Ax Jack Young 5/14/2026

WriteSAE: Sparse Autoencoders for Recurrent State

WriteSAE: First sparse autoencoder decomposing and editing state-space and hybrid recurrent LLM cache writes for interpretability.

Ax Yunfei Nie, Jiawei Wang, Ruobing Yan, Yuhan Wang, Zouxiaowei Ma, Yilun Wu 5/14/2026

Graph-Based Financial Fraud Detection with Calibrated Risk Scoring and Structural Regularization

Graph neural network approach for financial fraud detection with calibrated risk scoring and structural regularization.

Ax Andrew Y. Zhou, Sharvaree Vadgama, Sumanth Varambally, Peter Eckmann, Michael K. Gilson, Rose Yu 5/14/2026

ToolMol: Evolutionary Agentic Framework for Multi-objective Drug Discovery

ToolMol: LLM-based agentic framework using evolutionary algorithms for de novo drug discovery with improved molecular generation validity.

Ax Maximino Linares, Guillaume Doras, Thomas H\'elie 5/14/2026

Identifying the nonlinear string dynamics with port-Hamiltonian neural networks

Port-Hamiltonian neural networks for identifying nonlinear string dynamics, combining physical knowledge with neural networks.

Ax Thushari Hapuarachchi, Kaiqi Xiong 5/14/2026

SoK: A Comprehensive Analysis of the Current Status of Neural Tangent Generalization Attacks with Research Directions

Systematic review of neural tangent generalization attacks, a clean-label data poisoning technique for protecting unauthorized training data.

Ax Baris Askin, Muhammed Ustaomeroglu, Anupam Nayak, Gauri Joshi, Guannan Qu, Carlee Joe-Wong 5/14/2026

Emergent and Subliminal Misalignment Through the Lens of Data-Mediated Transfer

Study of emergent misalignment in fine-tuned LLMs, analyzing how harmful datasets cause behavioral spillover through data-mediated transfer.

Ax Ling-Qi Zhang, Kristin Branson 5/14/2026

Neurodata Without Boredom: Benchmarking Agentic AI for Data Reuse

Benchmark for evaluating agentic AI systems on data reuse and integration tasks across fragmented neuroscience datasets with diverse formats.

Ax Shixing Yu, Promit Ghosal, Kyra Gan 5/14/2026

Correcting Influence: Unboxing LLM Outputs with Orthogonal Latent Spaces

Method for token-level attribution of LLM outputs to training data using influence functions in orthogonal latent spaces for healthcare applications.

Ax Raj Kiran Gupta Katakam 5/14/2026

AGOP as Explanation: From Feature Learning to Per-Sample Attribution in Image Classifiers

AGOP-Weighted attribution method explaining image classifier predictions using Average Gradient Outer Product from feature learning theory.

Ax Benjamin Turtel, Paul Wilczewski, Kris Skotheim 5/14/2026

Training Large Language Models to Predict Clinical Events

Extension of Foresight Learning to clinical prediction using MIMIC-III longitudinal notes with natural language questions about future events.

Ax Sanya Murdeshwar, Sanjit Shashi, Kevin Bachelor, William Noid, Ashwin Lokapally, Razvan Marinescu 5/14/2026

Hessian Matching for Machine-Learned Coarse-Grained Molecular Dynamics

Machine learning framework for coarse-grained molecular dynamics augmenting force matching with Hessian matching for improved potential accuracy.

Ax Chien Van Nguyen, Chaitra Hegde, Van Cuong Pham, Ryan A. Rossi, Franck Dernoncourt, Thien Huu Nguyen 5/14/2026

Orthrus: Memory-Efficient Parallel Token Generation via Dual-View Diffusion

Orthrus framework combining autoregressive LLM fidelity with parallel diffusion model generation for efficient token generation.

Ax Leo Benac, Abhishek Sharma, Alihan Huyuk, Finale Doshi-Velez 5/14/2026

Quantifying Potential Observation Missingness in Inverse Reinforcement Learning

Method for quantifying missing observations in inverse reinforcement learning datasets, addressing data quality in behavioral analysis.

Ax Yunshu Wu, Jiayi Cheng, Longxuan Yu, Partha Thakuria, Rob Brekelmans, Evangelos E. Papalexakis, Greg Ver Steeg 5/14/2026

Discrete Stochastic Localization for Non-autoregressive Generation

Discrete Stochastic Localization framework for non-autoregressive sequence generation using continuous diffusion with unit-sphere embeddings.

Ax Kaiyang Li, Shaobo Han, Qing Su, Shihao Ji 5/14/2026

Bayesian Model Merging

Bayesian approach to merging multiple task-specific expert models without retraining, using strong anchor models as inductive bias.

Ax Truong Pham, Anay Majee, Rishabh Iyer 5/14/2026

SMA: Submodular Modality Aligner For Data Efficient Multimodal Learning

SMA framework using submodular optimization for efficient multimodal learning with limited paired data across modalities.

Ax Jordan F. McCann 5/14/2026

Descriptive Collision in Sparse Autoencoder Auto-Interpretability: When One Explanation Describes Many Features

Analysis of descriptive collision problem in sparse autoencoders where single explanations describe multiple features in language model interpretability.

Ax Blaise Delattre, Hengyu Wu, Paul Caillon, Wei Yang Bryan Lim, Yang Cao 5/14/2026

Certified Robustness under Heterogeneous Perturbations via Hybrid Randomized Smoothing

Randomized smoothing framework providing robustness certificates for multimodal models against heterogeneous perturbations across modalities.

Ax Huy Tran, Max Milkert, David Hyde 5/14/2026

ASAP: Amortized Doubly-Stochastic Attention via Sliced Dual Projection

ASAP method for efficient doubly-stochastic attention in Transformers via sliced dual projection, reducing computational cost in inference.

Ax Rohith Reddy Bellibatlu 5/14/2026

RISED: A Pre-Deployment Safety Evaluation Framework for Clinical AI Decision-Support Systems

Pre-deployment evaluation framework (RISED) for clinical AI systems covering reliability, equity, sensitivity, and deployability dimensions.

Ax Yilong Chen, Xueying Ding, Leman Akoglu 5/14/2026

VIP-COP: Context Optimization for Tabular Foundation Models

VIP-COP framework for optimizing context in tabular foundation models, enabling in-context learning on structured data with long sequences.

Ax Siyuan Liu (IIIS, Tsinghua University), Tinghong Chen (College of AI, Tsinghua University,Shanghai Qi Zhi Institute), Xinghan Li (IIIS, Tsinghua University), Yifei Wang (Amazon AGI SF Lab), Jingzhao Zhang (IIIS, Tsinghua University,Shanghai Qi Zhi Institute) 5/14/2026

Data Difficulty and the Generalization--Extrapolation Tradeoff in LLM Fine-Tuning

Systematic empirical and theoretical analysis of how data difficulty affects generalization and extrapolation in LLM fine-tuning.

Ax Changhao Li, Rushi Qiang, Jiawei Huang, Chenxiao Gao, Chao Zhang, Niao He, Bo Dai 5/14/2026

Revisiting DAgger in the Era of LLM-Agents

Study of DAgger algorithm for training long-horizon LLM agents, addressing covariate shift in multi-turn interactions with teacher supervision.

Ax Celine Lee, Jing Nathan Yan, Chen Liang, Jiaxin Shi, Yin Zhang, Jeremiah Liu, Pengcheng Yin, Fernando Pereira, Ed Chi, Derek Cheng, Alexander M. Rush, Ruoxi Wang 5/14/2026

The Efficiency Gap in Byte Modeling

Study of efficiency gap in byte-level language modeling comparing byte-level and masked diffusion approaches to traditional subword tokenization.

Ax Zhiyu Zhao, Xuejie Liu, Muhan Zhang, Anji Liu 5/14/2026

The Expressivity Boundary of Probabilistic Circuits: A Comparison with Large Language Models

Comparison of probabilistic circuits and transformers in autoregressive language modeling reveals expressivity gaps and bottlenecks.

Ax Zheng Wang, Yuang Liu, Yangkai Ding 5/14/2026

Reinforced Collaboration in Multi-Agent Flow Networks

MANGO framework for multi-agent LLM systems reducing error propagation through reinforced collaboration in workflow networks.

Ax Haodong Wu, Jiahao Zhang, Lijie Hu, Yongqi Zhang 5/14/2026

From Instance Selection to Fixed-Pool Data Recipe Search for Supervised Fine-Tuning

Fixed-pool data recipe search for SFT formulates supervised fine-tuning as optimization of operator combinations rather than instance ranking.

Ax Hongmin Li 5/14/2026

Separating Shortcut Transition from Cross-Family OOD Failure in a Minimal Model

Minimal model study separating shortcut feature reliance from OOD failure in binary classification.

Ax Hisashi Miyashita, Mgnite Inc 5/14/2026

Controlling Logical Collapse in LLMs via Algebraic Ontology Projection over F2

Algebraic Ontology Projection projects LLM hidden states into Galois Field F2 to verify and control logical relations.

Ax Feng Zhang, Xinhong Ma, Ziqiang Dong, Xi Leng, Jianfei Zhao, Xin Sun, Yang Yang, Guanjun Jiang 5/14/2026

Revisiting Reinforcement Learning with Verifiable Rewards from a Contrastive Perspective

Contrastive perspective on RLVR for LLM reasoning, reformulating GRPO as weighted positive-negative score difference.

Ax Adarsh Kumarappan, Ananya Mujoo 5/14/2026

Not Just RLHF: Why Alignment Alone Won't Fix Multi-Agent Sycophancy

Study showing multi-agent LLM sycophancy under disagreement stems from base model behavior, not RLHF, with activation patching analysis.

Ax Jihwan Kim, Chenglin Fan 5/14/2026

DP-Muon: Differentially Private Optimization via Matrix-Orthogonalized Momentum

DP-Muon formulates differentially private training using matrix-valued optimizer with momentum and orthogonalization.

Ax Rohan Surana, Gagan Mundada, Junda Wu, Xintong Li, Yizhu Jiao, Bowen Jin, Sizhe Zhou, Tong Yu, Ritwik Sinha, Jiawei Han, Jingbo Shang, Julian McAuley 5/14/2026

F-GRPO: Factorized Group-Relative Policy Optimization for Unified Candidate Generation and Ranking

F-GRPO algorithm optimizing LLMs for unified candidate generation and ranking in retrieval pipelines using group-relative policy optimization.

Ax Guiquan Sun, Xikun Zhang, Jingchao Ni, Dongjin Song 5/14/2026

\emph{DRIFT}: A Benchmark for Task-Free Continual Graph Learning with Continuous Distribution Shifts

DRIFT benchmark for task-free continual graph learning addressing catastrophic forgetting with continuous distribution shifts.

Ax Jing Yu Lim, Rushi Shah, Zarif Ikram, Samson Yu, Haozhe Ma, Tze-Yun Leong, Dianbo Liu 5/14/2026

JEDI: Joint Embedding Diffusion World Model for Online Model-Based Reinforcement Learning

Joint embedding diffusion world model for online model-based reinforcement learning balancing computational efficiency and performance.

Ax Claire Chen, Yuheng Zhang, Xinyu Liu, Zixuan Xie, Shuze Daniel Liu, Nan Jiang 5/14/2026

Offline Two-Player Zero-Sum Markov Games with KL Regularization

Theoretical framework for learning Nash equilibria in offline two-player zero-sum Markov games using KL regularization instead of explicit pessimism.

Ax Chunsan Hong, Sanghyun Lee, Chieh-Hsin Lai, Satoshi Hayakawa, Yuhta Takida, Yuki Mitsufuji, Seungryong Kim, Jong Chul Ye 5/14/2026

Understanding and Accelerating the Training of Masked Diffusion Language Models

Study analyzing why masked diffusion models train slower than autoregressive models and proposing methods to accelerate MDM training while maintaining performance.

Ax Yanggan Gu, Shuo Cai, Zihao Wang, Wenjun Wang, Yuanyi Wang, Pengkai Wang, Sirui Huang, Su Lu, Jianmin Wu, Hongxia Yang 5/14/2026

FeatCal: Feature Calibration for Post-Merging Models

FeatCal addresses performance gaps in merged models by analyzing feature drift between merged and expert models, proposing methods to calibrate features during model merging.

Ax Danny Wang, Ruihong Qiu, Zi Huang 5/14/2026

What Information Matters? Graph Out-of-Distribution Detection via Tri-Component Information Decomposition

Tide method for graph OOD detection via tri-component information decomposition to identify spurious signals in features and structure.

Ax Zihan Guan, Qiao Jin, Guangzhi Xiong, Fangyuan Chen, Mengxuan Hu, Qingyu Chen, Yifan Peng, Zhiyong Lu, Anil Vullikanti 5/14/2026

Large Language Models Lack Temporal Awareness of Medical Knowledge

Evaluation showing LLMs lack temporal awareness in medical knowledge because benchmarks are atemporal while medical knowledge continuously evolves.

Ax Minung Kim, Jeongmo Kim, Gwanwoo Choi, Seungyul Han 5/14/2026

Bridging Domain Gaps with Target-Aligned Generation for Offline Reinforcement Learning

Target-aligned Coverage Expansion framework for cross-domain offline reinforcement learning that leverages source data while reducing distribution mismatch.

Ax Aaditya L. Kachhadiya 5/14/2026

Local Inverse Geometry Can Be Amortized

Amortized learned surrogate (Deceptron) for solving nonlinear inverse problems by encoding curvature information into reusable reverse operator.

Ax Tien-Phat Nguyen, Truong Nguyen, Minh-Phuc Truong, Tuc Nguyen, James Bailey, Trung Le 5/14/2026

Spectral Flattening Is All Muon Needs: How Orthogonalization Controls Learning Rate and Convergence

Analysis of Muon optimizer showing spectral flattening mechanism enables larger learning rates and faster convergence through momentum orthogonalization.

Ax Chaokai Wu, Haofu Shi, Ningxuan Ma, Jianghong Ma, Xiaofeng Zhang 5/14/2026

MLGIB: Multi-Label Graph Information Bottleneck for Expressive and Robust Message Passing

Graph Information Bottleneck approach for multi-label graphs reducing over-squashing in deep message passing of graph neural networks.

Ax Changha Lee, Gyusang Cho 5/14/2026

ERPPO: Entropy Regularization-based Proximal Policy Optimization

Entropy regularization extension to multi-agent PPO addressing non-stationary observations in multi-dimensional cooperative reinforcement learning environments.

Ax Idan Barnea, Ofir Schlisselberg, Yishay Mansour 5/14/2026

Collaborating in Multi-Armed Bandits with Strategic Agents

Game-theoretic analysis of collaborative multi-agent bandit learning where strategic agents balance information sharing against free-riding incentives.

Ax Huiqi Deng, Yibo Li, Quanshi Zhang, Peng Zhang, Hongbin Pei, Xia Hu 5/14/2026

Understanding Generalization through Decision Pattern Shift

Decision Pattern Shift framework analyzing how deep neural network internal decision mechanisms evolve from training to test data for understanding generalization failures.

Ax Hung Le, Svetha Venkatesh 5/14/2026

Continual Fine-Tuning of Large Language Models via Program Memory

Parameter-efficient fine-tuning method for LLMs using program memory to balance rapid adaptation and knowledge retention in continual learning settings.

Ax Maxwell Standen, Junae Kim, Claudia Szabo 5/14/2026

Finding the Weakest Link: Adversarial Attack against Multi-Agent Communications

Adversarial attack methods against multi-agent reinforcement learning systems using Jacobian-based gradient information to identify vulnerable communications.

Ax Han Wang, Ying Wang, Bing Wang 5/14/2026

A Hybrid Tucker-LSTM Tensor Network Model for SOC Prediction in Electric Vehicles

arXiv paper: Hybrid Tucker-LSTM tensor network for battery state-of-charge prediction in electric vehicles.