Yichen Zhu 祝毅晨

yichen_zhu (at) foxmail.com


Short Bio

I left Midea Group in October 2025. Prior to that, I served as Head of the Embodied AI Department at Midea Group. I received my B.S. from University of Toronto. My goal is to build generalist robots. My research interests lie at the intersection of robot learning, vision-language model, and model compression. I believe in end-to-end model. I’m always up for exploring potential collaborations and/or engaging in insightful research conversations, feel free to get in touch!

News

  • [Oct. 2025] We are now focuing on in-the-wild, robotless data collection. We release ActiveUMI and HumaniodExo.
  • [Aug. 2025] DexVLA is accepted to CoRL 2025. ChatVLA is accepted to EMNLP 2025.
  • [Sep. 2025] Two papers, including ChatVLA-2, are accepted to NeurIPS.
  • [Aug. 2025] DexVLA is accepted to CoRL 2025. ChatVLA is accepted to EMNLP as Oral Presentation.
  • [June. 2025] CoA-VLA is accepted to ICCV 2025.
  • [May. 2025] We propose the first VLA model that preserves the reasoning capabilities of VLMs, enabling robots to perform open-world manipulation tasks without requiring any reasoning-specific training data. ChatVLA-2 is available here!
  • [May. 2025] We present WorldEval, a world model to evaluation real-world robot policy entirely online!
  • [May. 2025] DiffusionVLA is accepted to ICML 2025! Our proposed reasoning prediction technique is used by π0.5, who cited our latest DexVLA paper.
  • [Feb. 2025] We release DexVLA, a new framework that can complete laundry folding using only 15 hours of data.
  • [Feb. 2025] TinyVLA is accepted to IEEE RAL.
  • [Jan. 2025] Two of our papers, TinyVLA & ScaleDP, have been cited by Physical Intelligence (π0 & FAST).
  • [Jan. 2025] Our work, LLava-Phi , have been cited by Deepseek (Janus-Series, including Janus, Janus-Flow, and Janus-Pro).
  • [Jan. 2025] Two papers on diffusion policy (ScaleDP & Discrete Policy) are accepted to ICRA 2025.
  • [Dec. 2024] We introduce a new CoT-like method Chain-of-Affordance, check out our project page chain-of-affordance.github.io.
  • [Dec. 2024] Our paper Mipha, exploring the training strategy of multimodal small language model, is accepted to AAAI 2025.
  • [Sep. 2024] Three papers are accepted to NeurIPS 2024.

    • Highlight

  • Loganomaly is the 10th most-cited paper in IJCAI (2019–2024).
  • MM-SafetyBench is the 15th most-cited paper in ECCV 2024.
  • Mipha (LLava-Phi) is the 3rd most-cited paper in AAAI 2025.
  • TinyVLA is the 1st most-cited paper in IEEE RAL 2025.

    • Publications [Google Scholar][ResearchGate]

    1. Preprint
      HumanoidExo - Scalable Whole-Body Humanoid Manipulation via Wearable Exoskeleton
      Rui Zhong, Yizhe Sun, Junjie Wen, Jinming Li, Chuang Cheng, Wei Dai, Zhiwen Zeng, Huimin Lu†, Yichen Zhu* (co-first author & corresponding author), Yi Xu
      Preprint, 2025
      PDF Project Page

    2. Preprint
      ActiveUMI - Robotic Manipulation with Active Perception from Robot‑Free Human Demonstrations
      Qiyuan Zeng, Chengmeng Li, Jude St. John, Zhongyi Zhou, Junjie Wen, Guorui Feng1, Yichen Zhu* (co-first author & corresponding author), Yi Xu
      Preprint, 2025
      PDF Project Page

    3. NeurIPS
      ChatVLA-2 - Vision-Language-Action Model with Open-World Embodied Reasoning from Pretrained Knowledge
      Zhongyi Zhou, Yichen Zhu* (co-first author & corresponding author), Junjie Wen, Chaomin Shen, Yi Xu
      Conference on Neural Information Processing Systems (NeurIPS), 2025
      PDF Project Page

    4. Generated Video
      WorldEval - World Model as Real-World Robot Policies Evaluator
      Yaxuan Li, Yichen Zhu* (co-first author & corresponding author), Junjie Wen, Chaomin Shen, Yi Xu
      Preprint, 2025
      PDF Project Page

    5. Preprint
      PointVLA - Injecting the 3D World into Vision-Language-Action Models
      Chengmeng Li, Yichen Zhu* (co-first author & corresponding author), Yaxin Peng, Yan Peng, Feifei Feng
      Preprint, 2025
      PDF Project Page

    6. EMNLP Oral
      ChatVLA - Unified Multimodal Understanding and Robot Control with Vision-Language-Action Model
      Zhongyi Zhou, Yichen Zhu* (co-first author & corresponding author), Minjie Zhu, Junjie Wen, Ning Liu, Zhiyuan Xu, Weibin Meng, Ran Cheng, Yaxin Peng, Chaomin Shen, Feifei Feng
      Conference on Empirical Methods in Natural Language Processing (EMNLP), Oral, 2025
      PDF Project Page

    7. Preprint
      ObjectVLA - End-to-End Open-World Object Manipulation without Demonstration
      Minjie Zhu, Yichen Zhu* (co-first author & corresponding author), Jinming Li, Zhongyi Zhou, Junjie Wen, Xiaoyu Liu, Chaomin Shen, Yaxin Peng, Feifei Feng
      Preprint, 2025
      PDF Project Page

    8. Preprint
      DexVLA - Vision-Language Model with Plug-In Diffusion Expert for Visuomotor Policy Learning
      Junjie Wen*, Yichen Zhu* (co-first author & corresponding author), Jinming Li, Yaxin Peng, Chaomin Shen, Feifei Feng
      Conference on Robot Learning (CoRL), 2025
      PDF Project Page

    9. ICCV
      CoA-VLA - Improving Vision-Language-Action Models via Chain-of-Affordance
      Jinming Li*, Yichen Zhu* (co-first author & corresponding author), Zhibin Tang, Junjie Wen, Minjie Zhu, Xiaoyu Liu, Chengmeng Li, Ran Cheng, Yaxin Peng, Feifei Feng
      Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2025
      PDF Project Page

    10. ICML
      Diffusion-VLA - Scaling Robot Foundation Models via Unified Diffusion and Autoregression
      Junjie Wen*, Yichen Zhu* (co-first author & corresponding author), Minjie Zhu*, Zhibin Tang, Jinming Li, Zhongyi Zhou, Chengmeng Li, Xiaoyu Liu, Yaxin Peng, Chaomin Shen, Feifei Feng
      International Conference on Machine Learning (ICML), 2025
      PDF Project Page

    11. IEEE RAL
      TinyVLA - Towards Fast, Data-Efficient Vision-Language-Action Models for Robotic Manipulation
      Junjie Wen*, Yichen Zhu* (co-first author & corresponding author), Jinming Li, Minjie Zhu, Kun Wu, Zhiyuan Xu, Ran Cheng, Chaomin Shen, Yaxin Peng, Feifei Feng, Jian Tang
      IEEE Robotics and Automation Letters, 2025
      PDF Project Page

    12. ICRA
      Scaling Diffusion Policy in Transformer to 1 Billion Parameters for Robotic Manipulation
      Minjie Zhu*, Yichen Zhu* (co-first author & corresponding author), Jinming Li, Junjie Wen, Zhiyuan Xu, Ning Liu, Ran Cheng, Chaomin Shen, Yaxin Peng, Feifei Feng, Jian Tang
      IEEE International Conference on Robotics and Automation (ICRA), 2025
      PDF Project Page

    13. ICRA
      Discrete Policy - Learning Disentangled Action Space for Multi-Task Robotic Manipulation
      Kun Wu*, Yichen Zhu* (co-first author), Jinming Li, Junjie Wen, Ning Liu, Zhiyuan Xu, Qirui Qiu, Jian Tang
      IEEE International Conference on Robotics and Automation (ICRA), 2025
      PDF

    14. AAAI
      Mipha - A Comprehensive Overhaul of Multimodal Assistant with Small Language Models
      Minjie Zhu*, Yichen Zhu* (co-first author & corresponding author), Ning Liu, Xin Liu, Zhiyuan Xu, Chaomin Shen, Yaxin Peng
      Proceedings of the AAAI Conference on Artificial Intelligence (AAAI), 2025
      PDF

    15. NeurIPS
      Any2Policy - Learning Visuomotor Policy with Any-Modality
      Yichen Zhu, Zhicai Ou, Feifei Feng, Jian Tang
      Conference on Neural Information Processing Systems (NeurIPS), 2024
      PDF

    16. NeurIPS
      An Efficient Diffusion Transformer Framework Inspired by Human-like Sketching
      Xinwang Chen*, Ning Liu*, Yichen Zhu, Feifei Feng, Jian Tang
      Conference on Neural Information Processing Systems (NeurIPS), 2024

    17. NeurIPS
      A General Retrieval-Augmented Graph Learning Framework
      Xinke Jiang, Rihong Qiu, Yongxin Xu, Wentao Zhang, Yichen Zhu, Ruizhe Zhang, Yuchen Fang, Xu Chu, Junfeng Zhao, Yasha Wang
      Conference on Neural Information Processing Systems (NeurIPS), 2024

    18. ECCV
      Mm-safetybench - A Benchmark for Safety Evaluation of Multimodal Large Language Models
      Xin Liu*, Yichen Zhu* (co-first author), Jindong Gu, Yunshi Lan, Chao Yang, Yu Qiao
      European Conference on Computer Vision (ECCV), 2024
      PDF Project Page BibTex

    19. CVPR
      Retrieval-Augmented Embodied Agents
      Yichen Zhu, Zhicai Ou, Feifei Feng, Jian Tang
      IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2024
      PDF BibTex

    20. ICRA
      Language-Conditioned Robotic Manipulation with Fast and Slow Thinking
      Minjie Zhu*, Yichen Zhu* (co-first author), Jinming Li, Junjie Wen, Zhiyuan Xu, Zhengping Che, Chaomin Shen, Yaxin Peng, Dong Liu, Feifei Feng, Jian Tang
      IEEE International Conference on Robotics and Automation (ICRA), 2024
      PDF Project Page BibTex

    21. ICRA
      Object-Centric Instruction Augmentation for Robotic Manipulation
      Junjie Wen*, Yichen Zhu* (co-first author), Minjie Zhu, Jinming Li, Zhiyuan Xu, Zhengping Che, Chaomin Shen, Yaxin Peng, Dong Liu, Feifei Feng, Jian Tang
      IEEE International Conference on Robotics and Automation (ICRA), 2024
      PDF Project Page BibTex

    22. ACM MM@EMCLR
      Llava-phi - Efficient Multi-modal Assistant with Small Language Model
      Yichen Zhu, Minjie Zhu, Ning Liu, Zhicai Ou, Feifei Feng, Jian Tang
      ACM Multimedia - Efficient Multimedia Computing under Limited Resources, 2024
      PDF Project Page BibTex

    23. IJCAI
      Safety of Multimodal Large Language Models on Images and Texts
      Xin Liu, Yichen Zhu, Jindong Gu, Yunshi Lan, Chao Yang, Yu Qiao
      International Joint Conference on Artificial Intelligence (IJCAI), 2024
      PDF

    24. AAAI
      Exploring Gradient Explosion in Generative Adversarial Imitation Learning - A Probabilistic Perspective
      Wanying Wang*, Yichen Zhu* (co-first author), Yirui Zhou, Chaomin Shen, Zhiyuan Xu, Jian Tang, Yaxin Peng
      Proceedings of the AAAI Conference on Artificial Intelligence (AAAI), Oral, 2024
      PDF BibTex

    25. AAAI
      Early Pruning with Self-Distillation for Effcient Model Compression
      Dong Chen, Ning Liu, Yichen Zhu, Zhengping Che, Rui Ma, Fachao Zhang, Xiaofeng Mou, Yi Chang, Jian Tang
      Proceedings of the AAAI Conference on Artificial Intelligence (AAAI), 2024
      PDF

    26. CVPR
      ScaleKD - Distilling Scale-aware Knowledge in Small Object Detector
      Yichen Zhu, Qiqi Zhou, Zhicai Ou, Feifei Feng, Jian Tang
      IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2023
      PDF Project Page BibTex

    27. ECCV
      Label-Guided Auxiliary Training Improves 3D Object Detector
      Yaomin Huang*, Xinmei Liu*, Yichen Zhu, Zhiyuan Xu, Chaomin Shen, Zhengping Che, Guixu Zhang, Yaxin Peng, Feifei Feng, Jian Tang
      European Conference on Computer Vision (ECCV), 2022
      PDF

    28. NeurIPS
      Teach Less, Learn More - On the Undistillable Classes in Knowledge Distillation
      Yichen Zhu, Ning Liu, Zhiyuan Xu, Xin Liu, Weibin Meng, Louis Wang, Zhicai Ou, Jian Tang
      Advances in Neural Information Processing Systems (NeurIPS), Spotlight, 2022
      PDF BibTex

    29. ICCV
      Student Customized Knowledge Distillation - Bridging the Gap Between Student and Teacher
      Yichen Zhu, Yi Wang
      Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), Oral, 2021
      PDF BibTex

      30. Services

      • I was a reviewer/PC committee for CVPR 2020-2024, ECCV 2022-2024, ICCV 2021-2025, NeurIPS 2022-2025, ICLR 2022-2024, AAAI 2022-2025, CoRL 2024/2025, ICRA 2024/2025, ACL 2024/2025, EMNLP 2024, NAACL 2024, AISTATS 2024, BMVC 2024, ECML/PKDD 2024, ACCV 2024
      • I was a reviewer of IEEE TPAMI, TNNLS, TMLR, RAL, Neural Networks, Machine Learning, Information Fusion
      • I was a Session Chair of ECML/PKDD 2024