Ax Tianyu Liu, Jirui Qi, Mrinmaya Sachan, Ryan Cotterell, Raquel Fern\'andez, Arianna Bisazza 5/11/2026

Optimizing Language Models for Crosslingual Knowledge Consistency

Method using reinforcement learning with structured rewards to improve consistency of LLM knowledge across multiple languages.

Ax Chengwei Wu 5/11/2026

VDCook:DIY video data cook your MLLMs

VDCook platform for automated video dataset construction using natural language queries, retrieval, and synthesis for multimodal LLM training.

Ax Yanjun Chen, Yirong Sun, Hanlin Wang, Jinghan Wang, Xinming Zhang, Xiaoyu Shen, Wenjie Li, Wei Zhang 5/11/2026

Exact Is Easier: Credit Assignment for Cooperative LLM Agents

Research on credit assignment in cooperative multi-agent LLM systems, showing standard evaluation methods are flawed and proposing exact counterfactual approaches.

Ax Leonardo Cesani, Matteo Papini, Marcello Restelli 5/11/2026

How Log-Barrier Helps Exploration in Policy Optimization

Theoretical analysis of log-barrier regularization enabling exploration in stochastic gradient bandit optimization with convergence guarantees.

Ax Kaito Baba, Risa Kishikawa, Satoshi Kodera 5/11/2026

Multi-Modal Multi-Agent Reinforcement Learning for Radiology Report Generation

MARL-Rad uses multi-modal multi-agent RL for radiology report generation, training agentic workflows end-to-end within deployed clinical workflow.

Ax Gabriel Orlanski, Devjeet Roy, Alexander Yun, Changho Shin, Alex Gu, Albert Ge, Dyah Adila, Nicholas Roberts, Frederic Sala, Aws Albarghouthi 5/11/2026

SlopCodeBench: Benchmarking How Coding Agents Degrade Over Long-Horizon Iterative Tasks

SlopCodeBench evaluates coding agents on iterative 36-problem benchmark with 196 checkpoints where agents extend own solutions over time.

Ax Lihao Sun, Lewen Yan, Xiaoya Lu, Andrew Lee, Jie Zhang, Jing Shao 5/11/2026

Valence-Arousal Subspace in LLMs: Circular Emotion Geometry and Multi-Behavioral Control

Analysis of emotion vector organization in LLMs via valence-arousal subspace with circular geometry for emotion steering control.

Ax Daniele Solombrino, Antonio Andrea Gargiulo, Alessandro Zirilli, Luca Zhou, Adrian Robert Minut, Emanuele Rodol\`a 5/11/2026

Zero-Shot Quantization via Weight-Space Arithmetic

Zero-shot quantization via weight-space arithmetic extracts quantization vectors from donor tasks to improve post-training quantization without receiver training data.

Ax Gabriel Peyr\'e 5/11/2026

Muon Dynamics as a Spectral Wasserstein Flow

Analysis of Muon optimizer dynamics as spectral Wasserstein flow in mean-field regime, studying gradient normalization for deep learning.

Ax Guo Gan, Yuxuan Ding, Cong Chen, Yuwei Ren, Yin Huang, Hong Zhou 5/11/2026

Android Coach: Improve Online Agentic Training Efficiency with Single State Multiple Actions

Android Coach improves online RL agent training efficiency by processing multiple actions per state, addressing sample inefficiency in emulator-based learning.

Ax Xue Qin, Simin Luan, John See, Cong Yang, Zhijun Li 5/11/2026

Governed Capability Evolution: Lifecycle-Time Compatibility Checking and Rollback for AI-Component-Based Systems, with Embodied Agents as Case Study

Framework for versioned AI component governance in embodied agent systems with compatibility checking and rollback at deployment lifecycle.

Ax Andrey Bocharnikov, Ivan Ermakov, Denis Kuznedelev, Vyacheslav Zhdanovskiy, Yegor Yershov 5/11/2026

KV Cache Offloading for Context-Intensive Tasks

Technique for offloading KV cache to reduce memory and latency bottlenecks in long-context LLM inference while maintaining accuracy.

Ax Yecheng Wu, Song Han, Hai Cai 5/11/2026

Lightning OPD: Efficient Post-Training for Large Reasoning Models with Offline On-Policy Distillation

Lightning OPD enables efficient offline on-policy distillation for LLM post-training by precomputing teacher log-probabilities without live server.

Ax Sichun Luo, Yi Huang, Haochen Luo, Fengyuan Liu, Guanzhi Deng, Lei Li, Qinghua Yao, Zefa Hu, Junlan Feng, Qi Liu 5/11/2026

SeaEvo: Advancing Algorithm Discovery with Strategy Space Evolution

SeaEvo uses LLM-guided evolutionary search for algorithm discovery with persistent population-level state management and structured reasoning.

Ax Zichao Wei 5/11/2026

Structural Generalization on SLOG without Hand-Written Rules

Neural cellular automaton with discrete bottleneck learns compositional semantic parsing rules without hand-written rules, enabling structural generalization.

Ax Siyuan Huang, Xiaoye Qu, Yafu Li, Tong Zhu, Zefeng He, Muxin Fu, Daizong Liu, Wei-Long Zheng, Yu Cheng 5/11/2026

Persistent Visual Memory: Sustaining Perception for Deep Generation in LVLMs

Persistent Visual Memory module strengthens visual attention in autoregressive LVLMs to prevent signal dilution during long sequence generation.

Ax Daniel Song, Peter Ney, Cristina Menghini, Faizan Ahmad, Aidan Boyd, Nathaniel Li, Ziwen Han, Jean-Christophe Testud, Saisuke Okabayashi, Maeve Ryan, Jinpeng Miao, Hamza Kwisaba, Felix Binder, Spencer Whitman, Jim Gust, Esteban Arcaute, Dhaval Kapil, Jacob Kahn, Ayaz Minhas, Tristan Goodman, Lauren Deason, Alexander Vaughan, Shengjia Zhao, Summer Yue 5/11/2026

Code World Model Preparedness Report

Meta's Code World Model preparedness assessment for frontier risks including code generation, reasoning, and misalignment evaluation.

Ax Iman Sharifi, Hyeong Tae Kim, Maheed Hatem Ahmed, Mahsa Ghasemi, Peng Wei 5/11/2026

Separation Assurance between Heterogeneous Fleets of Small Unmanned Aerial Systems via Multi-Agent Reinforcement Learning

Multi-agent RL approach for tactical deconfliction between heterogeneous unmanned aerial system fleets in dense urban airspace.

Ax Ruiyang Qin, Qingzhuo Wang, Dongrui Liu, Qiang Li, Zhihua Wei, Wen Shen 5/11/2026

Multilingual Safety Alignment via Self-Distillation

Cross-lingual safety alignment framework using self-distillation to transfer safeguards from high-resource to low-resource languages in LLMs.

Ax Hongbo Jin, Rongpeng Zhu, Zhongjing Du, Xu Jiang, Jingqi Tian, Qiaoman Zhang, Jiayu Ding 5/11/2026

DGPO: Distribution Guided Policy Optimization for Fine Grained Credit Assignment

DGPO improves RL credit assignment for LLM alignment on complex reasoning via distribution-guided policy optimization with finer-grained step isolation.

Ax Siqiao Xue, Zihan Liao, Jin Qin, Ziyin Zhang, Yixiang Mu, Fan Zhou, Hang Yu 5/11/2026

Beyond Retrieval: A Multitask Benchmark and Model for Code Search

CoREB introduces a contamination-limited multitask benchmark for code search beyond first-stage retrieval, including reranking and developer-style queries.

Ax Bo Li, Chuan Wu, shaolin Zhu 5/11/2026

MACS: Modality-Aware Capacity Scaling for Efficient Multimodal MoE Inference

MACS improves multimodal MoE inference efficiency using modality-aware capacity scaling to address straggler effects in expert parallelism.

Ax Junran Wang, Xinjie Shen, Zehao Jin, Pan Li 5/11/2026

How Far Are VLMs from Privacy Awareness in the Physical World? An Empirical Study

Empirical study evaluating privacy awareness of Vision-Language Models when deployed as autonomous cognitive cores in physical environments.

Ax Dongfang Zhao 5/11/2026

EGA: Adapting Frozen Encoders for Vector Search with Bounded Out-of-Distribution Degradation

EGA proposes a residual adapter for vector search systems to handle distribution shift with frozen vision encoders.

Ax Huchen Yang, Xinghao Dong, Dan Negrut, Jin-Long Wu 5/11/2026

Active Learning for Communication Structure Optimization in LLM-Based Multi-Agent Systems

Active learning method optimizes communication structure in LLM-based multi-agent systems to improve performance and reduce token usage with limited budgets.

Ax Md Anwar Hossen, Fatema Siddika, Juan Pablo Munoz, Tanya Roosta, Ali Jannesari 5/11/2026

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

CRAFT proposes continual learning for LLMs via low-rank interventions on hidden representations to prevent catastrophic forgetting during fine-tuning.

Ax Kuanwei Lin, Wenhao Zhang, Ge Li 5/11/2026

VideoRouter: Query-Adaptive Dual Routing for Efficient Long-Video Understanding

VideoRouter addresses efficiency bottlenecks in long video understanding using query-adaptive dual routing to compress visual tokens dynamically.

Ax Guoxin Lu, Letian Sha, Qing Wang, Peijie Sun, Hao Zhou, Hua Dai, Fu Xiao 5/11/2026

Safety Anchor: Defending Harmful Fine-tuning via Geometric Bottlenecks

Safety Anchor defends LLM safety alignment against harmful fine-tuning by introducing geometric bottlenecks in parameter space.

Ax Hao Lin, Kunyang Lv, Xu Jiang, Jingqi Tian, Zhongjing Du, Jiayu Ding, Qiaoman Zhang, Hongbo Jin 5/11/2026

VISD: Enhancing Video Reasoning via Structured Self-Distillation

VISD: structured self-distillation for video reasoning combines RLVR and token-level credit assignment in VideoLLMs.

Ax Abdelghani Ghanem, Mounir Ghogho 5/11/2026

Entropy-Regularized Adjoint Matching for Offline Reinforcement Learning

Entropy-regularized adjoint matching for offline RL with flow-matching policies reduces popularity bias in behavior distribution.

Ax Roussel Desmond Nzoyem, Mauro Comi 5/11/2026

Render, Don't Decode: Weight-Space World Models with Latent Structural Disentanglement

NOVA world model framework representing state as neural network weights with latent structural disentanglement for interpretability.

Ax Dijia Zhan, Jinyi Li, Chenxi Zheng, Shaoyu Huang, Yong Li, Jie Tang, Xuemiao Xu 5/11/2026

NavOne: One-Step Global Planning for Vision-Language Navigation on Top-Down Maps

NavOne: top-down vision-language navigation using one-step global planning on maps for efficient path reasoning.

Ax Nan Jia, Haojin Yang, Xing Ma, Jiesong Lian, Shuailiang Zhang, Weipeng Zhang, Ke Zeng, Xunliang Cai, Zequn Sun 5/11/2026

Asymmetric On-Policy Distillation: Bridging Exploitation and Imitation at the Token Level

Asymmetric on-policy distillation improves token-level student learning by addressing variance and exploration issues.

Ax Xiang Li, Nan Jiang 5/11/2026

Q-MMR: Off-Policy Evaluation via Recursive Reweighting and Moment Matching

Q-MMR framework for off-policy evaluation in MDPs using moment matching and recursive reweighting of trajectories.

Ax Nithin Somasekharan, Rabi Pathak, Manushri Dhanakoti, Tingwen Zhang, Ling Yue, Andy Zhu, Shaowu Pan 5/11/2026

AI CFD Scientist: Toward Open-Ended Computational Fluid Dynamics Discovery with Physics-Aware AI Agents

AI CFD Scientist: open-source AI agent for autonomous computational fluid dynamics discovery using physics-aware reasoning.

Ax Fei Zuo, Zikang Zhou, Hao Cong, Xiaoyan Xi, Ho Fai Leung 5/11/2026

RateQuant: Optimal Mixed-Precision KV Cache Quantization via Rate-Distortion Theory

RateQuant applies rate-distortion theory for mixed-precision KV cache quantization with per-head bit allocation.

Ax Enshuai Zhou, Yifan Hao, Chao Wang, Rui Zhang, Di Huang, Jiaming Guo, Xing Hu, Zidong Du, Qi Guo, Yunji Chen 5/11/2026

LKV: End-to-End Learning of Head-wise Budgets and Token Selection for LLM KV Cache Eviction

LKV learns head-wise budgets and token selection for KV cache eviction in long-context LLM inference optimization.

Ax Yubo Jiang, Yitong An, Xin Yang, Abudukelimu Wuerkaixi, Xuxin Cheng, Fengying Xie, Zhiguo Jiang, Cao Liu, Ke Zeng, Haopeng Zhang 5/11/2026

Breaking the Illusion: When Positive Meets Negative in Multimodal Decoding

Positive-and-Negative Decoding framework reducing object hallucination in vision-language models via attention rebalancing.

Ax Chenxi Tao, Seung-Kyum Choi 5/11/2026

On the Role of Strain and Vorticity in Numerical Integration Error for Flow Matching

Analysis of strain and vorticity in velocity field Jacobians to understand integration error in flow matching models.

Ax Lorenzo Riccardo Allegrini, Geremia Pompei 5/11/2026

A Hierarchical Ensemble Pipeline for Anomaly Detection in ESA Satellite Telemetry

Hierarchical ensemble pipeline for anomaly detection in ESA satellite telemetry using shapelet and statistical features.

Ax Benjamin L. Badger, Ethan Roland 5/11/2026

Toeplitz MLP Mixers are Low Complexity, Information-Rich Sequence Models

Toeplitz MLP Mixer architecture for sequence modeling with O(dn log n) complexity as efficient transformer alternative.

Ax Obed Irakoze, Prasenjit Mitra 5/11/2026

Transformer-Based Wildlife Species Classification from Daily Movement Trajectories

Transformer-based sequence models for wildlife species classification from GPS movement trajectories, outperforming LSTMs and CNNs.

Ax Likhita Yerra (AIVANCITY School of AI and Data), Remi Uttejitha Allam (AIVANCITY School of AI and Data) 5/11/2026

Semantic State Abstraction Interfaces for LLM-Augmented Portfolio Decisions: Multi-Axis News Decomposition and RL Diagnostics

Semantic State Abstraction Interfaces for mapping news text to interpretable coordinates in LLM-augmented portfolio decision systems.

Ax Nadav Timor, Ravid Shwartz-Ziv, Micah Goldblum, Yann LeCun, David Harel 5/11/2026

On Training in Imagination

Analysis of model-based RL training on imagined trajectories, quantifying effects of dynamics and reward model errors on policy optimization.

Ax Jinqian Chen, Chang Liu, Jihua Zhu 5/11/2026

Beyond Factor Aggregation: Gauge-Aware Low-Rank Server Representations for Federated LoRA

Gauge-aware aggregation method for federated LoRA addressing representation dependence in decentralized LLM adaptation.

Ax Kuo-Chung Peng, Samuel Yen-Chi Chen, Jiun-Cheng Jiang, Chen-Yu Liu, En-Jui Kuo, Yun-Yuan Wang, Prayag Tiwari, Andrea Ceschini, Chi-Sheng Chen, Yu-Chao Hsu, Chun-Hua Lin, Tai-Yue Li, Antonello Rosato, Massimo Panella, Simon See, Saif Al-Kuwari, Kuan-Cheng Chen, Nan-Yow Chen, Hsi-Sheng Goan 5/11/2026

Gated QKAN-FWP: Scalable Quantum-inspired Sequence Learning

Quantum-inspired fast weight programmer combining KAN and quantum circuits for scalable sequence learning on NISQ devices.

Ax Abhijit Sen, Bikram Keshari Parida, Giridas Maiti, Mahima Arya, Denys I. Bondar 5/11/2026

Geometric Kolmogorov--Arnold Network (GeoKAN)

Geometric Kolmogorov-Arnold Networks using learned Riemannian metrics for geometry-aware approximation in input space warping.

Ax Zixi Li, Youzhen Li 5/11/2026

A Closed-Form Upper Bound for Admissible Learning-Rate Steps in Belief-Space Dynamics

Closed-form upper bound derivation for admissible learning-rate steps in belief-space dynamics using KL/Bregman geometry.

Ax Ismam Nur Swapnil, Aranya Saha, Tanvir Ahmed Khan, Mohammad Ariful Haque, Ser-Nam Lim 5/11/2026

Gradient Extrapolation-Based Policy Optimization

Gradient Extrapolation-Based Policy Optimization for efficient LLM reasoning improvement, reducing computational cost versus multi-step lookahead.

Ax Mohsen Dehghankar, Abolfazl Asudeh 5/11/2026

Sparse Attention as a Range Searching Problem: Towards an Inference-Efficient Index for KV Cache

Sparse attention indexing method for efficient LLM inference that guarantees zero false negatives in KV cache selection during decoding.