Ax Alaa Zniber, Ouassim Karrakchou, Mounir Ghogho 5/8/2026

Balancing Stability and Plasticity in Sequentially Trained Early-Exiting Neural Networks

Method addressing interference between sequentially trained early-exit classifiers in neural networks via stability-plasticity tradeoffs.

Ax Rahul Nandakumar, Deepayan Chakrabarti 5/8/2026

COPYCOP: Ownership Verification for Graph Neural Networks

Technique for verifying GNN model ownership and detecting unauthorized mimicry across different architectures and embeddings.

Ax Rahul Nandakumar, Ben Fauber, Deepayan Chakrabarti 5/8/2026

SPADE: Faster Drug Discovery by Learning from Sparse Data

ML method for drug discovery that learns from sparse ligand-protein binding data to accelerate candidate selection.

Ax David Leeftink, Max Hinne, Marcel van Gerven 5/8/2026

Neural Co-state Policies: Structuring Hidden States in Recurrent Reinforcement Learning

Research on interpretable hidden state structures in recurrent RL policies for partial observability using Pontryagin methods.

Ax Ahmad Aghapour, Erhan Bayraktar, Asaf Cohen 5/8/2026

Conditional Diffusion Under Linear Constraints: Langevin Mixing and Information-Theoretic Guarantees

Zero-shot conditional sampling with pretrained diffusion models for linear inverse problems, providing Langevin mixing guarantees with measurement consistency.

Ax Filip Rydin, Morteza Haghir Chehreghani, Bal\'azs Kulcs\'ar 5/8/2026

Two-Stage Learned Decomposition for Scalable Routing on Multigraphs

Neural routing method for vehicle routing problems on multigraphs using Node-Edge Policy Factorization to handle parallel edges with varying trade-offs at scale.

Ax Ion Matei, Maksym Zhenirovskyy, Anthony Wong 5/8/2026

Differentiable Parameter Optimization for DAEs with State-Dependent Events

Gradient-based parameter learning method for differential-algebraic equations with state-dependent events, handling implicit algebraic variables and discontinuous resets.

Ax Yiwei Zhang, Jeremiah Birrell, Reza Ebrahimi, Rouzbeh Behnia, Jason Pacheco, Elisa Bertino 5/8/2026

Information Theoretic Adversarial Training of Large Language Models

Information-theoretic adversarial training methods for improving LLM robustness against adversarial prompting with improved computational efficiency over existing approaches.

Ax Alexander DeLise, Nick Dexter 5/8/2026

Active Learning for Conditional Generative Compressed Sensing

Active learning framework for conditional generative compressed sensing combining prompt-conditioned generative models with adaptive sampling strategy selection.

Ax Pratik Deshmukh, Atirek Gupta 5/8/2026

On Semantic Loss Fine-Tuning Approach for Preventing Model Collapse in Causal Reasoning

Semantic loss fine-tuning approach preventing catastrophic model collapse when training Transformers on causal reasoning tasks like transitivity and d-separation.

Ax Othmane Kabal, Mounira Harzallah, Fabrice Guillet, Hideaki Takeda, Ryutaro Ichise 5/8/2026

Robustness of Graph Self-Supervised Learning to Real-World Noise: A Case Study on Text-Driven Biomedical Graphs

Study of Graph Self-Supervised Learning robustness to real-world noise from automatic text-extracted biomedical knowledge graphs, identifying failure modes.

Ax Othmane Kabal, Mounira Harzallah, Fabrice Guillet, Hideaki Takeda, Ryutaro Ichise 5/8/2026

A Unified Benchmark for Evaluating Knowledge Graph Construction Methods and Graph Neural Networks

Benchmark framework evaluating knowledge graph construction methods and GNN robustness to noise/fragmentation/semantic inconsistencies in automatically extracted graphs.

Ax Raimondo Fanale 5/8/2026

GRALIS: A Unified Canonical Framework for Linear Attribution Methods via Riesz Representation

GRALIS unified mathematical framework via Riesz representation theory establishing formal connections between major XAI attribution methods (GradCAM, SHAP, LIME, Integrated Gradients).

Ax Dillon Sandhu, Ronald Parr 5/8/2026

Approximate Next Policy Sampling: Replacing Conservative Target Policy Updates in Deep RL

Approximate Next Policy Sampling method solving conservative policy update problem in deep RL by estimating state-visitation distribution of improved policy.

Ax Taeyoung Kim, Joon-Hyuk Ko 5/8/2026

A Robust Foundation Model for Conservation Laws: Injecting Context into Flux Neural Operators via Recurrent Vision Transformers

Flux Neural Operator augmented with ViT-based context injection via hypernetworks for robust inference of conservation law dynamics across initial conditions.

Ax Xuwei Yang, David B. Emerson, Fatemeh Tavakoli, Anastasis Kratsios 5/8/2026

MEMOA: Massive Mixtures of Online Agents via Mean-Field Decentralized Nash Equilibria

MEMOA framework for scaling federated learning of massive agent populations using mean-field decentralized Nash equilibria with minimal communication overhead.

Ax William T. Redman, Erik C. Johnson, Brian Robinson 5/8/2026

Shortcut Solutions Learned by Transformers Impair Continual Compositional Reasoning

Study of how Transformers learn shortcut solutions that impair compositional reasoning in continual learning, examining failure modes in systematic generalization.

Ax Yuheng Lai, Garvesh Raskutti 5/8/2026

Online Localized Conformal Prediction

Online Localized Conformal Prediction framework providing valid uncertainty quantification for time-series via localized calibration under covariate heterogeneity.

Ax Linus Aronsson, Morteza Haghir Chehreghani 5/8/2026

Non-Myopic Active Feature Acquisition via Pathwise Policy Gradients

Non-Myopic Pathwise Policy Gradients (NM-PPG) method for active feature acquisition formulated as POMDP with sequential decision-making perspective.

Ax Shicheng Fan, Nour Elhendawy, Jianle Sun, Ke Fang, Kun Zhang, Yihang Wang, Lu Cheng 5/8/2026

MOSAIC: Module Discovery via Sparse Additive Identifiable Causal Learning for Scientific Time Series

MOSAIC method for discovering interpretable causal modules in scientific time series via sparse additive identifiable causal learning with post-hoc semantics.

Ax Tran Gia Bao Ngo, Zulfikar Alom, Federico Errica, Murat Kantarcioglu, Cuneyt Gurcan Akcora 5/8/2026

Adversarial Graph Neural Network Benchmarks: Towards Practical and Fair Evaluation

Benchmark framework for evaluating adversarial attacks and defenses on Graph Neural Networks, addressing inconsistent experimental settings in GNN robustness research.

Ax Tianyue Yang, Xiao Xue 5/8/2026

Towards Scalable One-Step Generative Modeling for Autoregressive Dynamical System Forecasting

One-step generative modeling for efficient autoregressive dynamical system forecasting with long-trajectory structure preservation.

Ax Nandiraju Gireesh, Yuanliang Ju, He Wang 5/8/2026

Adaptive Q-Chunking for Offline-to-Online Reinforcement Learning

Adaptive Q-chunking approach for offline-to-online reinforcement learning with state-dependent action chunk sizes.

Ax Quang-Huy Nguyen, Jiaqi Wang, Wei-shinn Ku 5/8/2026

FedeKD: Energy-Based Gating for Robust Federated Knowledge Distillation under Heterogeneous Settings

Federated knowledge distillation with energy-based gating for robustness in heterogeneous environments.

Ax Won-Jun Jang, Si-Hyeon Lee 5/8/2026

Accelerating LMO-Based Optimization via Implicit Gradient Transport

Optimization method accelerating linear minimization oracle-based optimizers via implicit gradient transport.

Ax Francisco Giral, Abhijeet Vishwasrao, Andrea Arroyo Ramo, Mahmoud Golestanian, Federica Tonti, Adrian Lozano-Duran, Steven L. Brunton, Sergio Hoyas, Hector Gomez, Soledad Le Clainche, Ricardo Vinuesa 5/8/2026

AeroJEPA: Learning Semantic Latent Representations for Scalable 3D Aerodynamic Field Modeling

Joint-Embedding Predictive Architecture for scalable 3D aerodynamic surrogate modeling with semantic latent representations.

Ax Lei Jiang, Adrian Ildefonso, Daniel Loveless, Fan Chen 5/8/2026

LLMSpace: Carbon Footprint Modeling for Large Language Model Inference on LEO Satellites

Carbon footprint modeling framework for LLM inference on solar-powered LEO satellites.

Ax Brett Barkley, Preston Culbertson, David Fridovich-Keil 5/8/2026

Scaling Pretrained Representations Enables Label-Free Out-of-Distribution Detection Without Fine-Tuning

Out-of-distribution detection using frozen pretrained model representations without fine-tuning.

Ax Ying Chen, Aoxi Li, Jihun Kim, Javad Lavaei 5/8/2026

Structural Correspondence and Universal Approximation in Diagonal plus Low-Rank Neural Networks

Theoretical analysis of expressive power in diagonal plus low-rank neural networks including LoRA structures.

Ax Yufeng Yang, Fangning Zhuo, Ziyi Chen, Heng Huang, Yi Zhou 5/8/2026

Distributionally Robust Multi-Objective Optimization

Distributionally robust multi-objective optimization formulation accounting for distributional shifts in data.

Ax Naveen Mysore 5/8/2026

Temporal Functional Circuits: From Spline Plots to Faithful Explanations in KAN Forecasting

Temporal Functional Circuits framework for interpreting Kolmogorov-Arnold Networks in time-series forecasting with mechanistic explanations.

Ax Amrit Nidhi 5/8/2026

Budgeted Attention Allocation: Cost-Conditioned Compute Control for Efficient Transformers

Proposes Budgeted Attention Allocation with head-gating mechanism to create multiple cost-quality operating points in transformer inference.

Ax Wei Yao, Wang Zhaoyang, Gengze Xu, Chen Qian, Dongrui Liu, Ziqiao Wang, Yong Liu, Yunbei Xu 5/8/2026

On the Blessing of Pre-training in Weak-to-Strong Generalization

Theoretical analysis proving pre-training is essential for weak-to-strong generalization in high-dimensional models using spiked Gaussian framework.

Ax Yui Hashimoto, Takayuki Nishio, Yuichi Kitagawa, Takahito Tanimura 5/8/2026

Enabling Federated Inference via Unsupervised Consensus Embedding

Framework for federated cooperative inference across models using unsupervised consensus embeddings while preserving data and parameter privacy.

Ax Md Anwar Hossen, Fatema Siddika, Juan Pablo Munoz, Tanya Roosta, Ali Jannesari 5/8/2026

CRAFT: Forgetting-Aware Intervention-Based Adaptation for Continual Learning

Introduces CRAFT framework for continual LLM adaptation using low-rank interventions on hidden representations to avoid catastrophic forgetting.

Ax Mei Wu, Wenchao Weng, Wenxin Su, Wenjie Tang, Wei Zhou 5/8/2026

CoMemNet: Contrastive Sampling with Memory Replay Network for Continual Traffic Prediction

Proposes CoMemNet combining non-topological space modeling with temporal learning for continual traffic prediction on dynamic streaming networks.

Ax Mohammad Al Ridhawi, Mahtab Haj Ali, Hussein Al Osman 5/8/2026

Multi-Dimensional Behavioral Evaluation of Agentic Stock Prediction Systems Using LLM Judges with Closed-Loop Reinforcement Learning Feedback

Behavioral evaluation framework for agentic LLM stock prediction systems using LLM judges and RL feedback to assess interdependent decision quality.

Ax Ivan Montero, Tomasz Jurczyk, Bhuwan Dhingra 5/8/2026

RVPO: Risk-Sensitive Alignment via Variance Regularization

Proposes RVPO for risk-sensitive multi-objective alignment in LLMs by penalizing reward variance to prevent constraint neglect in critic-less RLHF.

Ax Myoungjun Kim, Sangwoo Park, Yoseob Han, Jin-Hyun Ahn 5/8/2026

Adaptive Selection of LoRA Components in Privacy-Preserving Federated Learning

Proposes adaptive LoRA component selection in federated learning under differential privacy to reduce aggregation error and improve fine-tuning stability.

Ax Ziqing Wen, Zhouyang Liu, Jiahuan Wang, Ping Luo, Li Shen, Dongsheng Li, Tao Sun 5/8/2026

Revealing Modular Gradient Noise Imbalance in LLMs: Calibrating Adam via Signal-to-Noise Ratio

Analyzes gradient noise imbalance across LLM modules and proposes SNR-based Adam calibration to improve training convergence and stability.

Ax Nicholas Potteiger, Ankita Samaddar, Taylor T. Johnson, Xenofon Koutsoukos 5/8/2026

Reward Shaping and Action Masking for Compositional Tasks using Behavior Trees and LLMs

Combines behavior trees with LLMs for automated reward shaping and action masking in compositional RL tasks, improving subtask learning efficiency.

Ax Zhiyuan Zhai, Xin Wang 5/8/2026

Selective Rollout: Mid-Trajectory Termination for Multi-Sample Agent RL

Introduces Selective Rollout to reduce training costs in GRPO by early-terminating rollouts with zero reward variance in multi-turn dialogue agent training.

Ax Elad Hoffer, Yochai Blau, Ron Banner, Daniel Soudry, Boris Ginsburg 5/8/2026

Retrieval from Within: An Intrinsic Capability of Attention-Based Models

Proposes INTRA framework enabling attention-based models to perform retrieval-augmented generation using only internal representations without external systems.

Ax Zegu Zhang, Jianhua Peng, Jian Zhang 5/8/2026

A Testable Certificate for Constant Collapse in Teacher-Guided VAEs

Theoretical analysis of constant collapse in variational autoencoders using teacher-guided training, establishing exact thresholds for this failure mode.

Ax Shen Xu, Xiangwen Zhuge, Zhe Xu, Yingkun Hu, Zheng Yang, Yunhao Liu 5/8/2026

HCInfer: An Efficient Inference System via Error Compensation for Resource-Constrained Devices

Introduces HCInfer, an inference system using error compensation to enable efficient LLM deployment on memory-constrained devices without significant accuracy loss.

Ax Yulong Huang, Xiang Liu, Hongxiang Huang, Xiaopeng Lin, Zunchang Liu, Xiaowen Chu, Zeke Xie, Bojun Cheng 5/8/2026

MDN: Parallelizing Stepwise Momentum for Delta Linear Attention

Proposes MDN for parallelizing momentum-based optimization in linear attention models to improve convergence and information retention in long-sequence LLM scaling.

Ax Hua-Dong Xiong 5/8/2026

Hypothesis generation and updating in large language models

Studies how LLMs generate and update hypotheses from partial information using the number game task, analyzing inference optimality and mechanisms.

Ax Samuel Blad, Martin L\"angkvist, Amy Loutfi 5/8/2026

Measuring Learning Progress via Gradient-Momentum Coupling

Proposes Gradient-Momentum Coupling metric to measure learning progress in curiosity-driven RL exploration.

Ax Carlo Romeo, Girolamo Macaluso, Alessandro Sestini, Andrew D. Bagdanov 5/8/2026

SOPE: Stabilizing Off-Policy Evaluation for Online RL with Prior Data

Proposes SOPE to stabilize off-policy evaluation in online reinforcement learning with prior data without manual tuning.

Ax Junhao Cai, Dohun Kim, Dowon Kim, Sung Il Choi, Chengjun Jin, Juhyun Park, Changhee Joo 5/8/2026

Retain-Neutral Surrogates for Min-Max Unlearning

Addresses machine unlearning via retain-neutral surrogates in min-max formulation to remove training data influence.