About Me

I am a fifth-year Ph.D. student at Peking University, advised by Prof. Zongqing Lu. I received my Bachelor's degree in 2021, from the Turing Class at Peking University.

My research interest lies in Reinforcement Learning (RL) and Embodied AI. Currently, I am focusing on:

  • Scalable learning methods for dexterous hands and humanoid robots manipulation.
  • Efficient RL for open-world, embodied agents.

I am open to collaborations and discussions.

Selected Papers

DemoGrasp

DemoGrasp: Universal Dexterous Grasping from a Single Demonstration

Haoqi Yuan*, Ziye Huang*, Ye Wang, Chuan Mao, Chaoyi Xu, Zongqing Lu
arXiv 2025
Paper Project Code bibtex blog talk

SOTA method on universal sim-to-real grasping for any dexterous hand.

CrossDex

Cross-Embodiment Dexterous Grasping with Reinforcement Learning

Haoqi Yuan, Bohan Zhou, Yuhui Fu, Zongqing Lu
ICLR 2025
Paper Project Code bibtex blog talk slides poster

Train a human-hand policy that can generalize to different dexterous hands.

RLGPT

RL-GPT: Integrating Reinforcement Learning and Code-as-Policy

Shaoteng Liu, Haoqi Yuan, Minda Hu, Yanwei Li, Yukang Chen, Shu Liu, Zongqing Lu, Jiaya Jia
NeurIPS 2024 oral
paper project page bibtex poster

An LLM agent equipped with RL for continual learning in open-world environments.

MGPO

Pre-Trained Multi-Goal Transformers with Prompt Optimization for Efficient Online Adaptation

Haoqi Yuan, Yuhui Fu, Feiyang Xie, Zongqing Lu
NeurIPS 2024
paper code bibtex poster

Propose a multi-goal-conditioned Transformer policy for fast adaptation.

PTGM

Pre-Training Goal-Based Models for Sample-Efficient Reinforcement Learning

Haoqi Yuan, Zhancun Mu, Feiyang Xie, Zongqing Lu
ICLR 2024 oral (top 1.2%)
Paper Project Code bibtex talk slides poster

Leverage pre-training to bootstrap RL for open-world long-horizon tasks.

CORRO

Robust Task Representations for Offline Meta-Reinforcement Learning via Contrastive Learning

Haoqi Yuan, Zongqing Lu
ICML 2022
paper code bibtex talk slides poster

A contrastive learning method for robust task representations in offline meta-RL.

View All Papers ≫

Experience

BeingBeyond
BeingBeyond 智在无界 Research Intern (2025-)
BeingBeyond 智在无界
Research Intern (2025 - Present).
We are a start-up team on embodied AI and foundation models.
I am leading research on dexterous manipulation.
Representative works: Being-0, Being-H0, DemoGrasp.
BAAI
BAAI 北京智源 Research Intern (2023-2025)
BAAI 北京智源人工智能研究院
Research Intern (2023-2025).
I focus on RL research, including RL for open-world agents and RL for generalizable dexterous manipulation.
Representative works: PTGM, MGPO, Plan4MC, RL-GPT, CrossDex, ResDex, BiDexHD.
PKU
Hyperplane Lab (PKU) Research Intern (2019-2021)
Hyperplane Lab, Peking University
Research Intern (2019 - 2021). Advised by Prof. Hao Dong.
I study generative models and learning physical interactions.
Representative work: DMotion.

Selected Awards

  • Peking University President Scholarship | 北京大学校长奖学金 (2022, 2021)
  • National Scholarship | 国家奖学金 (2018)
  • Peking University Merit Student | 北京大学三好学生 (2025, 2018)
  • Second Class Award in Chinese Physics Olympiad (Finals) | 全国中学生物理竞赛决赛二等奖 (2016)
  • See CV for full list.

Services

Conference Reviewer

ICML'22,24,25; NeurIPS'22,23,24,25; ICLR'24,25,26; AAAI'23,24,25,26; CVPR'24,26; CORL'25; IROS'25.

Teaching Assistant

Deep Reinforcement Learning, Zongqing Lu, 2023 Spring
Computational Thinking in Social Science, Xiaoming Li, 2020 Autumn
Deep Generative Models, Hao Dong, 2020 Spring