 significantly outperforms Astromer 1 across all evaluated scenarios, including limited datasets of 20, 100, and 500 samples per class. The use of weighted per-sample embeddings, which integrate intermediate representations from Astromer's attention blocks, is particularly impactful. Notably, Astromer 2 achieves a 15% improvement in F1 score on the ATLAS dataset compared to prior models, showcasing robust generalization to new datasets. This enhanced performance, especially with minimal labeled data, underscores the potential of Astromer 2 for more efficient and scalable light curve analysis.
Authors: Tansheng Zhu, Hongyu Zhou, Ke Jin, Xusheng Xu, Qiufan Yuan, Lijie Ji
Abstract: Bayesian optimization is highly effective for optimizing expensive-to-evaluate black-box functions, but it faces significant computational challenges due to the high computational complexity of Gaussian processes, which results in a total time complexity that is quartic with respect to the number of iterations. To address this limitation, we propose the Bayesian Optimization by Kernel regression and density-based Exploration (BOKE) algorithm. BOKE uses kernel regression for efficient function approximation, kernel density for exploration, and integrates them into the confidence bound criteria to guide the optimization process, thus reducing computational costs to quadratic. Our theoretical analysis rigorously establishes the global convergence of BOKE and ensures its robustness in noisy settings. Through extensive numerical experiments on both synthetic and real-world optimization tasks, we demonstrate that BOKE not only performs competitively compared to Gaussian process-based methods and several other baseline methods but also exhibits superior computational efficiency. These results highlight BOKE's effectiveness in resource-constrained environments, providing a practical approach for optimization problems in engineering applications.
Authors: Eduardo Adame, Daniel Csillag, Guilherme Tegoni Goedert
Abstract: The increasing use of generative ML foundation models for image restoration tasks such as super-resolution calls for robust and interpretable uncertainty quantification methods. We address this need by presenting a novel approach based on conformal prediction techniques to create a 'confidence mask' capable of reliably and intuitively communicating where the generated image can be trusted. Our method is adaptable to any black-box generative model, including those locked behind an opaque API, requires only easily attainable data for calibration, and is highly customizable via the choice of a local image similarity metric. We prove strong theoretical guarantees for our method that span fidelity error control (according to our local image similarity metric), reconstruction quality, and robustness in the face of data leakage. Finally, we empirically evaluate these results and establish our method's solid performance.
Authors: Chris Zhuang, Debadyuti Mukherjee, Yingzhou Lu, Tianfan Fu, Ruqi Zhang
Abstract: Molecular discovery has brought great benefits to the chemical industry. Various molecule design techniques are developed to identify molecules with desirable properties. Traditional optimization methods, such as genetic algorithms, continue to achieve state-of-the-art results across multiple molecular design benchmarks. However, these techniques rely solely on random walk exploration, which hinders both the quality of the final solution and the convergence speed. To address this limitation, we propose a novel approach called Gradient Genetic Algorithm (Gradient GA), which incorporates gradient information from the objective function into genetic algorithms. Instead of random exploration, each proposed sample iteratively progresses toward an optimal solution by following the gradient direction. We achieve this by designing a differentiable objective function parameterized by a neural network and utilizing the Discrete Langevin Proposal to enable gradient guidance in discrete molecular spaces. Experimental results demonstrate that our method significantly improves both convergence speed and solution quality, outperforming cutting-edge techniques. For example, it achieves up to a 25% improvement in the top-10 score over the vanilla genetic algorithm. The code is publicly available at https://github.com/debadyuti23/GradientGA.
Authors: Kangning Cui, Rongkun Zhu, Manqi Wang, Wei Tang, Gregory D. Larsen, Victor P. Pauca, Sarra Alqahtani, Fan Yang, David Segurado, David Lutz, Jean-Michel Morel, Miles R. Silman
Abstract: Palms are ecologically and economically indicators of tropical forest health, biodiversity, and human impact that support local economies and global forest product supply chains. While palm detection in plantations is well-studied, efforts to map naturally occurring palms in dense forests remain limited by overlapping crowns, uneven shading, and heterogeneous landscapes. We develop PRISM (Processing, Inference, Segmentation, and Mapping), a flexible pipeline for detecting and localizing palms in dense tropical forests using large orthomosaic images. Orthomosaics are created from thousands of aerial images and spanning several to hundreds of gigabytes. Our contributions are threefold. First, we construct a large UAV-derived orthomosaic dataset collected across 21 ecologically diverse sites in western Ecuador, annotated with 8,830 bounding boxes and 5,026 palm center points. Second, we evaluate multiple state-of-the-art object detectors based on efficiency and performance, integrating zero-shot SAM 2 as the segmentation backbone, and refining the results for precise geographic mapping. Third, we apply calibration methods to align confidence scores with IoU and explore saliency maps for feature explainability. Though optimized for palms, PRISM is adaptable for identifying other natural objects, such as eastern white pines. Future work will explore transfer learning for lower-resolution datasets (0.5 to 1m).
Authors: Yunfei Wang, Ruoxi Jiang, Yingda Fan, Xiaowei Jia, Jens Eisert, Junyu Liu, Jin-Peng Liu
Abstract: A diffusion probabilistic model (DPM) is a generative model renowned for its ability to produce high-quality outputs in tasks such as image and audio generation. However, training DPMs on large, high-dimensional datasets such as high-resolution images or audio incurs significant computational, energy, and hardware costs. In this work, we introduce efficient quantum algorithms for implementing DPMs through various quantum ODE solvers. These algorithms highlight the potential of quantum Carleman linearization for diverse mathematical structures, leveraging state-of-the-art quantum linear system solvers (QLSS) or linear combination of Hamiltonian simulations (LCHS). Specifically, we focus on two approaches: DPM-solver-$k$ which employs exact $k$-th order derivatives to compute a polynomial approximation of $\epsilon_\theta(x_\lambda,\lambda)$; and UniPC which uses finite difference of $\epsilon_\theta(x_\lambda,\lambda)$ at different points $(x_{s_m}, \lambda_{s_m})$ to approximate higher-order derivatives. As such, this work represents one of the most direct and pragmatic applications of quantum algorithms to large-scale machine learning models, presumably taking substantial steps towards demonstrating the practical utility of quantum computing.
Authors: Zhihao Zhan, Wang Pang, Xiang Zhu, Yechao Bai
Abstract: In this work, we rethink the approach to video super-resolution by introducing a method based on the Diffusion Posterior Sampling framework, combined with an unconditional video diffusion transformer operating in latent space. The video generation model, a diffusion transformer, functions as a space-time model. We argue that a powerful model, which learns the physics of the real world, can easily handle various kinds of motion patterns as prior knowledge, thus eliminating the need for explicit estimation of optical flows or motion parameters for pixel alignment. Furthermore, a single instance of the proposed video diffusion transformer model can adapt to different sampling conditions without re-training. Empirical results on synthetic and real-world datasets illustrate the feasibility of diffusion-based, alignment-free video super-resolution.
Authors: Natalia Cherezova, Artur Jutman, Maksim Jenihhin
Abstract: The emergence of Deep Neural Networks (DNNs) in mission- and safety-critical applications brings their reliability to the front. High performance demands of DNNs require the use of specialized hardware accelerators. Systolic array architecture is widely used in DNN accelerators due to its parallelism and regular structure. This work presents a run-time reconfigurable systolic array architecture with three execution modes and four implementation options. All four implementations are evaluated in terms of resource utilization, throughput, and fault tolerance improvement. The proposed architecture is used for reliability enhancement of DNN inference on systolic array through heterogeneous mapping of different network layers to different execution modes. The approach is supported by a novel reliability assessment method based on fault propagation analysis. It is used for the exploration of the appropriate execution mode--layer mapping for DNN inference. The proposed architecture efficiently protects registers and MAC units of systolic array PEs from transient and permanent faults. The reconfigurability feature enables a speedup of up to $3\times$, depending on layer vulnerability. Furthermore, it requires $6\times$ fewer resources compared to static redundancy and $2.5\times$ fewer resources compared to the previously proposed solution for transient faults.
Authors: Denan Li, Jiyuan Yang, Xiangkai Chen, Lintao Yu, Shi Liu
Abstract: Machine learning force fields have emerged as promising tools for molecular dynamics (MD) simulations, potentially offering quantum-mechanical accuracy with the efficiency of classical MD. Inspired by foundational large language models, recent years have seen considerable progress in developing foundational atomistic models, sometimes referred to as universal force fields, designed to cover most elements in the periodic table. This Perspective adopts a practitioner's viewpoint to ask a critical question: Are these foundational atomistic models reliable for one of their most compelling applications, in particular simulating finite-temperature dynamics? Instead of a broad benchmark, we use the canonical ferroelectric-paraelectric phase transition in PbTiO$_3$ as a focused case study to evaluate prominent foundational atomistic models. Our findings suggest a potential disconnect between static accuracy and dynamic reliability. While 0 K properties are often well-reproduced, we observed that the models can struggle to consistently capture the correct phase transition, sometimes exhibiting simulation instabilities. We believe these challenges may stem from inherent biases in training data and a limited description of anharmonicity. These observed shortcomings, though demonstrated on a single system, appear to point to broader, systemic challenges that can be addressed with targeted fine-tuning. This Perspective serves not to rank models, but to initiate a crucial discussion on the practical readiness of foundational atomistic models and to explore future directions for their improvement.
Authors: Jian-Jian Jiang, Xiao-Ming Wu, Yi-Xiang He, Ling-An Zeng, Yi-Lin Wei, Dandan Zhang, Wei-Shi Zheng
Abstract: Bimanual robotic manipulation is an emerging and critical topic in the robotics community. Previous works primarily rely on integrated control models that take the perceptions and states of both arms as inputs to directly predict their actions. However, we think bimanual manipulation involves not only coordinated tasks but also various uncoordinated tasks that do not require explicit cooperation during execution, such as grasping objects with the closest hand, which integrated control frameworks ignore to consider due to their enforced cooperation in the early inputs. In this paper, we propose a novel decoupled interaction framework that considers the characteristics of different tasks in bimanual manipulation. The key insight of our framework is to assign an independent model to each arm to enhance the learning of uncoordinated tasks, while introducing a selective interaction module that adaptively learns weights from its own arm to improve the learning of coordinated tasks. Extensive experiments on seven tasks in the RoboTwin dataset demonstrate that: (1) Our framework achieves outstanding performance, with a 23.5% boost over the SOTA method. (2) Our framework is flexible and can be seamlessly integrated into existing methods. (3) Our framework can be effectively extended to multi-agent manipulation tasks, achieving a 28% boost over the integrated control SOTA. (4) The performance boost stems from the decoupled design itself, surpassing the SOTA by 16.5% in success rate with only 1/6 of the model size.
Authors: Lukas Aichberger, Alasdair Paren, Guohao Li, Philip Torr, Yarin Gal, Adel Bibi
Abstract: Recent advances in operating system (OS) agents have enabled vision-language models (VLMs) to directly control a user's computer. Unlike conventional VLMs that passively output text, OS agents autonomously perform computer-based tasks in response to a single user prompt. OS agents do so by capturing, parsing, and analysing screenshots and executing low-level actions via application programming interfaces (APIs), such as mouse clicks and keyboard inputs. This direct interaction with the OS significantly raises the stakes, as failures or manipulations can have immediate and tangible consequences. In this work, we uncover a novel attack vector against these OS agents: Malicious Image Patches (MIPs), adversarially perturbed screen regions that, when captured by an OS agent, induce it to perform harmful actions by exploiting specific APIs. For instance, a MIP can be embedded in a desktop wallpaper or shared on social media to cause an OS agent to exfiltrate sensitive user data. We show that MIPs generalise across user prompts and screen configurations, and that they can hijack multiple OS agents even during the execution of benign instructions. These findings expose critical security vulnerabilities in OS agents that have to be carefully addressed before their widespread deployment.
Authors: Jonathan Sauder, Viktor Domazetoski, Guilhem Banc-Prandi, Gabriela Perna, Anders Meibom, Devis Tuia
Abstract: Coral reefs are declining worldwide due to climate change and local stressors. To inform effective conservation or restoration, monitoring at the highest possible spatial and temporal resolution is necessary. Conventional coral reef surveying methods are limited in scalability due to their reliance on expert labor time, motivating the use of computer vision tools to automate the identification and abundance estimation of live corals from images. However, the design and evaluation of such tools has been impeded by the lack of large high quality datasets. We release the Coralscapes dataset, the first general-purpose dense semantic segmentation dataset for coral reefs, covering 2075 images, 39 benthic classes, and 174k segmentation masks annotated by experts. Coralscapes has a similar scope and the same structure as the widely used Cityscapes dataset for urban scene segmentation, allowing benchmarking of semantic segmentation models in a new challenging domain which requires expert knowledge to annotate. We benchmark a wide range of semantic segmentation models, and find that transfer learning from Coralscapes to existing smaller datasets consistently leads to state-of-the-art performance. Coralscapes will catalyze research on efficient, scalable, and standardized coral reef surveying methods based on computer vision, and holds the potential to streamline the development of underwater ecological robotics.
Authors: Emanuele Mezzi, Fabio Massacci, Katja Tuma
Abstract: Several recent works have argued that Large Language Models (LLMs) can be used to tame the data deluge in the cybersecurity field, by improving the automation of Cyber Threat Intelligence (CTI) tasks. This work presents an evaluation methodology that other than allowing to test LLMs on CTI tasks when using zero-shot learning, few-shot learning and fine-tuning, also allows to quantify their consistency and their confidence level. We run experiments with three state-of-the-art LLMs and a dataset of 350 threat intelligence reports and present new evidence of potential security risks in relying on LLMs for CTI. We show how LLMs cannot guarantee sufficient performance on real-size reports while also being inconsistent and overconfident. Few-shot learning and fine-tuning only partially improve the results, thus posing doubts about the possibility of using LLMs for CTI scenarios, where labelled datasets are lacking and where confidence is a fundamental factor.
Authors: Reza Esfandiarpoor, George Zerveas, Ruochen Zhang, Macton Mgonzo, Carsten Eickhoff, Stephen H. Bach
Abstract: Although synthetic data has changed various aspects of information retrieval (IR) pipelines, the main training paradigm remains: contrastive learning with binary relevance labels, where one positive document is compared against several negatives using the InfoNCE loss. This objective treats all documents that are not explicitly annotated as relevant on an equally negative footing, regardless of their actual degree of relevance, thus missing subtle nuances useful for ranking. To overcome this limitation, in this work, we forgo real documents and annotations and use large language models to directly generate synthetic documents that answer the MS MARCO queries according to several different levels of relevance. We also propose using Wasserstein distance as a more effective loss function for training transformer-based retrievers with graduated relevance labels. Our experiments on MS MARCO and BEIR benchmark show that our proposed approach outperforms conventional training with InfoNCE by a large margin. Without using any real documents, our method significantly improves self-supervised retrievers and is more robust to distribution shift compared to contrastive learning using real data. Our method also successfully integrates existing real data into the synthetic ranking context, further boosting the performance. Overall, we show that generating multi-level ranking contexts is a better approach to synthetic data generation for IR than just generating the standard positive and negative documents.
Authors: Mat\'eo Mahaut, Francesca Franzon
Abstract: Large Language Models (LLMs) can sometimes degrade into repetitive loops, persistently generating identical word sequences. Because repetition is rare in natural human language, its frequent occurrence across diverse tasks and contexts in LLMs remains puzzling. Here we investigate whether behaviorally similar repetition patterns arise from distinct underlying mechanisms and how these mechanisms develop during model training. We contrast two conditions: repetitions elicited by natural text prompts with those induced by in-context learning (ICL) setups that explicitly require copying behavior. Our analyses reveal that ICL-induced repetition relies on a dedicated network of attention heads that progressively specialize over training, whereas naturally occurring repetition emerges early and lacks a defined circuitry. Attention inspection further shows that natural repetition focuses disproportionately on low-information tokens, suggesting a fallback behavior when relevant context cannot be retrieved. These results indicate that superficially similar repetition behaviors originate from qualitatively different internal processes, reflecting distinct modes of failure and adaptation in language models.
Authors: Adriano Vinhas, Jo\~ao Correia, Penousal Machado
Abstract: The number of studies that combine Evolutionary Machine Learning and self-supervised learning has been growing steadily in recent years. Evolutionary Machine Learning has been shown to help automate the design of machine learning algorithms and to lead to more reliable solutions. Self-supervised learning, on the other hand, has produced good results in learning useful features when labelled data is limited. This suggests that the combination of these two areas can help both in shaping evolutionary processes and in automating the design of deep neural networks, while also reducing the need for labelled data. Still, there are no detailed reviews that explain how Evolutionary Machine Learning and self-supervised learning can be used together. To help with this, we provide an overview of studies that bring these areas together. Based on this growing interest and the range of existing works, we suggest a new sub-area of research, which we call Evolutionary Self-Supervised Learning and introduce a taxonomy for it. Finally, we point out some of the main challenges and suggest directions for future research to help Evolutionary Self-Supervised Learning grow and mature as a field.
Authors: Peihong Zhang, Zhixin Li, Rui Sang, Yuxuan Liu, Yiqiang Cai, Yizhou Tan, Shengchen Li
Abstract: The coupling signal refers to a latent physiological signal that characterizes the transformation from cardiac electrical excitation, captured by the electrocardiogram (ECG), to mechanical contraction, recorded by the phonocardiogram (PCG). By encoding the temporal and functional interplay between electrophysiological and hemodynamic events, it serves as an intrinsic link between modalities and offers a unified representation of cardiac function, with strong potential to enhance multi-modal cardiovascular disease (CVD) detection. However, existing coupling signal estimation methods remain highly vulnerable to noise, particularly in real-world clinical and physiological settings, which undermines their robustness and limits practical value. In this study, we propose Noise-Robust Multi-Modal Coupling Signal Estimation (NMCSE), which reformulates coupling signal estimation as a distribution matching problem solved via optimal transport. By jointly aligning amplitude and timing, NMCSE avoids noise amplification and enables stable signal estimation. When integrated into a Temporal-Spatial Feature Extraction (TSFE) network, the estimated coupling signal effectively enhances multi-modal fusion for more accurate CVD detection. To evaluate robustness under real-world conditions, we design two complementary experiments targeting distinct sources of noise. The first uses the PhysioNet 2016 dataset with simulated hospital noise to assess the resilience of NMCSE to clinical interference. The second leverages the EPHNOGRAM dataset with motion-induced physiological noise to evaluate intra-state estimation stability across activity levels. Experimental results show that NMCSE consistently outperforms existing methods under both clinical and physiological noise, highlighting it as a noise-robust estimation approach that enables reliable multi-modal cardiac detection in real-world conditions.
Authors: Binh H. Ho, Long Nguyen Chi, TrungTin Nguyen, Binh T. Nguyen, Van Ha Hoang, Christopher Drovandi
Abstract: Model-based clustering integrated with variable selection is a powerful tool for uncovering latent structures within complex data. However, its effectiveness is often hindered by challenges such as identifying relevant variables that define heterogeneous subgroups and handling data that are missing not at random, a prevalent issue in fields like transcriptomics. While several notable methods have been proposed to address these problems, they typically tackle each issue in isolation, thereby limiting their flexibility and adaptability. This paper introduces a unified framework designed to address these challenges simultaneously. Our approach incorporates a data-driven penalty matrix into penalized clustering to enable more flexible variable selection, along with a mechanism that explicitly models the relationship between missingness and latent class membership. We demonstrate that, under certain regularity conditions, the proposed framework achieves both asymptotic consistency and selection consistency, even in the presence of missing data. This unified strategy significantly enhances the capability and efficiency of model-based clustering, advancing methodologies for identifying informative variables that define homogeneous subgroups in the presence of complex missing data patterns. The performance of the framework, including its computational efficiency, is evaluated through simulations and demonstrated using both synthetic and real-world transcriptomic datasets.
Authors: Binh Duc Vu, Jan Kapar, Marvin Wright, David S. Watson
Abstract: We propose a principled method for autoencoding with random forests. Our strategy builds on foundational results from nonparametric statistics and spectral graph theory to learn a low-dimensional embedding of the model that optimally represents relationships in the data. We provide exact and approximate solutions to the decoding problem via constrained optimization, split relabeling, and nearest neighbors regression. These methods effectively invert the compression pipeline, establishing a map from the embedding space back to the input space using splits learned by the ensemble's constituent trees. The resulting decoders are universally consistent under common regularity assumptions. The procedure works with supervised or unsupervised models, providing a window into conditional or joint distributions. We demonstrate various applications of this autoencoder, including powerful new tools for visualization, compression, clustering, and denoising. Experiments illustrate the ease and utility of our method in a wide range of settings, including tabular, image, and genomic data.
Authors: Liangliang Zhang, Zhuorui Jiang, Hongliang Chi, Haoyang Chen, Mohammed Elkoumy, Fali Wang, Qiong Wu, Zhengyi Zhou, Shirui Pan, Suhang Wang, Yao Ma
Abstract: Knowledge Graph Question Answering (KGQA) systems rely on high-quality benchmarks to evaluate complex multi-hop reasoning. However, despite their widespread use, popular datasets such as WebQSP and CWQ suffer from critical quality issues, including inaccurate or incomplete ground-truth annotations, poorly constructed questions that are ambiguous, trivial, or unanswerable, and outdated or inconsistent knowledge. Through a manual audit of 16 popular KGQA datasets, including WebQSP and CWQ, we find that the average factual correctness rate is only 57 %. To address these issues, we introduce KGQAGen, an LLM-in-the-loop framework that systematically resolves these pitfalls. KGQAGen combines structured knowledge grounding, LLM-guided generation, and symbolic verification to produce challenging and verifiable QA instances. Using KGQAGen, we construct KGQAGen-10k, a ten-thousand scale benchmark grounded in Wikidata, and evaluate a diverse set of KG-RAG models. Experimental results demonstrate that even state-of-the-art systems struggle on this benchmark, highlighting its ability to expose limitations of existing models. Our findings advocate for more rigorous benchmark construction and position KGQAGen as a scalable framework for advancing KGQA evaluation.
Authors: Fengxiang Wang, Mingshuo Chen, Xuming He, Yueying Li, YiFan Zhang, Feng Liu, Zijie Guo, Zhenghao Hu, Jiong Wang, Jingyi Xu, Zhangrui Li, Fenghua Ling, Ben Fei, Weijia Li, Long Lan, Wenjing Yang, Wenlong Zhang, Lei Bai
Abstract: Existing benchmarks for multimodal learning in Earth science offer limited, siloed coverage of Earth's spheres and their cross-sphere interactions, typically restricting evaluation to the human-activity sphere of atmosphere and to at most 16 tasks. These limitations: \textit{narrow-source heterogeneity (single/few data sources), constrained scientific granularity, and limited-sphere extensibility}. Therefore, we introduce \textbf{OmniEarth-Bench}, the first multimodal benchmark that systematically spans all six spheres: atmosphere, lithosphere, oceanosphere, cryosphere, biosphere, and human-activity sphere, and cross-spheres. Built with a scalable, modular-topology data inference framework and native multi-observation sources and expert-in-the-loop curation, OmniEarth-Bench produces 29,855 standardized, expert-curated annotations. All annotations are organized into a four-level hierarchy (Sphere, Scenario, Ability, Task), encompassing 109 expert-curated evaluation tasks. Experiments on 9 state-of-the-art MLLMs reveal that even the most advanced models struggle with our benchmarks, where none of them reach 35\% accuracy, revealing systematic gaps in Earth-system cognitive ability. The dataset and evaluation code were released at OmniEarth-Bench (https://anonymous.4open.science/r/OmniEarth-Bench-B1BD).
URLs: https://anonymous.4open.science/r/OmniEarth-Bench-B1BD).
Authors: Mu Qiao
Abstract: Identifying evolutionary correspondences between cell types across species is a fundamental challenge in comparative genomics and evolutionary biology. Existing approaches often rely on either reference-based matching, which imposes asymmetry by designating one species as the reference, or projection-based matching, which may increase computational complexity and obscure biological interpretability at the cell-type level. Here, we present OT-MESH, an unsupervised computational framework leveraging entropy-regularized optimal transport (OT) to systematically determine cross-species cell type homologies. Our method uniquely integrates the Minimize Entropy of Sinkhorn (MESH) technique to refine the OT plan, transforming diffuse transport matrices into sparse, interpretable correspondences. Through systematic evaluation on synthetic datasets, we demonstrate that OT-MESH achieves near-optimal matching accuracy with computational efficiency, while maintaining remarkable robustness to noise. Compared to other OT-based methods like RefCM, OT-MESH provides speedup while achieving comparable accuracy. Applied to retinal bipolar cells (BCs) and retinal ganglion cells (RGCs) from mouse and macaque, OT-MESH accurately recovers known evolutionary relationships and uncovers novel correspondences, one of which was independently validated experimentally. Thus, our framework offers a principled, scalable, and interpretable solution for evolutionary cell type mapping, facilitating deeper insights into cellular specialization and conservation across species.
Authors: George Orfanides, Tim Hoheisel, Marwa El Halabi
Abstract: Submodular functions, defined on continuous or discrete domains, arise in numerous applications. We study the minimization of the difference of two submodular (DS) functions, over both domains, extending prior work restricted to set functions. We show that all functions on discrete domains and all smooth functions on continuous domains are DS. For discrete domains, we observe that DS minimization is equivalent to minimizing the difference of two convex (DC) functions, as in the set function case. We propose a novel variant of the DC Algorithm (DCA) and apply it to the resulting DC Program, obtaining comparable theoretical guarantees as in the set function case. The algorithm can be applied to continuous domains via discretization. Experiments demonstrate that our method outperforms baselines in integer compressive sensing and integer least squares.
Authors: Felix Wagner, Pramit Saha, Harry Anthony, J. Alison Noble, Konstantinos Kamnitsas
Abstract: Safe deployment of machine learning (ML) models in safety-critical domains such as medical imaging requires detecting inputs with characteristics not seen during training, known as out-of-distribution (OOD) detection, to prevent unreliable predictions. Effective OOD detection after deployment could benefit from access to the training data, enabling direct comparison between test samples and the training data distribution to identify differences. State-of-the-art OOD detection methods, however, either discard the training data after deployment or assume that test samples and training data are centrally stored together, an assumption that rarely holds in real-world settings. This is because shipping the training data with the deployed model is usually impossible due to the size of training databases, as well as proprietary or privacy constraints. We introduce the Isolation Network, an OOD detection framework that quantifies the difficulty of separating a target test sample from the training data by solving a binary classification task. We then propose Decentralized Isolation Networks (DIsoN), which enables the comparison of training and test data when data-sharing is impossible, by exchanging only model parameters between the remote computational nodes of training and deployment. We further extend DIsoN with class-conditioning, comparing a target sample solely with training data of its predicted class. We evaluate DIsoN on four medical imaging datasets (dermatology, chest X-ray, breast ultrasound, histopathology) across 12 OOD detection tasks. DIsoN performs favorably against existing methods while respecting data-privacy. This decentralized OOD detection framework opens the way for a new type of service that ML developers could provide along with their models: providing remote, secure utilization of their training data for OOD detection services. Code: https://github.com/FelixWag/DIsoN
Authors: Michael Amir, Matteo Bettini, Amanda Prorok
Abstract: The success of teams in robotics, nature, and society often depends on the division of labor among diverse specialists; however, a principled explanation for when such diversity surpasses a homogeneous team is still missing. Focusing on multi-agent task allocation problems, we study this question from the perspective of reward design: what kinds of objectives are best suited for heterogeneous teams? We first consider an instantaneous, non-spatial setting where the global reward is built by two generalized aggregation operators: an inner operator that maps the $N$ agents' effort allocations on individual tasks to a task score, and an outer operator that merges the $M$ task scores into the global team reward. We prove that the curvature of these operators determines whether heterogeneity can increase reward, and that for broad reward families this collapses to a simple convexity test. Next, we ask what incentivizes heterogeneity to emerge when embodied, time-extended agents must learn an effort allocation policy. To study heterogeneity in such settings, we use multi-agent reinforcement learning (MARL) as our computational paradigm, and introduce Heterogeneity Gain Parameter Search (HetGPS), a gradient-based algorithm that optimizes the parameter space of underspecified MARL environments to find scenarios where heterogeneity is advantageous. Across different environments, we show that HetGPS rediscovers the reward regimes predicted by our theory to maximize the advantage of heterogeneity, both validating HetGPS and connecting our theoretical insights to reward design in MARL. Together, these results help us understand when behavioral diversity delivers a measurable benefit.
Authors: Piotr Gai\'nski, Oussama Boussif, Andrei Rekesh, Dmytro Shevchuk, Ali Parviz, Mike Tyers, Robert A. Batey, Micha{\l} Koziarski
Abstract: Template-based molecular generation offers a promising avenue for drug design by ensuring generated compounds are synthetically accessible through predefined reaction templates and building blocks. In this work, we tackle three core challenges in template-based GFlowNets: (1) minimizing synthesis cost, (2) scaling to large building block libraries, and (3) effectively utilizing small fragment sets. We propose Recursive Cost Guidance, a backward policy framework that employs auxiliary machine learning models to approximate synthesis cost and viability. This guidance steers generation toward low-cost synthesis pathways, significantly enhancing cost-efficiency, molecular diversity, and quality, especially when paired with an Exploitation Penalty that balances the trade-off between exploration and exploitation. To enhance performance in smaller building block libraries, we develop a Dynamic Library mechanism that reuses intermediate high-reward states to construct full synthesis trees. Our approach establishes state-of-the-art results in template-based molecular generation.
Authors: Qingyue Jiao, Yongcan Tang, Jun Zhuang, Jason Cong, Yiyu Shi
Abstract: Machine learning-assisted diagnosis shows promise, yet medical imaging datasets are often scarce, imbalanced, and constrained by privacy, making data augmentation essential. Classical generative models typically demand extensive computational and sample resources. Quantum computing offers a promising alternative, but existing quantum-based image generation methods remain limited in scale and often face barren plateaus. We present MediQ-GAN, a quantum-inspired GAN with prototype-guided skip connections and a dual-stream generator that fuses classical and quantum-inspired branches. Its variational quantum circuits inherently preserve full-rank mappings, avoid rank collapse, and are theory-guided to balance expressivity with trainability. Beyond generation quality, we provide the first latent-geometry and rank-based analysis of quantum-inspired GANs, offering theoretical insight into their performance. Across three medical imaging datasets, MediQ-GAN outperforms state-of-the-art GANs and diffusion models. While validated on IBM hardware for robustness, our contribution is hardware-agnostic, offering a scalable and data-efficient framework for medical image generation and augmentation.
Authors: Hassan Baker, Matthew S. Emigh, Austin J. Brockmeier
Abstract: As a computer vision task, automatic object segmentation remains challenging in specialized image domains without massive labeled data, such as synthetic aperture sonar images, remote sensing, biomedical imaging, etc. In any domain, obtaining pixel-wise segmentation masks is expensive. In this work, we propose a method for training a masking network to perform binary object segmentation using weak supervision in the form of image-wise presence or absence of an object of interest, which provides less information but may be obtained more quickly from manual or automatic labeling. A key step in our method is that the segmented objects can be placed into background-only images to create realistic images of the objects with counterfactual backgrounds. To create a contrast between the original and counterfactual background images, we propose to first cluster the background-only images and then, during learning, create counterfactual images that blend objects segmented from their original source backgrounds to backgrounds chosen from a targeted cluster. One term in the training loss is the divergence between these counterfactual images and the real object images with backgrounds of the target cluster. The other term is a supervised loss for background-only images. While an adversarial critic could provide the divergence, we use sample-based divergences. We conduct experiments on side-scan and synthetic aperture sonar in which our approach succeeds compared to previous unsupervised segmentation baselines that were only tested on natural images. Furthermore, to show generality we extend our experiments to natural images, obtaining reasonable performance with our method that avoids pretrained networks, generative networks, and adversarial critics. The code for this work can be found at \href{GitHub}{https://github.com/bakerhassan/WSOS}.
Authors: Edan Toledo, Karen Hambardzumyan, Martin Josifoski, Rishi Hazra, Nicolas Baldwin, Alexis Audran-Reiss, Michael Kuchnik, Despoina Magka, Minqi Jiang, Alisia Maria Lupidi, Andrei Lupu, Roberta Raileanu, Kelvin Niu, Tatiana Shavrina, Jean-Christophe Gagnon-Audet, Michael Shvartsman, Shagun Sodhani, Alexander H. Miller, Abhishek Charnalia, Derek Dunfield, Carole-Jean Wu, Pontus Stenetorp, Nicola Cancedda, Jakob Nicolaus Foerster, Yoram Bachrach
Abstract: AI research agents are demonstrating great potential to accelerate scientific progress by automating the design, implementation, and training of machine learning models. We focus on methods for improving agents' performance on MLE-bench, a challenging benchmark where agents compete in Kaggle competitions to solve real-world machine learning problems. We formalize AI research agents as search policies that navigate a space of candidate solutions, iteratively modifying them using operators. By designing and systematically varying different operator sets and search policies (Greedy, MCTS, Evolutionary), we show that their interplay is critical for achieving high performance. Our best pairing of search strategy and operator set achieves a state-of-the-art result on MLE-bench lite, increasing the success rate of achieving a Kaggle medal from 39.6% to 47.7%. Our investigation underscores the importance of jointly considering the search strategy, operator design, and evaluation methodology in advancing automated machine learning.
Authors: Brennen A. Hill, Mant Koh En Wei, Thangavel Jishnuanandh
Abstract: Robust coordination is critical for effective decision-making in multi-agent systems, especially under partial observability. A central question in Multi-Agent Reinforcement Learning (MARL) is whether to engineer communication protocols or learn them end-to-end. We investigate this dichotomy using embodied world models. We propose and compare two communication strategies for a cooperative task-allocation problem. The first, Learned Direct Communication (LDC), learns a protocol end-to-end. The second, Intention Communication, uses an engineered inductive bias: a compact, learned world model, the Imagined Trajectory Generation Module (ITGM), which uses the agent's own policy to simulate future states. A Message Generation Network (MGN) then compresses this plan into a message. We evaluate these approaches on goal-directed interaction in a grid world, a canonical abstraction for embodied AI problems, while scaling environmental complexity. Our experiments reveal that while emergent communication is viable in simple settings, the engineered, world model-based approach shows superior performance, sample efficiency, and scalability as complexity increases. These findings advocate for integrating structured, predictive models into MARL agents to enable active, goal-driven coordination.
Authors: Kang-Hyun Lee, Faez Ahmed
Abstract: A major obstacle to establishing reliable structure-property (SP) linkages in materials engineering is the scarcity of diverse 3D microstructure datasets. Limited dataset availability and insufficient control over the analysis and design space restrict the variety of achievable microstructure morphologies, hindering progress in solving the inverse (property-to-structure) design problem. To address these challenges, we introduce MicroLad, a latent diffusion framework specifically designed for reconstructing 3D microstructures from 2D data. Trained on 2D images and employing multi-plane denoising diffusion sampling in the latent space, the framework reliably generates stable and coherent 3D volumes that remain statistically consistent with the original data. While this reconstruction capability enables dimensionality expansion (2D-to-3D) for generating statistically equivalent 3D samples from 2D data, effective exploration of microstructure design requires methods to guide the generation process toward specific objectives. To achieve this, MicroLad integrates score distillation sampling (SDS), which combines a differentiable score loss with microstructural descriptor-matching and property-alignment terms. This approach updates encoded 2D slices of the 3D volume in the latent space, enabling robust inverse-controlled 2D-to-3D microstructure generation. Consequently, the method facilitates exploration of an expanded 3D microstructure analysis and design space in terms of both microstructural descriptors and material properties.
Authors: Imran S. A. Khan, Emmanuel G. Blanchard, S\'ebastien George
Abstract: This paper introduces the Future Atmospheric Conditions Training System (FACTS), a novel platform that advances climate resilience education through place-based, adaptive learning experiences. FACTS combines real-time atmospheric data collected by IoT sensors with curated resources from a Knowledge Base to dynamically generate localized learning challenges. Learner responses are analyzed by a Generative AI powered server, which delivers personalized feedback and adaptive support. Results from a user evaluation indicate that participants found the system both easy to use and effective for building knowledge related to climate resilience. These findings suggest that integrating IoT and Generative AI into atmospherically adaptive learning technologies holds significant promise for enhancing educational engagement and fostering climate awareness.
Authors: Brennen Hill
Abstract: The capacity of an embodied agent to understand, predict, and interact with its environment is fundamentally contingent on an internal world model. This paper introduces a novel framework for investigating the formation and adaptation of such world models within a biological substrate: human neural organoids. We present a curriculum of three scalable, closed-loop virtual environments designed to train these biological agents and probe the underlying synaptic mechanisms of learning, such as long-term potentiation (LTP) and long-term depression (LTD). We detail the design of three distinct task environments that demand progressively more sophisticated world models for successful decision-making: (1) a conditional avoidance task for learning static state-action contingencies, (2) a one-dimensional predator-prey scenario for goal-directed interaction, and (3) a replication of the classic Pong game for modeling dynamic, continuous-time systems. For each environment, we formalize the state and action spaces, the sensory encoding and motor decoding mechanisms, and the feedback protocols based on predictable (reward) and unpredictable (punishment) stimulation, which serve to drive model refinement. In a significant methodological advance, we propose a meta-learning approach where a Large Language Model automates the generative design and optimization of experimental protocols, thereby scaling the process of environment and curriculum design. Finally, we outline a multi-modal evaluation strategy that moves beyond task performance to directly measure the physical correlates of the learned world model by quantifying synaptic plasticity at electrophysiological, cellular, and molecular levels. This work bridges the gap between model-based reinforcement learning and computational neuroscience, offering a unique platform for studying embodiment, decision-making, and the physical basis of intelligence.
Authors: Brennen Hill
Abstract: Recent advances in agent development have focused on scaling model size and raw interaction data, mirroring successes in large language models. However, for complex, long-horizon multi-agent tasks such as robotic soccer, this end-to-end approach often fails due to intractable exploration spaces and sparse rewards. We propose that an effective world model for decision-making must model the world's physics and also its task semantics. A systematic review of 2024 research in low-resource multi-agent soccer reveals a clear trend towards integrating symbolic and hierarchical methods, such as Hierarchical Task Networks (HTNs) and Bayesian Strategy Networks (BSNs), with multi-agent reinforcement learning (MARL). These methods decompose complex goals into manageable subgoals, creating an intrinsic curriculum that shapes agent learning. We formalize this trend into a framework for Hierarchical Task Environments (HTEs), which are essential for bridging the gap between simple, reactive behaviors and sophisticated, strategic team play. Our framework incorporates the use of Large Language Models (LLMs) as generative world models of tasks, capable of dynamically generating this scaffolding. We argue that HTEs provide a mechanism to guide exploration, generate meaningful learning signals, and train agents to internalize hierarchical structure, enabling the development of more capable and general-purpose agents with greater sample efficiency than purely end-to-end approaches.
Authors: Benjamin Prada, Ankur Mali
Abstract: Classical circuit complexity characterizes parallel computation in purely combinatorial terms, ignoring the physical constraints that govern real hardware. The standard classes $\mathbf{NC}$, $\mathbf{AC}$, and $\mathbf{TC}$ treat unlimited fan-in, free interconnection, and polynomial gate counts as feasible -- assumptions that conflict with geometric, energetic, and thermodynamic realities. We introduce the family of \textit{realizable circuit classes} $\mathbf{RC}_d$, which model computation embedded in physical $d$-dimensional space. Each circuit in $\mathbf{RC}_d$ obeys conservative realizability laws: volume scales as $\mathcal{O}(t^d)$, cross-boundary information flux is bounded by $\mathcal{O}(t^{d-1})$ per unit time, and growth occurs through local, physically constructible edits. These bounds apply to all causal systems, classical or quantum. Within this framework, we show that algorithms with runtime $\omega(n^{d/(d-1)})$ cannot scale to inputs of maximal entropy, and that any $d$-dimensional parallel implementation offers at most a polynomial speed-up of degree $(d-1)$ over its optimal sequential counterpart. In the limit $d\to\infty$, $\mathbf{RC}_\infty(\mathrm{polylog})=\mathbf{NC}$, recovering classical parallelism as a non-physical idealization. By unifying geometry, causality, and information flow, $\mathbf{RC}_d$ extends circuit complexity into the physical domain, revealing universal scaling laws for computation.
Authors: Kaihua Ding
Abstract: Reliable evaluation of AI systems remains a fundamental challenge when ground truth labels are unavailable, particularly for systems generating natural language outputs like AI chat and agent systems. Many of these AI agents and systems focus on entity-centric tasks. In enterprise contexts, organizations deploy AI systems for entity linking, data integration, and information retrieval where verification against gold standards is often infeasible due to proprietary data constraints. Academic deployments face similar challenges when evaluating AI systems on specialized datasets with ambiguous criteria. Conventional evaluation frameworks, rooted in supervised learning paradigms, fail in such scenarios where single correct answers cannot be defined. We introduce VB-Score, a variance-bounded evaluation framework for entity-centric AI systems that operates without ground truth by jointly measuring effectiveness and robustness. Given system inputs, VB-Score enumerates plausible interpretations through constraint relaxation and Monte Carlo sampling, assigning probabilities that reflect their likelihood. It then evaluates system outputs by their expected success across interpretations, penalized by variance to assess robustness of the system. We provide formal theoretical analysis establishing key properties including range, monotonicity, and stability along with concentration bounds for Monte Carlo estimation. Through case studies on AI systems with ambiguous inputs, we demonstrate that VB-Score reveals robustness differences hidden by conventional evaluation frameworks, offering a principled measurement framework for assessing AI system reliability in label-scarce domains.
Authors: Nicol\`o Dal Fabbro, Milad Mesbahi, Renato Mendes, Jo\~ao Borges de Sousa, George J. Pappas
Abstract: We study the problem of long-term (multiple days) mapping of a river plume using multiple autonomous underwater vehicles (AUVs), focusing on the Douro river representative use-case. We propose an energy - and communication - efficient multi-agent reinforcement learning approach in which a central coordinator intermittently communicates with the AUVs, collecting measurements and issuing commands. Our approach integrates spatiotemporal Gaussian process regression (GPR) with a multi-head Q-network controller that regulates direction and speed for each AUV. Simulations using the Delft3D ocean model demonstrate that our method consistently outperforms both single- and multi-agent benchmarks, with scaling the number of agents both improving mean squared error (MSE) and operational endurance. In some instances, our algorithm demonstrates that doubling the number of AUVs can more than double endurance while maintaining or improving accuracy, underscoring the benefits of multi-agent coordination. Our learned policies generalize across unseen seasonal regimes over different months and years, demonstrating promise for future developments of data-driven long-term monitoring of dynamic plume environments.
Authors: Khalid Mehtab Khan, Anagha Kulkarni
Abstract: Identifying cultural capital (CC) themes in student reflections can offer valuable insights that help foster equitable learning environments in classrooms. However, themes such as aspirational goals or family support are often woven into narratives, rather than appearing as direct keywords. This makes them difficult to detect for standard NLP models that process sentences in isolation. The core challenge stems from a lack of awareness, as standard models are pre-trained on general corpora, leaving them blind to the domain-specific language and narrative context inherent to the data. To address this, we introduce AWARE, a framework that systematically attempts to improve a transformer model's awareness for this nuanced task. AWARE has three core components: 1) Domain Awareness, adapting the model's vocabulary to the linguistic style of student reflections; 2) Context Awareness, generating sentence embeddings that are aware of the full essay context; and 3) Class Overlap Awareness, employing a multi-label strategy to recognize the coexistence of themes in a single sentence. Our results show that by making the model explicitly aware of the properties of the input, AWARE outperforms a strong baseline by 2.1 percentage points in Macro-F1 and shows considerable improvements across all themes. This work provides a robust and generalizable methodology for any text classification task in which meaning depends on the context of the narrative.
Authors: Jahidul Arafat, Sanjaya Poudel
Abstract: Nanopore sequencing enables real-time long-read DNA sequencing with reads exceeding 10 kilobases, but inherent error rates of 12-15 percent present significant computational challenges for read alignment. The critical seed chaining step must connect exact k-mer matches between reads and reference genomes while filtering spurious matches, yet state-of-the-art methods rely on fixed gap penalty functions unable to adapt to varying genomic contexts including tandem repeats and structural variants. This paper presents RawHash3, a hybrid framework combining graph neural networks with classical dynamic programming for adaptive seed chaining that maintains real-time performance while providing statistical guarantees. We formalize seed chaining as graph learning where seeds constitute nodes with 12-dimensional feature vectors and edges encode 8-dimensional spatial relationships including gap consistency. Our architecture employs three-layer EdgeConv GNN with confidence-based method selection that dynamically switches between learned guidance and algorithmic fallback. Comprehensive evaluation on 1,000 synthetic nanopore reads with 5,200 test seeds demonstrates RawHash3 achieves 99.94 percent precision and 40.07 percent recall, representing statistically significant 25.0 percent relative improvement over baseline with p less than 0.001. The system maintains median inference latency of 1.59ms meeting real-time constraints, while demonstrating superior robustness with 100 percent success rate under 20 percent label corruption versus baseline degradation to 30.3 percent. Cross-validation confirms stability establishing graph neural networks as viable approach for production genomics pipelines.
Authors: Chun Chet Ng, Wei Zeng Low, Yin Yin Boon
Abstract: Despite accounting for 96.1% of all businesses in Malaysia, access to financing remains one of the most persistent challenges faced by Micro, Small, and Medium Enterprises (MSMEs). Newly established or young businesses are often excluded from formal credit markets as traditional underwriting approaches rely heavily on credit bureau data. This study investigates the potential of bank statement data as an alternative data source for credit assessment to promote financial inclusion in emerging markets. Firstly, we propose a cash flow-based underwriting pipeline where we utilise bank statement data for end-to-end data extraction and machine learning credit scoring. Secondly, we introduce a novel dataset of 611 loan applicants from a Malaysian lending institution. Thirdly, we develop and evaluate credit scoring models based on application information and bank transaction-derived features. Empirical results show that the use of such data boosts the performance of all models on our dataset, which can improve credit scoring for new-to-lending MSMEs. Lastly, we intend to release the anonymised bank transaction dataset to facilitate further research on MSMEs financial inclusion within Malaysia's emerging economy.
Authors: Haowei Hua (Princeton University), Hong Jiao (University of Maryland, College Park), Xinyi Wang (University of Maryland, College Park & Beijing Normal University)
Abstract: BERT and its variants are extensively explored for automated scoring. However, a limit of 512 tokens for these encoder-based models showed the deficiency in automated scoring of long essays. Thus, this research explores generative language models for automated scoring of long essays via summarization and prompting. The results revealed great improvement of scoring accuracy with QWK increased from 0.822 to 0.8878 for the Learning Agency Lab Automated Essay Scoring 2.0 dataset.
Authors: Stefan M. Fischer, Johannes Kiechle, Laura Daza, Lina Felsner, Richard Osuala, Daniel M. Lang, Karim Lekadir, Jan C. Peeken, Julia A. Schnabel
Abstract: In this work, we introduce Progressive Growing of Patch Size, an automatic curriculum learning approach for 3D medical image segmentation. Our approach progressively increases the patch size during model training, resulting in an improved class balance for smaller patch sizes and accelerated convergence of the training process. We evaluate our curriculum approach in two settings: a resource-efficient mode and a performance mode, both regarding Dice score performance and computational costs across 15 diverse and popular 3D medical image segmentation tasks. The resource-efficient mode matches the Dice score performance of the conventional constant patch size sampling baseline with a notable reduction in training time to only 44%. The performance mode improves upon constant patch size segmentation results, achieving a statistically significant relative mean performance gain of 1.28% in Dice Score. Remarkably, across all 15 tasks, our proposed performance mode manages to surpass the constant patch size baseline in Dice Score performance, while simultaneously reducing training time to only 89%. The benefits are particularly pronounced for highly imbalanced tasks such as lesion segmentation tasks. Rigorous experiments demonstrate that our performance mode not only improves mean segmentation performance but also reduces performance variance, yielding more trustworthy model comparison. Furthermore, our findings reveal that the proposed curriculum sampling is not tied to a specific architecture but represents a broadly applicable strategy that consistently boosts performance across diverse segmentation models, including UNet, UNETR, and SwinUNETR. In summary, we show that this simple yet elegant transformation on input data substantially improves both Dice Score performance and training runtime, while being compatible across diverse segmentation backbones.
Authors: Tongyi DeepResearch Team, Baixuan Li, Bo Zhang, Dingchu Zhang, Fei Huang, Guangyu Li, Guoxin Chen, Huifeng Yin, Jialong Wu, Jingren Zhou, Kuan Li, Liangcai Su, Litu Ou, Liwen Zhang, Pengjun Xie, Rui Ye, Wenbiao Yin, Xinmiao Yu, Xinyu Wang, Xixi Wu, Xuanzhong Chen, Yida Zhao, Zhen Zhang, Zhengwei Tao, Zhongwang Zhang, Zile Qiao, Chenxi Wang, Donglei Yu, Gang Fu, Haiyang Shen, Jiayin Yang, Jun Lin, Junkai Zhang, Kui Zeng, Li Yang, Hailong Yin, Maojia Song, Ming Yan, Minpeng Liao, Peng Xia, Qian Xiao, Rui Min, Ruixue Ding, Runnan Fang, Shaowei Chen, Shen Huang, Shihang Wang, Shihao Cai, Weizhou Shen, Xiaobin Wang, Xin Guan, Xinyu Geng, Yingcheng Shi, Yuning Wu, Zhuo Chen, Zijian Li, Yong Jiang
Abstract: We present Tongyi DeepResearch, an agentic large language model, which is specifically designed for long-horizon, deep information-seeking research tasks. To incentivize autonomous deep research agency, Tongyi DeepResearch is developed through an end-to-end training framework that combines agentic mid-training and agentic post-training, enabling scalable reasoning and information seeking across complex tasks. We design a highly scalable data synthesis pipeline that is fully automatic, without relying on costly human annotation, and empowers all training stages. By constructing customized environments for each stage, our system enables stable and consistent interactions throughout. Tongyi DeepResearch, featuring 30.5 billion total parameters, with only 3.3 billion activated per token, achieves state-of-the-art performance across a range of agentic deep research benchmarks, including Humanity's Last Exam, BrowseComp, BrowseComp-ZH, WebWalkerQA, xbench-DeepSearch, FRAMES and xbench-DeepSearch-2510. We open-source the model, framework, and complete solutions to empower the community.
Authors: John-Joseph Brady, Benjamin Cox, Yunpeng Li, V\'ictor Elvira
Abstract: State-space models (SSMs) are a widely used tool in time series analysis. In the complex systems that arise from real-world data, it is common to employ particle filtering (PF), an efficient Monte Carlo method for estimating the hidden state corresponding to a sequence of observations. Applying particle filtering requires specifying both the parametric form and the parameters of the system, which are often unknown and must be estimated. Gradient-based optimisation techniques cannot be applied directly to standard particle filters, as the filters themselves are not differentiable. However, several recently proposed methods modify the resampling step to make particle filtering differentiable. In this paper, we present an implementation of several such differentiable particle filters (DPFs) with a unified API built on the popular PyTorch framework. Our implementation makes these algorithms easily accessible to a broader research community and facilitates straightforward comparison between them. We validate our framework by reproducing experiments from several existing studies and demonstrate how DPFs can be applied to address several common challenges with state space modelling.
Authors: Nelson Matti\'e, Arturo Sanchez-Azofeifa, Pablo Crespo-Peremarch, Juan-Ygnacio L\'opez-Hern\'andez
Abstract: According to the Paris Climate Change Agreement, all nations are required to submit reports on their greenhouse gas emissions and absorption every two years by 2024. Consequently, forests play a crucial role in reducing carbon emissions, which is essential for meeting these obligations. Recognizing the significance of forest conservation in the global battle against climate change, Article 5 of the Paris Agreement emphasizes the need for high-quality forest data. This study focuses on enhancing methods for mapping aboveground biomass in tropical dry forests. Tropical dry forests are considered one of the least understood tropical forest environments; therefore, there is a need for accurate approaches to estimate carbon pools. We employ a comparative analysis of AGB estimates, utilizing different discrete and full-waveform laser scanning datasets in conjunction with Ordinary Least Squares and Bayesian approaches SVM. Airborne Laser Scanning, Unmanned Laser Scanning, and Space Laser Scanning were used as independent variables for extracting forest metrics. Variable selection, SVM regression tuning, and cross-validation via a machine-learning approach were applied to account for overfitting and underfitting. The results indicate that six key variables primarily related to tree height: Elev\.minimum, Elev\.L3, lev\.MAD\.mode, Elev\.mode, Elev\.MAD\.median, and Elev\.skewness, are important for AGB estimation using ALSD and ULSD, while Leaf Area Index, canopy coverage and height, terrain elevation, and full-waveform signal energy emerged as the most vital variables. AGB values estimated from ten permanent tropical dry forest plots in Costa Rica Guanacaste province ranged from 26.02 Mg/ha to 175.43 Mg/ha. The SVM regressions demonstrated a 17.89 error across all laser scanning systems, with SLSF W exhibiting the lowest error 17.07 in estimating total biomass per plot.
Authors: Michael D. Moffitt
Abstract: The Abstraction and Reasoning Corpus remains one of the most compelling and challenging benchmarks for tracking progress toward achieving Artificial General Intelligence. In contrast to other evaluation datasets designed to assess an agent's task-specific skills or accumulated knowledge, the ARC-AGI suite is specifically targeted at measuring skill acquisition efficiency, a trait that has (so far) been lacking in even the most sophisticated machine learning systems. For algorithms that require extensive intra-task exemplars, a significant constraint imposed by ARC-AGI is the modest cardinality of its demonstration set, comprising a small number of $\langle$ input, output $\rangle$ grids per task specifying the corresponding transformation. To embellish the space of viable sample pairs, this paper introduces ARC-GEN, an open-source procedural generator aimed at extending the original ARC-AGI training dataset as faithfully as possible. Unlike prior efforts, our generator is both exhaustive (covering all four-hundred tasks) and mimetic (more closely honoring the distributional properties and characteristics embodied in the initial ARC-AGI-1 release). We also discuss the use of this generator in establishing a static benchmark suite to verify the correctness of programs submitted to the 2025 Google Code Golf Championship.
Authors: Peter Atandoh, Jie Zou, Weikang Guo, Jiwei Wei, Zheng Wang
Abstract: Sentiment analysis using deep learning and pre-trained language models (PLMs) has gained significant traction due to their ability to capture rich contextual representations. However, existing approaches often underperform in scenarios involving nuanced emotional cues, domain shifts, and imbalanced sentiment distributions. We argue that these limitations stem from inadequate semantic grounding, poor generalization to diverse linguistic patterns, and biases toward dominant sentiment classes. To overcome these challenges, we propose CISEA-MRFE, a novel PLM-based framework integrating Contextual Instruction (CI), Semantic Enhancement Augmentation (SEA), and Multi-Refined Feature Extraction (MRFE). CI injects domain-aware directives to guide sentiment disambiguation; SEA improves robustness through sentiment-consistent paraphrastic augmentation; and MRFE combines a Scale-Adaptive Depthwise Encoder (SADE) for multi-scale feature specialization with an Emotion Evaluator Context Encoder (EECE) for affect-aware sequence modeling. Experimental results on four benchmark datasets demonstrate that CISEA-MRFE consistently outperforms strong baselines, achieving relative improvements in accuracy of up to 4.6% on IMDb, 6.5% on Yelp, 30.3% on Twitter, and 4.1% on Amazon. These results validate the effectiveness and generalization ability of our approach for sentiment classification across varied domains.
Authors: Simone Sarrocco, Philippe C. Cattin, Peter M. Maloca, Paul Friedrich, Philippe Valmaggia
Abstract: Purpose: To evaluate deep learning (DL) models for enhancing vitreous optical coherence tomography (OCT) image quality and reducing acquisition time. Methods: Conditional Denoising Diffusion Probabilistic Models (cDDPMs), Brownian Bridge Diffusion Models (BBDMs), U-Net, Pix2Pix, and Vector-Quantised Generative Adversarial Network (VQ-GAN) were used to generate high-quality spectral-domain (SD) vitreous OCT images. Inputs were SD ART10 images, and outputs were compared to pseudoART100 images obtained by averaging ten ART10 images per eye location. Model performance was assessed using image quality metrics and Visual Turing Tests, where ophthalmologists ranked generated images and evaluated anatomical fidelity. The best model's performance was further tested within the manually segmented vitreous on newly acquired data. Results: U-Net achieved the highest Peak Signal-to-Noise Ratio (PSNR: 30.230) and Structural Similarity Index Measure (SSIM: 0.820), followed by cDDPM. For Learned Perceptual Image Patch Similarity (LPIPS), Pix2Pix (0.697) and cDDPM (0.753) performed best. In the first Visual Turing Test, cDDPM ranked highest (3.07); in the second (best model only), cDDPM achieved a 32.9% fool rate and 85.7% anatomical preservation. On newly acquired data, cDDPM generated vitreous regions more similar in PSNR to the ART100 reference than true ART1 or ART10 B-scans and achieved higher PSNR on whole images when conditioned on ART1 than ART10. Conclusions: Results reveal discrepancies between quantitative metrics and clinical evaluation, highlighting the need for combined assessment. cDDPM showed strong potential for generating clinically meaningful vitreous OCT images while reducing acquisition time fourfold. Translational Relevance: cDDPMs show promise for clinical integration, supporting faster, higher-quality vitreous imaging. Dataset and code will be made publicly available.
Authors: Yoshihiro Maruyama
Abstract: We propose CatEquiv, a category-equivariant neural network for Human Activity Recognition (HAR) from inertial sensors that systematically encodes temporal, amplitude, and structural symmetries. We introduce a symmetry category that jointly represents cyclic time shifts, positive gain scalings, and the sensor-hierarchy poset, capturing the categorical symmetry structure of the data. CatEquiv achieves equivariance with respect to the categorical symmetry product. On UCI-HAR under out-of-distribution perturbations, CatEquiv attains markedly higher robustness compared with circularly padded CNNs and plain CNNs. These results demonstrate that enforcing categorical symmetries yields strong invariance and generalization without additional model capacity.