Ax Xi Yang, Aurelie Lozano, Naoki Abe, Bhavya, Saurabh Jha, Noah Zheutlin, Rohan R. Arora, Yu Deng, Daby M. Sow 3/24/2026

A Context Engineering Framework for Improving Enterprise AI Agents based on Digital-Twin MDP

Framework for improving LLM-based enterprise agents via offline RL and digital-twin simulation without extensive real-world data.

Ax Xiao Han, Yuzheng Fan, Sendong Zhao, Haochun Wang, Bing Qin 3/24/2026

GSEM: Graph-based Self-Evolving Memory for Experience Augmented Clinical Reasoning

GSEM: Graph-based memory framework for clinical reasoning agents that organizes experiences with relational structure.

Ax Syed Usama Imtiaz, Mitra Nasr Azadani, Nasrin Alamdari 3/24/2026

SpecTM: Spectral Targeted Masking for Trustworthy Foundation Models

SpecTM: Physics-informed masking approach for foundation models in Earth observation with trustworthiness constraints.

Ax Jack W O'Sullivan, Mohammad Asadi, Lennart Elbe, Akshay Chaudhari, Tahoura Nedaee, Francois Haddad, Michael Salerno, Li Fe-Fei, Ehsan Adeli, Rima Arnaout, Euan A Ashley 3/24/2026

MARCUS: An agentic, multimodal vision-language model for cardiac diagnosis and management

MARCUS: Agentic multimodal vision-language model for automated cardiac diagnosis from ECGs and echocardiograms.

Ax Wenwen Li, Yuanyuan Tian, Sizhe Wang, Amber Wutich, Paul Westerhoff, Sarah Porter, Anais Roque, Jobayer Hossain, Patrick Thomson, Rhett Larson, Michael Hanemann 3/24/2026

Measuring Research Convergence in Interdisciplinary Teams Using Large Language Models and Graph Analytics

Framework using LLMs and graph analytics to analyze how interdisciplinary research teams develop shared knowledge over time.

Ax Xu Zhao, Xiting Wang, Weiran Shen 3/24/2026

Enhancing Safety of Large Language Models via Embedding Space Separation

Safety enhancement technique exploiting linear separability of harmful and safe query embeddings in LLMs to improve robustness against adversarial prompts.

Ax Hyunjun Jeon, Kyuyoung Kim, Jinwoo Shin 3/24/2026

RedacBench: Can AI Erase Your Secrets?

RedacBench: comprehensive benchmark for evaluating LLM-based redaction systems' ability to remove sensitive information from unstructured text.

Ax Hengwei Ye, Yuanting Guan, Yuxuan Ge, Tianying Zhu, Zhenhan Guan, Yijia Zhong, Yijing Zhang, Han Zhang, Yingna Wu, Zheng Tian 3/24/2026

Children's Intelligence Tests Pose Challenges for MLLMs? KidGym: A 2D Grid-Based Reasoning Benchmark for MLLMs

KidGym benchmark inspired by Wechsler Intelligence Scales evaluating multimodal LLMs on 2D grid-based reasoning tasks measuring interpretable abilities.

Ax Roy Uziel, Omer Belhasin, Itay Levi, Akhiad Bercovich, Ran El-Yaniv, Ran Zilberstein, Michael Elad 3/24/2026

CRoCoDiL: Continuous and Robust Conditioned Diffusion for Language

CRoCoDiL: continuous semantic space diffusion model for language addressing token dependencies and semantic coherence in masked diffusion models.

Ax Bahare Riahi, Ally Limke, Xiaoyi Tian, Viktoriia Storozhevykh, Sayali Patukale, Tahreem Yasir, Khushbu Singh, Jennifer Chiu, Nicholas lytle, Tiffany Barnes, Veronica Catete 3/24/2026

Exploring Teacher-Chatbot Interaction and Affect in Block-Based Programming

Qualitative study of middle school teachers using LLM chatbots in block-based programming environments examining interaction patterns and affect.

Ax Tatiana Chakravorti, Pranav Narayanan Venkit, Sourojit Ghosh, Sarah Rajtmajer 3/24/2026

Beyond Detection: Governing GenAI in Academic Peer Review as a Sociotechnical Challenge

Mixed-method study examining governance of GenAI in academic peer review through social media analysis and area chair interviews.

Ax Michael Hersche, Nicolas Menet, Ronan Tanios, Abbas Rahimi 3/24/2026

Locally Coherent Parallel Decoding in Diffusion Language Models

Locally Coherent Parallel Decoding method improving diffusion language models' token generation by capturing joint dependencies instead of marginal distributions.

Ax Kushal Khemani 3/24/2026

Inference Energy and Latency in AI-Mediated Education: A Learning-per-Watt Analysis of Edge and Cloud Models

Empirical comparison of on-device LLM inference configurations measuring latency-energy-learning tradeoffs for AI tutoring using Phi-3 Mini.

Ax Drake Mullens, Stella Shen 3/24/2026

The Arrival of AGI? When Expert Personas Exceed Expert Benchmarks

Analysis of expert personas in LLMs revealing why Wharton study found null results; demonstrates structural predictability of outcome through baseline contamination.

Ax Eric Gong, Nathan E. Sanders, Bruce Schneier 3/24/2026

Characterizing the ability of LLMs to recapitulate Americans' distributional responses to public opinion polling questions across political issues

Framework evaluating LLMs' ability to simulate Americans' political opinion distributions across policy issues for polling augmentation.

Ax Ahmed Abouelazm, Jonas Michel, Daniel Bogdoll, Philip Sch\"orner, J. Marius Z\"ollner 3/24/2026

Beyond Scalar Rewards: Distributional Reinforcement Learning with Preordered Objectives for Safe and Reliable Autonomous Driving

Preordered Multi-Objective MDP framework for autonomous driving addressing conflicting objectives like safety and efficiency without scalar reward collapse.

Ax Dang Nguyen, Harvey Yiyun Fu, Peter West, Chenhao Tan, Ari Holtzman 3/24/2026

Email in the Era of LLMs

HR Simulator game studies LLM email generation for workplace communication, comparing human and model outputs on social task performance.

Ax Saadi Lahlou (London School of Economics and Political Science, Paris Institute for Advanced Study), Annabelle Gouttebroze (London School of Economics and Political Science), Atrina Oraee (London School of Economics and Political Science), Julian Madera (London School of Economics and Political Science) 3/24/2026

Writing literature reviews with AI: principles, hurdles and some lessons learned

Qualitative study comparing literature reviews generated by LLMs with varying corpus selections, revealing biases and factual gaps in outputs.

Ax Michal Olak, Tommaso Boccato, Matteo Ferrante 3/24/2026

Decoding the decoder: Contextual sequence-to-sequence modeling for intracortical speech decoding

Sequence-to-sequence decoder for brain-computer interfaces converting intracortical neural activity to speech with robustness improvements.

Ax Tianshu Zhang, Huan Sun 3/24/2026

SciNav: A General Agent Framework for Scientific Coding Tasks

SciNav: Agent framework built on LLMs for autonomous scientific coding tasks with objective evaluation through executable benchmarks.

Ax Zequn Liu, Kehan Wu, Shufang Xie, Zekun Guo, Wei Zhang, Tao Qin, Renhe Liu, Yingce Xia 3/24/2026

Deciphering Scientific Reasoning Steps from Outcome Data for Molecule Optimization

DESRO framework uses LLMs to infer intermediate scientific reasoning steps from experimental outcomes to enable automated scientific discovery and molecule optimization.

Ax Islam Guven, Mehmet Parlak 3/24/2026

JCAS-MARL: Joint Communication and Sensing UAV Networks via Resource-Constrained Multi-Agent Reinforcement Learning

Multi-agent reinforcement learning framework for coordinating UAV networks with joint communication and sensing under resource constraints for monitoring missions.

Ax Saeed Khaki, Nima Safaei, Kamal Ginotra 3/24/2026

Understanding Pruning Regimes in Vision-Language Models Through Domain-Aware Layer Selection

Domain-aware layer pruning analysis of vision-language models with focus on math reasoning and perception coupling.

Ax Zhuofeng Li, Dongfu Jiang, Xueguang Ma, Haoxiang Zhang, Ping Nie, Yuyu Zhang, Kai Zou, Jianwen Xie, Yu Zhang, Wenhu Chen 3/24/2026

OpenResearcher: A Fully Open Pipeline for Long-Horizon Deep Research Trajectory Synthesis

Fully open-source pipeline for synthesizing long-horizon research trajectories for training deep research agents using reproducible methods.

Ax Alex Popa, Adrian Taylor, Ranwa Al Mallah 3/24/2026

Learning Communication Between Heterogeneous Agents in Multi-Agent Reinforcement Learning for Autonomous Cyber Defence

Multi-agent RL with learned communication protocols for heterogeneous agent coordination in autonomous cyber defense systems.

Ax Jussi Keppo, Yuze Li, Gerry Tsoukalas, Nuo Yuan 3/24/2026

On the Fragility of AI Agent Collusion

Empirical analysis of algorithmic collusion fragility in LLM agents with heterogeneous parameters using 2000+ compute hours of experiments.

Ax Dong Li, Zhengzhang Chen, Xujiang Zhao, Linlin Yu, Zhong Chen, Yi He, Haifeng Chen, Chen Zhao 3/24/2026

MARLIN: Multi-Agent Reinforcement Learning for Incremental DAG Discovery

Multi-agent RL approach for incremental causal DAG discovery from observational data with improved efficiency for online applications.

Ax Jing Liu, Zhenchao Ma, Han Yu, Bobo Ju, Wenliang Yang, Chengfang Li, Bo Hu, Liang Song 3/24/2026

Collaborative Adaptive Curriculum for Progressive Knowledge Distillation

Collaborative knowledge distillation with adaptive curriculum for heterogeneous distributed learning in edge-based visual analytics.

Ax Yusen Wu, Xiaotie Deng 3/24/2026

HCAG: Hierarchical Abstraction and Retrieval-Augmented Generation on Theoretical Repositories with LLMs

RAG method using hierarchical code abstraction and architecture-guided retrieval for complex theory-driven codebases in algorithmic game theory.

Ax Shaolin Wang, Yi Mei, Haoyang Che, He Jiang, Shui Yu, Ying Gu 3/24/2026

From Human Interfaces to Agent Interfaces: Rethinking Software Design in the Age of AI-Native Systems

Argues for redesigning software interfaces from human-centric to agent-centric as LLM agents become primary consumers of systems.

Ax Aayam Bansal, Ishaan Gangwani 3/24/2026

Reason-to-Transmit: Deliberative Adaptive Communication for Cooperative Perception

Framework for cooperative autonomous agents to decide what V2X network messages to transmit based on reasoning about receiver benefit.

Ax Rohan Siva, Kai Cheung, Lichi Li, Ganesh Sundaram 3/24/2026

kRAIG: A Natural Language-Driven Agent for Automated DataOps Pipeline Generation

Natural language-driven LLM agent (kRAIG) for automating data pipeline generation and ETL workflow construction.

Ax Sarat Mudunuri, Jian Wan, Ally Qin, Srinivasan Manoharan 3/24/2026

Semantic Tool Discovery for Large Language Models: A Vector-Based Approach to MCP Tool Selection

Vector-based semantic approach to automatically discover and select relevant tools from large MCP server toolsets for LLM agents.

Ax Sasan Mansouri, Edoardo Pilla, Mark Wahrenburg, Fabian Woebbeking 3/24/2026

Bypassing Document Ingestion: An MCP Approach to Financial Q&A

Comparison of Model Context Protocol (MCP) versus RAG for financial Q&A, showing direct system integration outperforms document retrieval.

Ax Shasha Yu, Fiona Carroll, Barry L. Bentley 3/24/2026

The Causal Impact of Tool Affordance on Safety Alignment in LLM Agents

Empirical study of how executable tool access changes safety alignment in LLM agents, moving beyond text-centric safety evaluations.

Ax Fujian Jia, Jiwen Gu, Cheng Lu, Dezhi Zhao, Mengjiang Huang, Yuanzhi Lu, Xin Liu, Kang Liu 3/24/2026

GIP-RAG: An Evidence-Grounded Retrieval-Augmented Framework for Interpretable Gene Interaction and Pathway Impact Analysis

GIP-RAG: Retrieval-augmented framework for interpretable gene interaction and pathway analysis enabling multi-step reasoning across biological networks.

Ax Artem Maryanskyy 3/24/2026

When Agents Disagree: The Selection Bottleneck in Multi-Agent LLM Pipelines

Analysis of multi-agent LLM pipeline contradictions: identifies selection bottleneck threshold determining whether agent diversity improves output quality.

Ax Oleksii Bychkov 3/24/2026

Bounded Coupled AI Learning Dynamics in Tri-Hierarchical Drone Swarms

Theoretical analysis of coupled learning dynamics in heterogeneous tri-hierarchical drone swarm systems operating at different timescales.

Ax Yu-Ning Qiu, Lin-Feng Zou, Jiong-Da Wang, Xue-Rong Yuan, Wang-Zhou Dai 3/24/2026

Procedural Refinement by LLM-driven Algorithmic Debugging for ARC-AGI-2

LLM-driven algorithmic debugging framework for code generation using formal program debugging theory to improve ARC-AGI-2 task performance.

Ax Yao Tian, Zhoujin Tian, Xi Zhao, Ruiyuan Zhang, Xiaofang Zhou 3/24/2026

GEM: A Native Graph-based Index for Multi-Vector Retrieval

GEM: Native graph-based indexing algorithm for multi-vector retrieval enabling finer-grained semantic matching beyond single-vector approaches.

Ax Zijian Lu, Yiping Zuo, Yupeng Nie, Xin He, Weibei Fan, Chen Dai 3/24/2026

ContractSkill: Repairable Contract-Based Skills for Multimodal Web Agents

ContractSkill: Framework for creating repairable multimodal web agent skills with explicit contracts specifying preconditions, steps, and success criteria.

Ax Yizhe Zhao, Yongjian Fu, Zihao Feng, Hao Pan, Yongheng Deng, Yaoxue Zhang, Ju Ren 3/24/2026

MANA: Towards Efficient Mobile Ad Detection via Multimodal Agentic UI Navigation

MANA: Agentic multimodal framework using UI navigation and reasoning for detecting mobile ads, combining static and runtime analysis.

Ax Yuxuan He, Chaiming Huang, Yifan Wu, Hongjun Wang, Chenkui Shen, Jifan Zhang, Long Li 3/24/2026

Leum-VL Technical Report

Leum-VL technical report on multimodal model for analyzing short video structure and temporal organization beyond scene description.

Ax Vicen\c{c} Torra, Maria Bras-Amor\'os 3/24/2026

Memory poisoning and secure multi-agent systems

Analysis of memory poisoning attacks on agentic AI and multi-agent systems, covering semantic, episodic, and short-term memory vulnerabilities.

Ax Xuanwang Zhang, Yuteng Han, Jinnan Qi, Mulong Xie, Zhen Wu, Xinyu Dai 3/24/2026

WebNavigator: Global Web Navigation via Interaction Graph Retrieval

WebNavigator uses interaction graph retrieval to enable web navigation agents to overcome topological blindness via global environment structure.

Ax Christopher J. Agostino, Nayan D'Souza 3/24/2026

ALARA for Agents: Least-Privilege Context Engineering Through Portable Composable Multi-Agent Teams

ALARA framework for managing multi-agent systems with least-privilege context isolation and composable team coordination.

Ax Men Niu, Xinxin Fan, Quanliang Jing, Shaoye Luo, Yunfeng Lu 3/24/2026

CAMA: Exploring Collusive Adversarial Attacks in c-MARL

Study of adversarial attacks targeting cooperative multi-agent reinforcement learning systems in real-world applications.

Ax Y. Sungtaek Ju 3/24/2026

SymCircuit: Bayesian Structure Inference for Tractable Probabilistic Circuits via Entropy-Regularized Reinforcement Learning

SymCircuit uses RL to learn probabilistic circuit structures via entropy-regularized search instead of greedy algorithms.

Ax Yichun Xu, Navjot K. Khaira, Tejinder Singh 3/24/2026

KV Cache Optimization Strategies for Scalable and Efficient LLM Inference

Research on optimizing KV cache memory usage in Transformer-based LLMs to handle longer context windows efficiently during inference.

Ax Ata Poyraz Turna, Asrin Efe Yorulmaz, Tamer Ba\c{s}ar 3/24/2026

Meta-Learning for Repeated Bayesian Persuasion

Meta-learning algorithms for repeated Bayesian persuasion exploiting structural similarity across strategic interaction tasks.