Hao Wen (温皓)

Hao Wen (温皓)

Feel free to contact me.

About me

I am Hao Wen (温皓), a Ph.D. student at Institute for AI Industry Research (AIR), Tsinghua University, expecting to graduate in December 2026. I am advised by Prof. Ya-Qin Zhang and Prof. Yunhao Liu, with Prof. Yunxin Liu and Prof. Yuanchun Li as my research mentors.

My research lies at the intersection of AI agents and systems. I build agents that operate devices through natural language, with an emphasis on reliable task execution and efficient reasoning under practical resource constraints. My work spans mobile GUI agents, exploration and environment knowledge, and budget-aware and parallel reasoning. More broadly, I am interested in how agents can learn from interaction and use feedback to improve their decisions in digital and physical environments.

Google Scholar · GitHub

Research Interests

  • Agents that learn through interaction. Acquiring useful environment knowledge from exploration, demonstrations, and failed attempts, and turning that knowledge into more reliable task execution.
  • Efficient reasoning and execution. Allocating computation across planning, verification, and recovery, with a focus on budget-aware reasoning, parallel inference, and deployment on resource-constrained devices.

Honors and Awards

Publications and Preprints

Mobile Agents and Systems

  • LLM-Explorer: Towards Efficient and Affordable LLM-based Exploration for Mobile Apps
    Shanhui Zhao, Hao Wen, Wenjie Du, Cheng Liang, Yunxin Liu, Xiaozhou Ye, Ye Ouyang, Yuanchun Li.
    ACM MobiCom 2025
    Paper · Code
    Uses LLMs to maintain exploration knowledge while generating actions without an LLM call at every step.

  • AutoDroid-V2: Boosting SLM-based GUI Agents via Code Generation
    Hao Wen, Shizuo Tian, Borislav Pavlov, Wenjie Du, Yixuan Li, Ge Chang, Shanhui Zhao, Jiacheng Liu, Yunxin Liu, Ya-Qin Zhang, Yuanchun Li.
    ACM MobiSys 2025 · Best Artifact Award
    Paper · Code
    Enables small language models to automate mobile tasks by generating executable scripts from app-specific documentation.

  • AutoDroid: LLM-powered Task Automation in Android
    Hao Wen, Yuanchun Li, Guohong Liu, Shanhui Zhao, Tao Yu, Toby Jia-Jun Li, Shiqi Jiang, Yunhao Liu, Yaqin Zhang, Yunxin Liu.
    ACM MobiCom 2024
    Paper · Code · Project
    Combines LLM reasoning with app knowledge acquired through automated exploration to complete natural-language tasks on Android.

  • AdaptiveNet: Post-deployment Neural Architecture Adaptation for Diverse Edge Environments
    Hao Wen, Yuanchun Li, Zunshuai Zhang, Shiqi Jiang, Xiaozhou Ye, Ye Ouyang, Yaqin Zhang, Yunxin Liu.
    ACM MobiCom 2023
    Paper · Code

Efficient LLM Reasoning

  • ParaThinker: Native Parallel Thinking as a New Paradigm to Scale LLM Test-time Compute
    Hao Wen, Yifan Su, Feifei Zhang, Yunxin Liu, Yunhao Liu, Ya-Qin Zhang, Yuanchun Li.
    Preprint, 2025
    Paper · Code
    Trains language models to explore multiple reasoning paths in parallel and synthesize their results.

  • BudgetThinker: Empowering Budget-aware LLM Reasoning with Control Tokens
    Hao Wen, Xinrui Wu, Yi Sun, Feifei Zhang, Liye Chen, Jie Wang, Yunxin Liu, Yunhao Liu, Ya-Qin Zhang, Yuanchun Li.
    Preprint, 2025
    Paper · Code
    Enables language models to adapt their reasoning to an explicit token budget through control tokens and budget-aware training.

Other Publications and Preprints

  • Personal LLM Agents: Insights and Survey about the Capability, Efficiency and Security
    Yuanchun Li, Hao Wen, Weijun Wang, Xiangyu Li, Yizhen Yuan, Guohong Liu, Jiacheng Liu, Wenxing Xu, Xiang Wang, Yi Sun, Rui Kong, Yile Wang, Hanfei Geng, Jian Luan, Xuefeng Jin, Zilong Ye, Guanjing Xiong, Fan Zhang, Xiang Li, Mengwei Xu, Zhijun Li, Peng Li, Yang Liu, Ya-Qin Zhang, Yunxin Liu.
    Preprint, 2024
    Paper · Resources

  • Generative Model for Models: Rapid DNN Customization for Diverse Tasks and Resource Constraints
    Wenxing Xu, Yuanchun Li, Jiacheng Liu, Yi Sun, Zhengyang Cao, Yixuan Li, Hao Wen, Yunxin Liu.
    Preprint, 2023
    Paper

  • Retrieval-based Battery Degradation Prediction for Battery Energy Storage System Operations
    Yixuan Li, Qirui Yang, Hao Wen, Huiwen Zheng, Weimin Liu, Hui Li, Yuanchun Li, Yunxin Liu.
    IEEE GreenCom 2023

  • DroidBot-GPT: GPT-powered UI Automation for Android
    Hao Wen, Hongming Wang, Jiaxuan Liu, Yuanchun Li.
    Paper · Code

Education

  • Tsinghua University, 2022–2026 (expected)
    Ph.D. student; research at the Institute for AI Industry Research (AIR). Expected graduation: December 2026.
  • Tsinghua University, 2018–2022
    B.S., Department of Automation.

Research Experience

  • Shanghai Artificial Intelligence Laboratory, Summer 2021
    Research Assistant Intern.