Pushi Zhang (张蒲石)

Researcher at Microsoft Research

Google Scholar | Email: pushizhang@microsoft.com

Short Bio

I’m now a researcher at Machine Learning Group, Microsoft Research (Asia) under Li Zhao’s group. My current research interests include reinforcement learning and foundation models for embodied AI. I was also responsible for the project of Vision-based Game-Testing AI, in collaboration with Microsoft XBox team.

News

• 2025-08: Our project villa-X: Enhancing Latent Action Modeling in Vision-Language-Action Models is released.
• 2024-10: Our project IGOR: Image-GOal Representations are the Atomic Control Units for Foundation Models in Embodied AI is released.
• 2023-11: Our paper Distributional Pareto-Optimal Multi-Objective Reinforcement Learning is accepted by NeurIPS 2023. Our paper IG-Net: Image-Goal Network for Offline Visual Navigation on A Large-Scale Game Map is accepted by Robot Learning Worshop at NeurIPS 2023.

Publications

1. villa-X: Enhancing Latent Action Modeling in Vision-Language-Action Models
   Arxiv 2025 [pdf] [code] [website]
2. PIG-Nav: Key Insights for Pretrained Image Goal Navigation Models
   Arxiv 2025 [pdf] [demo] [code (released soon)]
3. What Do Latent Action Models Actually Learn?
   Arxiv 2025 [pdf]
4. IG-Net: Image-Goal Network for Offline Visual Navigation on A Large-Scale Game Map
   Baiting Zhu*, Pushi Zhang*, Xin-Qiang Cai*, Li Zhao, Masashi Sugiyama, Jiang Bian.
   NeurIPS 2023 Robot Learning Workshop [pdf] [demo]
5. Distributional Pareto-Optimal Multi-Objective Reinforcement Learning
   Xin-Qiang Cai*, Pushi Zhang*, Li Zhao, Jiang Bian, Masashi Sugiyama, Ashley J. Llorens.
   NeurIPS 2023 [code] [pdf]
6. Asking Before Action: Gather Information in Embodied Decision Making with Language Models
   Xiaoyu Chen, Shenao Zhang, Pushi Zhang, Li Zhao, Jianyu Chen.
   Arxiv 2023 [pdf]
7. An Adaptive Deep RL Method for Non-Stationary Environments with Piecewise Stable Context
   Xiaoyu Chen, Xiangming Zhu, Yufeng Zheng, Pushi Zhang, Li Zhao, Wenxue Cheng, Peng Cheng, Yongqiang Xiong, Tao Qin, Jianyu Chen, Tie-Yan Liu.
   NeurIPS 2022 [pdf]
8. Distributional Reinforcement Learning for Multi-Dimensional Reward Functions
   Pushi Zhang, Xiaoyu Chen, Li Zhao, Wei Xiong, Tao Qin, Tie-Yan Liu.
   NeurIPS 2021 [code] [pdf]
9. Independence-aware Advantage Estimation
   Pushi Zhang, Li Zhao, Guoqing Liu, Jiang Bian, Minlie Huang, Tao Qin, Tie-Yan Liu.
   IJCAI 2021 [pdf]
10. Demonstration Actor Critic
    Guoqing Liu, Li Zhao, Pushi Zhang, Jiang Bian, Tao Qin, Nenghai Yu, Tie-Yan Liu.
    Neurocomputing 2021 [pdf]

About Myself

I hope to make a positive impact by conducting research that contributes meaningfully to society. My goal is to create knowledge and solutions that benefit the world and inspire others along the way.

Last update: 2025.8.13