Ax Oliver Makins, Orazio Angelini, Zohreh Shams, Mary Phuong 7/9/2026

Multi-Agent AI Control: Distributed Attacks Hamper Per-Instance Monitors

Empirical study of multi-agent AI control examining how distributed attacks across shared infrastructure can bypass per-instance safety monitors.

Ax Xufeng Zhao, Fuzhi Yang, Jianhui Chen, Li Gao, Zhang Meng, Jie Gao, Yao Zheng, Wenyu Liu, Menglin Yang, Minqi Gu, Yaru Zhao, Honglin Han, Shihui Su, Zixiao Tang, Liu Liu, Mu Xu, Yang Cai, Wenbin Tang 7/9/2026

Behavior Foundations for Quadruped Robots: ABot-C0 Technical Report

Technical report on motion controllers for quadruped robots trained from motion-capture data to bridge semantic reasoning and physical execution.

Ax Chethan Krishnamurthy Ramanaik, Tobias Callies, Michael Hecht, Eirini Ntoutsi 7/9/2026

On Adversarial Vulnerability of Vision-Language Models through the Lens of Intermediate Spectral Subspaces

Analysis of adversarial vulnerability in vision-language transformers through spectral structure of intermediate linear transformations.

Ax Tanay Sodha, Aditya Sharma, Ramya Hebbalaguppe, Vinti Agarwal, Pranav Murthy Yeluripaty 7/9/2026

When Prompts Ignore Structure: Graph-Based Attribute Reasoning for Calibrated VLMs

Graph-based attribute reasoning approach for vision-language models that improves calibration by leveraging relational structure of class attributes.

Ax Bojie Li, Noah Shi 7/9/2026

RLVP: Penalize the Path, Reward the Outcome

RL framework for training agents in costly real-world interactions that penalizes constraint violations during paths while rewarding final outcomes.

Ax Andrea Scarinci, Virginia Negri, Brayan Impata, Suleiman Khan, Victor Martinez, Marcello Federico 7/9/2026

SynthAVE: Scalable Synthetic Labeling for E-Commerce with LLM-Arena Validation

LLM-based synthetic label generation system for e-commerce attribute extraction with integrated quality control across product categories.

Ax Vin\'icius Gabriel Angelozzi, H\'eber H. Arcolezi 7/9/2026

Where to Intervene? Benchmarking Fairness-Aware Learning on Differentially Private Synthetic Tabular Data

Benchmark study examining fairness-aware learning on differentially private synthetic tabular data in high-stakes ML deployments.

Ax Harry Owiredu-Ashley 7/9/2026

Beyond Attack-Success Rate: Action-Graded Severity Scale for Tool-Using AI Agents

Severity grading framework for agentic red-teaming that replaces binary attack-success rates with ordinal harm scale for tool-using agents.

Ax Florian Fuchs, Jessy Gosselin-Grant, Boris Skuin, Michele Petteni, Alessandro Sestini, Joakim Bergdahl, Amir Baghi, Linus Gissl\'en 7/9/2026

Reward-Adaptive Iterative Discovery: A Case Study on Automated Game Testing for NHL26

Reward-adaptive discovery method for automated game testing that reduces re-testing effort by iteratively finding behavioral exploits in AI systems.

Ax Jaris K\"uken, Shi Bin Hoo, Martin Mr\'az, Frank Hutter, Lennart Purucker 7/9/2026

TimEE: End-to-end Time Series Classification via In-Context Learning

In-context learning approach for time series classification that eliminates separate feature encoder training and enables label exploitation at inference.

Ax Feng He, Zhenting Wang, Qifan Wang, Qiang Guan, Dongfang Liu, Ruixiang Tang, Qiankun Li 7/9/2026

HIVE: Understanding Post-Hallucination Reasoning in Vision Language Models

Study of how hallucinated content propagates through reasoning stages in vision-language models and affects downstream inference.

Ax Zhenyu Hou, Yujiang Li, Jie Tang, Yuxiao Dong 7/9/2026

Single-Rollout Asynchronous Optimization for Agentic Reinforcement Learning

Asynchronous reinforcement learning system for LLM post-training optimized for long-horizon agentic tasks with improved training stability.

Ax Ricardo Maia Avelino, Rita Sevastjanova, Tom Van Mele, Philippe Block, Mennatallah El-Assady 7/9/2026

Creativity from Friction: Human-AI Interaction for Exploratory Structural Design

Interactive AI agent framework for structural design that explores alternatives and refines solutions while satisfying spatial, mechanical, and cost constraints.

Ax Maximilian Andreas Hoefler, Karsten Mueller, Wojciech Samek 7/9/2026

Collaborative Synthetic Data Generation for Knowledge Transfer in Federated Learning

Federated learning approach using collaborative synthetic data generation for knowledge transfer across distributed clients with divergent data distributions.

Ax Kaicong Huang, Meng Ma, Ruimin Ke 7/9/2026

CARLA-GS: Decoupling Representation, Reasoning, and Physics Simulation for Autonomous Driving Corner-Case Synthesis

Multi-component simulator for autonomous driving that synthesizes corner cases combining visual representation, scene reasoning, and vehicle control.

Ax Mubarak Raji, Masooda Bashir 7/9/2026

Towards Agentic AI Governance: A Preliminary Assessment

Systematic review of governance challenges and frameworks for agentic AI systems capable of autonomous planning and task execution.

Ax Sahil Kale 7/9/2026

Future Confidence Distillation in Large Language Models

Method for improving confidence estimation in LLMs by tracking confidence evolution during generation for better deployment in tool use and adaptive systems.

Ax Jordan Painter, Dipankar Srirag, Adarsh Kappiyath, Diptesh Kanojia, Aditya Joshi, Lu Yin 7/9/2026

DiaLLM: An Investigation into the Robustness-Generation Gap in English Dialect Adaptation

DiaLLM: Method for improving dialectal English generation in open-weight LLMs through continual pretraining and alignment.

Ax Eric Zhu, Abhinav Shrivastava, Soumik Mukhopadhyay 7/9/2026

Selective Timestep Weighting and Advantage-Based Replay for Sample-Efficient Diffusion RLHF

Sample-efficient RLHF method for diffusion models using selective timestep weighting and advantage-based replay.

Ax Victor Giannakouris, Immanuel Trummer 7/9/2026

Breaking Database Lock-in: Agentic Regeneration of High Performance Storage Readers for Database Bypass

Jailbreak: Agentic approach to bypass database engines by directly reading storage files for high-performance columnar analytics.

Ax Yair Feldman, Linxi Zhao, Nathan Godey, Dongyoung Go, Yilun Hua, Kilian Q. Weinberger, Jennifer J. Sun, Yoav Artzi 7/9/2026

Co-LMLM: Continuous-Query Limited Memory Language Models

Continuous-query limited memory language models that externalize factual knowledge to knowledge bases during pretraining and generation.

Ax Chen Tang, Yizhou Wang, Jianyu Wu, Lintao Wang, Shixiang Tang, Pengze Li, Encheng Su, Jun Yao, Jiabei Xiao, Yuqi Shi, Jielan Li, Hongxia Hao, Zhangyang Gao, Fang Wu, Ben Fei, Xiangyu Yue, Pan Tan, Bozitao Zhong, Jinouwen Zhang, Aoran Wang, Yan Lu, Jiaheng Liu, Xinzhu Ma, Liang Hong, Mingyue Zheng, Phil Torr, Bowen Zhou, Wanli Ouyang, Lei Bai 7/9/2026

Accurate, Interdisciplinary and Transparent Structure-property Understanding with Deep Native Structural Reasoning

Framework for mechanistically explaining structure-property relationships using deep learning with physical constraints and scientific principles.

Ax Alexander Tuisov, Yonatan Vernik, Alexander Shleyfman 7/9/2026

Successor-Generator Planning with LLM-generated Heuristics

Method for domain-independent planning that uses LLMs to automatically synthesize heuristics from problem definitions.

Ax Laurens Engwegen, Max Weltevrede, Caroline Horsch, Daan Brinks, Wendelin B\"ohmer 7/9/2026

Shared Modular Recurrence in Contextual MDPs for Universal Morphology Control

Deep reinforcement learning approach for universal robot control using modular recurrence and contextual MDPs across different morphologies.

Ax Satiyabooshan Murugaboopathy, Connor T. Jerzak, Adel Daoud 7/9/2026

Platonic Representations for Poverty Mapping: Unified Vision-Language Codes or Agent-Induced Novelty?

Study using satellite imagery and LLM-generated text descriptions to investigate socioeconomic indicators in poverty mapping.

Ax Kaijian Zou, Aaron Xiong, Yunxiang Zhang, Frederick Zhang, Yueqi Ren, Jirong Yang, Ayoung Lee, Shitanshu Bhushan, Lu Wang 7/9/2026

LiveOIBench: Can Large Language Models Outperform Human Contestants in Informatics Olympiads?

LiveOIBench: Large-scale benchmark of competitive programming problems for evaluating LLM coding capabilities with comprehensive test coverage.

Ax Jaehyung Lee, Justin Ely, Kent Zhang, Akshaya Ajith, Charles Rhys Campbell, Kamal Choudhary 7/9/2026

AGAPI-Agents: An Open-Access Agentic AI Platform for Accelerated Materials Design on AtomGPT.org

AGAPI-Agents: Open-source agentic AI platform integrating LLMs with 28 scientific tools for accelerated materials design.

Ax Dongshen Peng, Yi Wang, Austin Schoeffler, Sun-ha Hong, Brian Suffoletto, David Kim, Carl Preiksaitis, Christian Rose 7/9/2026

SycoEval-EM: Sycophancy Evaluation of Large Language Models in Simulated Clinical Encounters for Emergency Care

Multi-agent simulation framework evaluating LLM robustness to adversarial persuasion in simulated clinical emergency medicine scenarios.

Ax Katherine Elkins, Jon Chun 7/9/2026

Framing Instability in LLM Ethical Stance: Auditing Negation Sensitivity in Moral Dilemmas

Audit of 16 LLMs showing instability in ethical stances when moral dilemmas are reframed as negations versus prescriptions.

Ax Kate H. Bentley, Luca Belli, Adam M. Chekroud, Emily J. Ward, Emily R. Dworkin, Emily Van Ark, Kelly M. Johnston, Will Alexander, Millard Brown, Matt Hawrilenko 7/9/2026

AI Chatbot Suicide Risk Detection and Response: Human Validation Study of the Open-Source VERA-MH Safety Evaluation

VERA-MH benchmark for validating AI chatbot safety in suicide risk detection with human evaluation studies.

Ax Jin Wang, Hui Ma, Yajun Zhang, Xinjun Pei, Ming Yan, Fei Xing, Yikun Chen 7/9/2026

An Adaptive Differentially Private Federated Learning Framework

Federated learning framework addressing device heterogeneity and non-IID data with adaptive differential privacy mechanisms.

Ax Joseph Bingham, Netanel Arussy, Dvir Aran 7/9/2026

SOMtime the World Ain$'$t Fair: Violating Fairness Using Self-Organizing Maps

Study showing sensitive attributes emerge in unsupervised embeddings even when withheld from training using self-organizing maps.

Ax Nivasini Ananthakrishnan, Meena Jagadeesan 7/9/2026

Power and Limitations of Aggregation in Compound AI Systems

Theoretical analysis of aggregating multiple LLM responses in compound AI systems and whether this unlocks new capabilities.

Ax Yiyang Fang, Wenke Huang, Pei Fu, Yihao Yang, Kehua Su, Zhenbo Luo, Jian Luan, Mang Ye 7/9/2026

EMO-R3: Reflective Reinforcement Learning for Emotional Reasoning in Multimodal Large Language Models

Method for improving emotional reasoning in multimodal LLMs using reflective reinforcement learning for better emotion understanding.

Ax David Baumgartner, Eliezer de Souza da Silva, I\~nigo Urteaga 7/9/2026

Anomaly detection in time-series via inductive biases in the latent space of conditional normalizing flows

Deep generative models for anomaly detection in multivariate time-series using normalizing flows with inductive biases in latent space.

Ax Richard Servajean, Philippe Servajean 7/9/2026

Measuring the metacognition of AI

Methods for measuring metacognitive capabilities of AI systems to assess reliability and manage uncertainty in decision-making workflows.

Ax Spandan Garg, Vikram Nitin, Yufan Huang 7/9/2026

Terminus-4B: Can a Smaller Model Replace Frontier LLMs at Agentic Execution Tasks?

Research on smaller 4B parameter models for agentic execution tasks using subagent architectural patterns to handle specialized subtasks like debugging and terminal execution.

Ax Hoyoung Lee, Suhwan Park, Seunghan Lee, Jun Seo, Jaehoon Lee, Sungdong Yoo, Minjae Kim, CheolWon Na, Zhangyang Wang, Zach Golkhou, Minkyu Kim, Sotirios Sabanis, Alejandro Lopez-Lira, Dhagash Mehta, Soonyoung Lee, Chanyeol Choi, Wonbin Ahn, Yongjae Lee 7/9/2026

When Summaries Distort Decisions: Information Fidelity in LLM-Compressed Financial Analysis

Study of how LLM-compressed financial summaries can distort investment decisions and information fidelity in agentic systems.

Ax Shei Pern Chua, Hao Wu, Qianli Ma, Fangzhao Wu 7/9/2026

HARC: Coupling Harmfulness and Refusal Directions for Robust Safety Alignment

Analysis of how aligned LLMs internally represent safety through harmfulness and refusal directions for robust alignment.

Ax Zitong Shi, Yixuan Tang, Anthony Kum Hoe Tung 7/9/2026

A-TMA: Decoupling State-Aware Memory Failures in Long-Term Agent Memory

Study of memory failures in LLM agents where conflicting state facts coexist, and methods to decouple them.

Ax Shuo Ren, Zijin Cheng, Yaohui Han, Libo Shen, Leilei Jin, Wanting Tian, Rongliang Fu, Chao Wang, Bei Yu, Tsung-Yi Ho 7/9/2026

AgenticPD: A Stage-Aware Agentic Framework for Physical Design QoR Optimization

Stage-aware agentic framework for physical design optimization avoiding full re-runs after parameter changes.

Ax Yunhan Xu, Qifeng Wu, Xunjin Li, Yuanwei Bin, Qingsong Yao, Jianghang Gu, Guan Wang, Weihao Lv, Huiyu Yang, Wenfa Luo, Jiao Xiang, Yuntian Chen, Shiyi Chen 7/9/2026

ArtisanCAD: An Industrial-Level CAD Agent with Expert-Grounded Knowledge Distillation

CAD agent for industrial component design using knowledge distillation to handle ambiguous specifications and parametric modeling.

Ax Jihao Liu, Guoxiong Gao, Zeming Sun, Bin Wu, Shurui Liu, Jiedong Jiang, Haocheng Ju, Leheng Chen, Ronnie Cheng, Xiping Zhang, Bin Dong 7/9/2026

Danus: Orchestrating Mathematical Reasoning Agents with Fact-Graph Memory

System for orchestrating mathematical reasoning agents using fact-graph memory for research-level problem solving.

Ax Zhiqiang He, Zhi Liu 7/9/2026

Silent Neuron Theory and Plasticity Preservation for Deep Reinforcement Learning in Adaptive Video Streaming

Study of neural plasticity and generalization in deep reinforcement learning for adaptive video streaming.

Ax Juyi Lin, Amir Taherin, Arash Akbari, Arman Akbari, Lei Lu, Guangyu Chen, Taskin Padir, Xiaomeng Yang, Weiwei Chen, Yiqian Li, Xue Lin, David Kaeli, Pu Zhao, Yanzhi Wang 7/9/2026

VOTE: Vision-Language-Action Optimization with Trajectory Ensemble Voting

Framework for Vision-Language-Action models reducing inference latency and improving robotic manipulation through trajectory ensemble voting.

Ax Changcun Huang 7/9/2026

Understanding Two-Layer Neural Networks with Smooth Activation Functions

Theoretical analysis of training solutions in two-layer neural networks with smooth activation functions.

Ax Luis Roque, Vitor Cerqueira, Carlos Soares, Luis Torgo 7/9/2026

L-GTA: Latent Generative Modeling for Time Series Augmentation

Generative model using VAE with Bi-LSTM for time series data augmentation in forecasting and classification.

Ax Abhishek Kolari, Mohammadhossein Khojasteh, Yifan Jiang, Floris den Hengst, Filip Ilievski 7/9/2026

A Study of Commonsense Reasoning over Visual Object Properties

Study of visual reasoning about object properties and physical attributes in VQA tasks.

Ax Fucai Ke, Joy Hsu, Zhixi Cai, Zixian Ma, Xin Zheng, Xindi Wu, Sukai Huang, Weiqing Wang, Pari Delir Haghighi, Gholamreza Haffari, Ranjay Krishna, Jiajun Wu, Hamid Rezatofighi 7/9/2026

Explain Before You Answer: A Survey on Compositional Visual Reasoning

Survey on compositional visual reasoning in multimodal AI systems for decomposing scenes and multi-step inference.

Ax Xinzhe Huang, Wenjing Hu, Tianhang Zheng, Kedong Xiu, Hongsheng Hu, Xiaojun Jia, Di Wang, Zhan Qin, Kui Ren 7/9/2026

NonTextual Target Attack

Study of gradient-based jailbreak attacks on LLMs using adversarial suffixes without fixed target constraints.