Sunrisulfr的菠萝屋
0%

About Me

Hello!

I am Mingyang Sun (孙铭洋).

I am a Ph.D. candidate in Artificial Intelligence at the Shanghai Innovation Institute (SII), under the supervision of Prof. Weinan Zhang and Prof. Cewu Lu. Concurrently, I am a joint Ph.D. student at Zhejiang University – Westlake University, under the supervision of Prof. Donglin Wang at the Machine Intelligence Laboratory (MiLAB). I previously earned both my M.S. and B.S. degrees in Computer Science and Technology from Dalian University of Technology.

Research: My current research lies at the intersection of Embodied AI and Intelligent Agent, with a focus on building embodied systems that achieve long-term autonomy in open-ended environments. I investigate how perception, action, memory, and decision-making can be integrated into a unified framework to support a consistent sense of self in embodied agents. Reinforcement Learning (RL) has been a core methodological foundation of my work since my master’s studies. Building upon RL, I explore mechanisms such as long-horizon learning and embodied interaction to advance scalable and adaptive robotic intelligence.

Links: GitHub Google Scholar
E-mail: sunmingyang@zju.edu.cn
Internship: Huawei Noah's Ark Lab Ant Group - Robby
Reviewer: TPAMI CVPR ICML ICLR

If you're interested in learning more or discussing potential collaborations, please don't hesitate to interact with me.

Profile Photo

Research Interests:

Embodied Agents

  • Pillar 1: Digital Gene

    • Analytic Concepts for Manipulation (2025)
    • Analytic Concepts for VLA (2025)
    • Analytic Concepts for Reward (2025)

  • Pillar 2: Persistent Agent

    • System 3 Layer for Artificial Life (2025)
    • Memory Disign for Conginitive Architecture(2025)

Reinforcement Learning

  • Pillar 1: Generative Policy Optimization

    • Online RL for Diffusion/FLow Policy (2025)
    • Motion Generation through RL(2022)

  • Pillar 2:Multi-Agent Reinforcement Learning

    • Evolutionary Multi-Agent RL (2021-2023)
    • Multi-Agent RL Exploration (2022)

Research

Publications

Selected Papers

  1. Mingyang Sun, et al. “Score-Based Diffusion Policy Compatible with Reinforcement Learning via Optimal Transport.” Proceedings of the International Conference on Machine Learning (ICML 2025). CCF-A

  2. Mingyang Sun, et al. “Co-speech Gesture Synthesis by Reinforcement Learning with Contrastive Pre-trained Rewards.” Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2023). CCF-A

  3. Yaqing Hou†, Mingyang Sun†, et al. “A Multi-agent Cooperative Learning System with Evolution of Social Roles” IEEE Transactions on Evolutionary Computation (导师共一) 中科院1区， IF：14.3

  4. Mingyang Sun, et al. “Improving Cooperative Multi-Agent Exploration via Surprise Minimization and Social Influence Maximization” the 22nd International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2023) CCF-B

  5. Yaqing Hou, Mingyang Sun, et al. “Behavior Reasoning for Opponent Agents in Multi-agent Learning Systems,” in IEEE Transactions on Emerging Topics in Computational Intelligence 中科院2区

  6. Mingyang Sun, et al. “Memetic Multi-agent optimization with Problem Reformulation by Coordinate Rotation,” 2020 IEEE Congress on Evolutionary Computation (CEC)

Preprints & Under Submission

  1. “Multiform Evolution for High-Dimensional Problems with Low Effective Dimensionality.” arXiv:2401.00168 (1st Author)

  2. “Sophia: A Persistent Agent Framework of Artificial Life.” arXiv:2512.18202 (1st Author)

  3. “Explicit Spatial Guidance from Object-Centric Concept Analysis for Vision-Language-Action Model Fine-Tuning.” Under Review for CVPR 2026. (1st Author)

  4. “Integrating Analytical Physics with LLM-Guided Reward Evolution in Reinforcement Learning.” Under Review for IEEE TEVC. (1st Author)

  5. “Executable Analytic Concepts as the Missing Link Between VLM Insight and Precise Manipulation.” Under Review for IJCAI 2026. (1st Author)

  6. “Iterative Refinement of Flow Policies in Probability Space for Online Reinforcement Learning.” Under Review for ICML 2026. (1st Author)

  7. “Reward-Aligned Imagination with Analytic Concepts in World Models.” Under Review for ICML 2026. (Co-first Author)