Ax Jessica M. Lundin, Usman Nasir Nakakana, Guillaume Chabot-Couture 5/18/2026

From Guidelines to Guarantees: A Graph-Based Evaluation Harness for Domain-Specific Evaluation of LLMs

Graph-based evaluation harness transforming clinical guidelines into queryable knowledge graphs for domain-specific LLM evaluation.

Ax Sihang Zeng, Yujuan Fu, Sitong Zhou, Zixuan Yu, Lucas Jing Liu, Jun Wen, Matthew Thompson, Ruth Etzioni, Meliha Yetisgen 5/18/2026

Traj-CoA: Patient Trajectory Modeling via Chain-of-Agents for Lung Cancer Risk Prediction

Traj-CoA: Multi-agent system using chain-of-agents to model patient trajectories from EHR data for lung cancer risk prediction.

Ax Augusto B. Corr\^ea, Andr\'e G. Pereira, Jendrik Seipp 5/18/2026

Frontier Large Language Models Rival State-of-the-Art Planners

Frontier LLMs now solve complex planning tasks previously thought beyond their capability, validated on International Planning Competition benchmarks.

Ax Siva Sai, Rajkumar Buyya 5/18/2026

Quantum Artificial Intelligence for Mission-Critical Systems: Foundations, Architectural Elements, and Future Directions

Survey of quantum AI applications for mission-critical systems. Architectural overview without novel empirical results.

Ax Jordan Abi Nader, David Lee, Nathaniel Dennler, Andreea Bobu 5/18/2026

QuickLAP: Quick Language-Action Preference Learning for Semi-Autonomous Systems

QuickLAP: Bayesian framework fusing physical and language feedback to learn reward functions for robotic systems in real time.

Ax Peter Jansen, Samiah Hassan, Pragnya Narasimha 5/18/2026

CodeDistiller: Automatically Generating Code Libraries for Scientific Coding Agents

CodeDistiller automatically generates code libraries from scientific discovery agents to expand their reliable capabilities beyond parametric knowledge.

Ax Yun Qu, Qi Wang, Yixiu Mao, Heming Zou, Yuhang Jiang, Weijie Liu, Clive Bai, Kai Yang, Yangkun Chen, Saiyong Yang, Xiangyang Ji 5/18/2026

Small Generalizable Prompt Predictive Models Can Steer Efficient RL Post-Training of Large Reasoning Models

Small generalizable prompt predictive models steer efficient RL post-training of large reasoning models by prioritizing informative prompts.

Ax Shutong Fan, Lan Zhang, Xiaoyong Yuan 5/18/2026

When AI Persuades: Adversarial Explanation Attacks on Human Trust in AI-Assisted Decision Making

Studies adversarial attacks on human trust through fluent LLM-generated explanations that manipulate user perception and decision-making in AI-assisted settings.

Ax Hengyuan Hu, Tingchen Fu, Minqi Jiang, Alexander H Miller, Yoram Bachrach, Jakob Nicolaus Foerster 5/18/2026

Asking the Right Questions: Improving Reasoning with Generated Stepping Stones

Method for improving LLM reasoning on hard tasks by having models generate intermediate stepping stones like simplifications and subproblems.

Ax Wael Hafez, Cameron Reid, Amit Nazeri 5/18/2026

The Informational Cost of Agency: A Bounded Measure of Interaction Efficiency for Deployed Reinforcement Learning

Bipredictability metric quantifies interaction efficiency in deployed RL systems using information-theoretic bounds on agent-environment loop conversion.

Ax Giulio Frey, Kawin Ethayarajh 5/18/2026

Mecha-nudges for Machines

Studies mecha-nudging: how internet environments can be systematically modified to influence AI agent decisions without degrading human decision-making.

Ax Zhixin Han, Yanzhi Zhang, Chuyang Wei, Maohang Gao, Xiawei Yue, Kefei Chen, Yu Zhuang, Haoxiang Guan, Jiyan He, Jian Li, Yitong Duan, Yu Shi, Mengting Hu, Shuxin Zheng 5/18/2026

FutureWorld: A Live Reinforcement Learning Environment for Predictive Agents with Real-World Outcome Rewards

FutureWorld live RL environment for predicting real-world events with LLM-based agents, enabling continual learning from actual outcomes without answer leakage.

Ax Fangming Cui, Ruixiao Zhu, Cheng Fang, Sunan Li, Jiahong Li 5/18/2026

Rethinking Agentic Reinforcement Learning In Large Language Models

Analysis of paradigm shift toward agentic reinforcement learning in LLMs for autonomous agents tackling complex open-ended tasks beyond narrow environments.

Ax Aleksandr Bowkis, Marie Davidsen Buhl, Jacob Pfau, Geoffrey Irving 5/18/2026

Automated alignment is harder than you think

Argues that automating alignment research with AI agents risks producing misleading safety assessments even without deliberate deception due to fundamental misunderstandings.

Ax Keqi Han, Songlin Zhao, Yao Su, Xiang Li, Yixuan Yuan, Lifang He, Carl Yang 5/18/2026

Towards a Virtual Neuroscientist: Autonomous Neuroimaging Analysis via Multi-Agent Collaboration

Multi-agent collaboration system for autonomous neuroimaging analysis that reasons about objectives and adapts workflows dynamically like human researchers.

Ax Prasad Mahadik, Adrians Skapars 5/18/2026

Do Linear Probes Generalize Better in Persona Coordinates?

Investigation of whether linear probes trained in persona coordinate spaces generalize better for monitoring harmful language model behaviors.

Ax Ruiyi Yang, Lihuan Li, Hao Xue, Flora D. Salim 5/18/2026

STAR: Failure-Aware Markovian Routing for Multi-Agent Spatiotemporal Reasoning

STAR framework routes among heterogeneous specialist agents in spatiotemporal reasoning by modeling different failure modes using Markovian routing.

Ax Jinhang Xu, Qiyuan Zhu, Yujun Wu, Zirui Wang, Dongxu Zhang, Marcia Tian, Yiling Duan, Siyuan Li, Jingxuan Wei, Sirui Han, Yike Guo, Odin Zhang, Conghui He, Cheng Tan 5/18/2026

NanoResearch: Co-Evolving Skills, Memory, and Policy for Personalized Research Automation

NanoResearch system co-evolves skills, memory, and policy for personalized research automation using multi-agent LLMs adapted to individual researcher preferences.

Ax Moein Hasani, Hamidreza Shahidi, Trace Levinson, Yuan Zhong, Guanghua Shu, Vinesh Gudla, Tejaswi Tenneti 5/18/2026

A Cascaded Generative Approach for e-Commerce Recommendations

Cascaded generative approach for personalized e-commerce recommendations that assembles dynamic storefronts with semantic cohesion across placements.

Ax Zahra Zanjani Foumani, Alberto Castelo, Shuang Xie, Ted Chaiwachirasak, Han Li, Lingyun Wang 5/18/2026

SimPersona: Learning Discrete Buyer Personas from Raw Clickstreams for Grounded E-Commerce Agents

SimPersona framework learns discrete buyer personas from clickstream data to enable e-commerce agents to capture heterogeneous buyer behavior instead of collapsing to average policies.

Ax Joy Bose 5/18/2026

Falkor-IRAC: Graph-Constrained Generation for Verified Legal Reasoning in Indian Judicial AI

Falkor-IRAC graph-constrained generation framework for verified legal reasoning in Indian judicial AI using LLMs with structured symbolic reasoning.

Ax Haoran Zhang, Luxin Xu, Zhilin Wang, Runquan Gui, Shunkai Zhang, Haodi Lei, Zihao He, Bingsu He, Chicheng Qin, Tong Zhu, Xiaoye Qu, Yang Yang, Yu Cheng, Yafu Li 5/18/2026

$\pi$-Bench: Evaluating Proactive Personal Assistant Agents in Long-Horizon Workflows

π-Bench benchmark for evaluating proactive personal assistant agents in long-horizon workflows with underspecified user requests.

Ax Shihao Qi, Jie Ma, Rui Xing, Wei Guo, Xiao Huang, Zhitao Gao, Jianhao Deng, Jun Liu, Lingling Zhang, Bifan Wei, Boqian Yang, Pinghui Wang, Jianwen Sun, Jing Tao, Yaqiang Wu, Hui Liu, Yu Yao, Tongliang Liu 5/18/2026

Beyond Individual Intelligence: Surveying Collaboration, Failure Attribution, and Self-Evolution in LLM-based Multi-Agent Systems

Survey of LLM-based multi-agent systems covering collaboration mechanisms, failure attribution, and self-evolution capabilities for complex workflows.

Ax Yongcun Song, Ziqi Wang, Enrique Zuazua 5/18/2026

Approximate and Weighted Data Reconstruction Attack in Federated Learning

Approximate and weighted data reconstruction attack against federated learning demonstrating privacy vulnerabilities in horizontal FL scenarios.

Ax Ravidu Suien Rammuni Silva, Jordan J. Bird 5/18/2026

FM-G-CAM: A Holistic Approach for Explainable AI in Computer Vision

FM-G-CAM extension of Grad-CAM for holistic explainability in CNN predictions beyond single target class focus.

Ax Ziyu Liu, Zeyi Sun, Yuhang Zang, Wei Li, Pan Zhang, Xiaoyi Dong, Yuanjun Xiong, Dahua Lin, Jiaqi Wang 5/18/2026

RAR: Retrieving And Ranking Augmented MLLMs for Visual Recognition

RAR framework combines CLIP's broad recognition with MLLMs' fine-grained classification via retrieval and ranking for improved visual recognition.

Ax Yue Liu, Xiaoxin He, Miao Xiong, Jinlan Fu, Shumin Deng, Yingwei Ma, Jiaheng Zhang, Bryan Hooi 5/18/2026

FlipAttack: Jailbreak LLMs via Flipping

FlipAttack exploits left-to-right text comprehension weakness in black-box LLMs using left-side noise to disguise harmful prompts.

Ax ChonLam Lao, Jiaqi Gao, Jiamin Cao, Zhipeng Zhang, Pengcheng Zhang, Jiangfei Duan, Zhilong Zheng, Yu Guan, Yichi Xu, Yong Li, Zhengping Qian, Aditya Akella, Minlan Yu, Ennan Zhai, Dennis Cai, Jingren Zhou 5/18/2026

TrainMover: An Interruption-Resilient Runtime for ML Training

TrainMover runtime enables interruption-resilient LLM training using elastic/standby machines with minimal downtime and zero memory overhead.

Ax Wenyun Li, Wenjie Huang, Chen Sun 5/18/2026

Shaping Sparse Rewards in Reinforcement Learning: A Semi-supervised Approach

Semi-supervised learning approach with data augmentation for reward shaping in sparse reward reinforcement learning environments.

Ax Wei-Yao Wang, Zhao Wang, Helen Suzuki, Yoshiyuki Kobayashi 5/18/2026

Seeing is Understanding: Unlocking Causal Attention into Modality-Mutual Attention for Multimodal LLMs

Modality-Mutual Attention mechanism addresses vision-language misalignment in multimodal LLMs to improve factual alignment between textual responses and visual inputs.

Ax Yash Akhauri, Ahmed F AbouElhamayed, Yifei Gao, Chi-Chih Chang, Sameh Gobriel, Nilesh Jain, Mohamed S. Abdelfattah 5/18/2026

TokenButler: Token Importance is Predictable

TokenButler predicts token importance in KV-Cache to identify critical tokens and reduce memory/computation bottlenecks in LLM inference.

Ax Andrea W Wen-Yi, Unso Eun Seo Jo, David Mimno 5/18/2026

Do Chinese models speak Chinese languages?

Comparative analysis of multilingual capabilities in Chinese open-weight LLMs versus US/European models, examining pre-training data curation strategies.

Ax Yuyang Xu, Renjun Hu, Haochao Ying, Jian Wu, Xing Shi, Wei Lin 5/18/2026

Large Language Models Could Be Rote Learners

Study investigates benchmark contamination in LLM evaluation, showing models can achieve inflated performance through rote learning of test data.

Ax Ruicheng Ao, Gan Luo, David Simchi-Levi, Xinshang Wang 5/18/2026

Optimizing LLM Inference: Fluid-Guided Online Scheduling with Memory Constraints

Fluid-Guided Online Scheduling optimizes LLM inference token scheduling under memory constraints to reduce latency and GPU costs in multi-user serving.

Ax Xiaoqiang Lin, Arun Verma, Zhongxiang Dai, Daniela Rus, See-Kiong Ng, Bryan Kian Hsiang Low 5/18/2026

ActiveDPO: Active Direct Preference Optimization for Sample-Efficient Alignment

ActiveDPO uses active learning to reduce annotation costs for direct preference optimization, enabling more sample-efficient LLM alignment for downstream tasks.

Ax Yuxin Ren, Maxwell D Collins, Miao Hu, Huanrui Yang 5/18/2026

FAR: Function-preserving Attention Replacement for IMC-friendly Inference

FAR framework replaces transformer attention mechanism with function-preserving alternatives optimized for in-memory computing devices to reduce latency and bandwidth overhead.

Ax Aditya Majumdar, Wenbo Zhang, Kashvi Prawal, Amulya Yadav 5/18/2026

The Hardness of Achieving Impact in AI for Social Impact Research: A Ground-Level View of Challenges & Opportunities

arXiv paper analyzing challenges and opportunities in AI for Social Impact research addressing UN Sustainable Development Goals through interdisciplinary approaches.

Ax Prabhat Nagarajan, Martha White, Marlos C. Machado 5/18/2026

Deep Double Q-learning

arXiv paper introducing Double DQN improvements to address maximization bias in deep reinforcement learning with decoupled action-selection and evaluation.

Ax Zeyu Huang, Tianhao Cheng, Zihan Qiu, Zili Wang, Yinghui Xu, Edoardo M. Ponti, Ivan Titov 5/18/2026

Blending Supervised and Reinforcement Fine-Tuning with Prefix Sampling

arXiv paper on blending supervised and reinforcement fine-tuning techniques with prefix sampling to balance generalization and performance in LLM post-training.

Ax Shengji Tang, Jianjian Cao, Weihao Lin, Jiale Hong, Bo Zhang, Shuyue Hu, Lei Bai, Tao Chen, Wanli Ouyang, Peng Ye 5/18/2026

A Scalable Multi-LLM Collaboration System with Retrieval-based Selection and Exploration-Exploitation-Driven Enhancement

arXiv paper proposing SMCS, a scalable multi-LLM collaboration system with retrieval-based selection and exploration-exploitation enhancement for coordinating multiple open-source LLMs.

Ax Mehul Damani, Isha Puri, Stewart Slocum, Idan Shenfeld, Leshem Choshen, Yoon Kim, Jacob Andreas 5/18/2026

Beyond Binary Rewards: Training LMs to Reason About Their Uncertainty

Trains LMs via RL to reason about uncertainty in natural language chains, going beyond binary reward functions for better calibration.

Ax Valeria Pais, Luis Oala, Daniele Faccio, Marco Aversa 5/18/2026

Autoguided Online Data Curation for Diffusion Model Training

Integrates autoguidance and online data selection methods to improve time and sample efficiency of diffusion model training.

Ax Sasha Cui, Zhongren Chen 5/18/2026

Painless Activation Steering: An Automated, Lightweight Approach for Post-Training Large Language Models

Automated activation steering for post-training LLMs without hand-crafted prompts, providing cheap and controllable steering alternative.

Ax Xiaolong Fu, Lichen Ma, Zipeng Guo, ShiPing Dong, Lan Yang, Tan Lit Sin, Gaojing Zhou, Yu He, Jingling Fu, Shizhe Zhou, Junshi Huang, Jason Li 5/18/2026

Dynamic-TreeRPO: Breaking the Independent Trajectory Bottleneck with Structured Sampling

Dynamic-TreeRPO improves text-to-image generation with structured tree-based sampling for reinforcement learning in flow matching models.

Ax Parth Asawa, Alan Zhu, Abigail O'Neill, Matei Zaharia, Alexandros G. Dimakis, Joseph E. Gonzalez 5/18/2026

How to Train Your Advisor: Steering Black-Box LLMs with Advisor Models

Advisor Models trains small open-weight models to generate dynamic natural language advice improving black-box LLM performance by up to 27.4%.

Ax Taehan Kim, Sangdae Nam 5/18/2026

SAE-RNA: A Sparse Autoencoder Model for Interpreting RNA Language Model Representations

Applies sparse autoencoders to RNA language model representations for interpretability, extending protein LM analysis to RNA domain.

Ax Qing Huang, Zhipei Xu, Xuanyu Zhang, Xiangyu Yu, Jian Zhang 5/18/2026

UniShield: An Adaptive Multi-Agent Framework for Unified Forgery Image Detection and Localization

UniShield is an adaptive multi-agent framework for detecting and localizing forgery in images using domain-specific methods.

Ax Yutao Wu, Xiao Liu, Yinghui Li, Yifeng Gao, Yifan Ding, Jiale Ding, Xiang Zheng, Xingjun Ma 5/18/2026

ADMIT: Few-shot Knowledge Poisoning Attacks on RAG-based Fact Checking

ADMIT demonstrates few-shot knowledge poisoning attacks on RAG-based fact-checking systems, showing how adversarial content tricks LLMs.

Ax Mengqi Li, Lei Zhao, Anthony Man-Cho So, Ruoyu Sun, Xiao Li 5/18/2026

A Model Can Help Itself: Reward-Free Self-Training for LLM Reasoning

Self-evolving Post-Training (SePT) method enabling LLMs to improve reasoning performance without external rewards using self-generated responses.

Ax Qiwei Ma, Xukun Lu, Wang Liu, Puhong Duan, Xudong Kang, Shutao Li 5/18/2026

SARVLM: A Vision Language Foundation Model for Semantic Understanding in SAR Imagery

Vision-language foundation model for SAR imagery using self-supervised learning and masked image modeling for semantic understanding.