Publications

Publications by categories in reversed chronological order. The asterisks(*) mean that authors are equal contribution.

Preprints

  1. du2026adaexplore.png
    AdaExplore: Failure-Driven Adaptation and Diversity-Preserving Search for Efficient Kernel Generation
    Weihua Du, Jingming Zhuo, Yixin Dong, Andre Wang He, Weiwei Sun, Zeyu Zheng, Manupa Karunaratne, Ivan Fox, Tim Dettmers, Tianqi Chen, and 2 more authors
    Preprint, 2026
    arXiv Website
  2. liu2026consensus.png
    The Consensus Trap: Rescuing Multi-Agent LLMs from Adversarial Majorities via Token-Level Collaboration
    Jiayuan Liu, Shiyi Du, Weihua Du, Mingyu Guo, and Vincent Conitzer
    Preprint, 2026
    arXiv
  3. zhou2026mindsim2realgapuser.png
    Mind the Sim2Real Gap in User Simulation for Agentic Tasks
    Xuhui Zhou*, Weiwei Sun*, Qianou Ma, Yiqing Xie, Jiarui Liu, Weihua Du, S. Welleck, Yiming Yang, Graham Neubig, Sherry Tongshuang Wu, and 1 more author
    Preprint, 2026
    arXiv
  4. yang2026gradalign.png
    GradAlign: Gradient-Aligned Data Selection for LLM Reinforcement Learning
    Ningyuan Yang*, Weihua Du*, Weiwei Sun, Sean Welleck, and Yiming Yang
    Preprint, 2026
    arXiv Website
  5. sun2025personalized.png
    Training Proactive and Personalized LLM Agents
    Weiwei Sun, Xuhui Zhou, Weihua Du, Xingyao Wang, Sean Welleck, Graham Neubig, Maarten Sap, and Yiming Yang
    Preprint, 2025
    arXiv Website

Proceedings

2025

  1. lu2025scaling.png
    Scaling LLM Multi-turn RL with End-to-end Summarization-based Context Management
    Miao Lu, Weiwei Sun, Weihua Du, Zhan Ling, Xuesong Yao, Kang Liu, and Jiecao Chen
    ACL 2026, 2025
    arXiv
  2. du2025codegym.png
    Generalizable End-to-End Tool-Use RL with Synthetic CodeGym
    Weihua Du, Hailei Gong, Zhan Ling, Kang Liu, Lingfeng Shen, Xuesong Yao, Yufei Xu, Dingyuan Shi, Yiming Yang, and Jiecao Chen
    ICLR 2026, 2025
    arXiv Website Dataset
  3. han2025zerotuning.png
    ZeroTuning: Unlocking the Initial Token’s Power to Enhance Large Language Models Without Training
    Feijiang Han, Xiaodong Yu, Jianheng Tang, Delip Rao, Weihua Du, and Lyle Ungar
    ICLR 2026, 2025
    arXiv
  4. du2025agentic.png
    Agentic-R1: Distilled Dual-Strategy Reasoning
    Weihua Du, Pranjal Aggarwal, Sean Welleck, and Yiming Yang
    EMNLP 2025, 2025
    arXiv Website Dataset
  5. du2025optimizing.png
    Optimizing Temperature for Language Models with Multi-Sample Inference
    Weihua Du, Yiming Yang, and Sean Welleck
    ICML 2025, 2025
    arXiv Website

2024

  1. du2024constrained.png
    Constrained Human-AI Cooperation: An Inclusive Embodied Social Intelligence Challenge
    Weihua Du*, Qiushi Lyu*, Jiaming Shan, Zhenting Qi, Hongxin Zhang, Sunli Chen, Andi Peng, Tianmin Shu, Kwonjoon Lee, Behzad Dariush, and 1 more author
    NeurIPS 2024 D&B Track, 2024
    arXiv Website
  2. chen2023eval.png
    T-Eval: Evaluating the Tool Utilization Capability of Large Language Models Step by Step
    Zehui Chen*, Weihua Du*, Wenwei Zhang*, Kuikun Liu, Jiangning Liu, Miao Zheng, Jingming Zhuo, Songyang Zhang, Dahua Lin, Kai Chen, and 1 more author
    ACL 2024, 2024
    arXiv Website
  3. zhou2023hazard.png
    HAZARD Challenge: Embodied Decision Making in Dynamically Changing Environments
    Qinhong Zhou*, Sunli Chen*, Yisong Wang, Haozhe Xu, Weihua Du, Hongxin Zhang, Yilun Du, Joshua B Tenenbaum, and Chuang Gan
    ICLR 2024, 2024
    arXiv Openreview Website

2023

  1. zhang2023building.png
    Building Cooperative Embodied Agents Modularly with Large Language Models
    Hongxin Zhang*, Weihua Du*, Jiaming Shan, Qinhong Zhou, Yilun Du, Joshua B Tenenbaum, Tianmin Shu, and Chuang Gan
    ICLR 2024, also NeurIPS 2023 FMDM, 2023
    arXiv Openreview Website
  2. neurips2023peerreview.png
    Iteratively Learn Diverse Strategies with State Distance Information
    Wei Fu, Weihua Du, Jingwei Li, Sunli Chen, Jingzhao Zhang, and Yi Wu
    NeurIPS 2023, 2023
    arXiv Openreview
  3. ijcai2023p407.png
    Automatic Truss Design with Reinforcement Learning
    Weihua Du*, Jinglun Zhao*, Chao Yu, Xingcheng Yao, Zimeng Song, Siyang Wu, Ruifeng Luo, Zhiyuan Liu, Xianzhong Zhao, and Yi Wu
    IJCAI 2023, 2023
    arXiv