Heyang Yu
Heyang Yu
Heyang Yu
College of AI
Tsinghua University
About Me

I’m a Ph.D. student in the College of AI at Tsinghua University, advised by Prof. Yiming Li and Prof. Yuan Yao.

My research focuses on embodied AI and AI agents, with an emphasis on reinforcement learning–based post-training for robotics.

I previously interned at the THUNLP Lab, New York University (NYU), and Ant Group.

Research Interests
Embodied AI

Studying the integration of perception, action, and cognition in intelligent agents situated in the physical world.

Physical AI Agent

Exploring the development and application of intelligent agents in physical world, especially for robotics.

Reinforcement Learning

Investigating algorithms and techniques for training agents to make sequential decisions in complex environments.

Education
  • Tsinghua University
    Department of Computer Science and Technology
    B.E.
    2022 - 2026
  • Tsinghua University
    College of AI
    Ph.D.
    2026 - present
Publications (view all )
Thinking in 360°: Humanoid Visual Search in the Wild
Thinking in 360°: Humanoid Visual Search in the Wild

Heyang Yu, Y. Han, X. Zhang, B. Yin, B. Chang, X. Han, X. Liu, J. Zhang, M. Pavone, et al.

CVPR 2026

[Paper]

Thinking in 360°: Humanoid Visual Search in the Wild

Heyang Yu, Y. Han, X. Zhang, B. Yin, B. Chang, X. Han, X. Liu, J. Zhang, M. Pavone, et al.

CVPR 2026

[Paper]

MindBlock: Probing Spatial Assembly and Structure in Unified Multimodal Models
MindBlock: Probing Spatial Assembly and Structure in Unified Multimodal Models

B. Yin, J. Liu, H. Yin, Heyang Yu, T. Zhang, Z. Li, C. Li, J. Yang, M. Li, C. Feng, Y. Li

ECCV 2026

[Scholar]

MindBlock: Probing Spatial Assembly and Structure in Unified Multimodal Models

B. Yin, J. Liu, H. Yin, Heyang Yu, T. Zhang, Z. Li, C. Li, J. Yang, M. Li, C. Feng, Y. Li

ECCV 2026

[Scholar]

Enhancing Open-Domain Task-Solving Capability of LLMs via Autonomous Tool Integration from GitHub
Enhancing Open-Domain Task-Solving Capability of LLMs via Autonomous Tool Integration from GitHub

B. Lyu, X. Cong, Heyang Yu, P. Yang, C. Qian, Z. Wang, Y. Qin, Y. Ye, Y. Lu, C. Qian, et al.

ACL 2025

[Scholar]

Enhancing Open-Domain Task-Solving Capability of LLMs via Autonomous Tool Integration from GitHub

B. Lyu, X. Cong, Heyang Yu, P. Yang, C. Qian, Z. Wang, Y. Qin, Y. Ye, Y. Lu, C. Qian, et al.

ACL 2025

[Scholar]

AgentVerse: Facilitating Multi-Agent Collaboration and Exploring Emergent Behaviors
AgentVerse: Facilitating Multi-Agent Collaboration and Exploring Emergent Behaviors

W. Chen, Y. Su, J. Zuo, C. Yang, C. Yuan, C. M. Chan, Heyang Yu, Y. Lu, Y. H. Hung, et al.

ICLR 2024

[Paper]

AgentVerse: Facilitating Multi-Agent Collaboration and Exploring Emergent Behaviors

W. Chen, Y. Su, J. Zuo, C. Yang, C. Yuan, C. M. Chan, Heyang Yu, Y. Lu, Y. H. Hung, et al.

ICLR 2024

[Paper]

All publications