Dechen Gao

🎓 CS PhD Candidate | 🏫 UC Davis | 📨 dechengao at gmail dot com

Welcome! I am Dechen Gao, a PhD candidate in Computer Science at the University of California, Davis. I received my Master of Science in Computer Science from UC Davis in 2023, and later transferred to the CS PhD program with the support of my advisors Dr. Junshan Zhang and Dr. Iman Soltani. I am grateful to work with my advisors, and collaborate with all my labmates and DARE and LARA.

My research focuses on robot learning, particularly in reinforcement learning, imitation learning, world models, and foundation models. I am always open to discussions adn collaborations on these topics - feel free to drop me an email if you are interested!

I interned at Meta Reality Labs Research (2024), Meta Instagram(2022), TikTok (2024), and Amazon (2025).

Dechen Gao

News

Jul 01, 2025 We present VITA (accepted to ICLR 2026), vision-to-action flow matching policy, which evolves latent images, instead of random Gaussian, to actions. VITA is the first MLP-only policy that handles complex real-world bimanual tasks such as ALOHA.
May 01, 2025 Imitation or reinforcement policy fine-tuning? You may need both! Check out IN-RIL which interleaves between two objectives for better sample-efficiency and stability.
Sep 01, 2024 Are robots able to actively locate perspectives to solve tasks just like human do? Check out our ICRA work AV-ALOHA that applies active vision for bimanual manipulation!
May 01, 2024 We open-sourced CarDreamer , the first world model based learning platform for autonomous driving!

Experience

  • 2026.06 - 2026.09 Redmond, WA
    Meta - Research Scientist Intern
    Robot policy learning. Dexterous manipulation.
    Logo for Meta
  • 2025.06 - 2025.09 Seattle, WA
    Amazon - Applied Scientist Intern
    Logo for Amazon
  • 2024.06 - 2024.09 San Jose, CA
    TikTok - Machine Learning Engineering Intern
    Logo for TikTok
  • 2022.06 - 2022.09 Menlo Park, CA
    Meta - Software Engineering Intern
    Logo for Meta

Education

  • 2023.03 - Present Davis, California, US
    University of California, Davis - Ph.D. in Computer Science
    • Foundation models
    • World models
    • Reinforcement learning
    • Robotics
    Logo for University of California, Davis
  • 2021.09 - 2023.06 Davis, California, US
    University of California, Davis - Master of Computer Science
    Logo for University of California, Davis
  • 2017.09 - 2021.06 Shandong, China
    Shandong University - Bachelor of Computer Science
    Logo for Shandong University

Selected Publications

  1. ICLR 2026

    VITA: Vision-to-Action Flow Matching Policy

    Dechen Gao, Boqi Zhao, Andrew Lee, Ian Chuang, Hanchu Zhou, Hang Wang, Zhe Zhao, Junshan Zhang, Iman Soltani

    Project Paper Code
    BibTeX 
    @misc{gao2025vitavisiontoactionflowmatching,
  title = {VITA: Vision-to-Action Flow Matching Policy},
  author = {Gao, Dechen and Zhao, Boqi and Lee, Andrew and Chuang, Ian and Zhou, Hanchu and Wang, Hang and Zhao, Zhe and Zhang, Junshan and Soltani, Iman},
  year = {2025},
  eprint = {2507.13231},
  archiveprefix = {arXiv},
  primaryclass = {cs.CV},
}
  2. Preprint 2026

    DuoMind: Enabling Distributed Multi-Robot Coordination with Semantic Communication

    Hanchu Zhou, Dechen Gao, Hang Wang, Brendan Lynch, Boqi Zhao, Qiyao Ma, Raman Goyal, Junshan Zhang

    Project Paper Code
    BibTeX 
    @misc{zhou2026duomind,
  title = {DuoMind: Enabling Distributed Multi-Robot Coordination with Semantic Communication},
  author = {Zhou, Hanchu and Gao, Dechen and Wang, Hang and Lynch, Brendan and Zhao, Boqi and Ma, Qiyao and Goyal, Raman and Zhang, Junshan},
  year = {2026},
}
  4. Preprint 2025

    IN-RIL: Interleaved Reinforcement and Imitation Learning for Policy Fine-Tuning

    Dechen Gao, Hang Wang, Hanchu Zhou, Nejib Ammar, Shatadal Mishra, Ahmadreza Moradipari, Iman Soltani, Junshan Zhang

    Paper
    BibTeX 
    @article{gao2025ril,
  title = {IN-RIL: Interleaved Reinforcement and Imitation Learning for Policy Fine-Tuning},
  author = {Gao, Dechen and Wang, Hang and Zhou, Hanchu and Ammar, Nejib and Mishra, Shatadal and Moradipari, Ahmadreza and Soltani, Iman and Zhang, Junshan},
  journal = {arXiv preprint arXiv:2505.10442},
  year = {2025},
}
  7. IEEE IoTJ 2024

    Cardreamer: Open-source learning platform for world model based autonomous driving

    Dechen Gao, Shuangyu Cai, Hanchu Zhou, Hang Wang, Iman Soltani, Junshan Zhang

    Project Paper
    BibTeX 
    @article{gao2024cardreamer,
  title = {Cardreamer: Open-source learning platform for world model based autonomous driving},
  author = {Gao, Dechen and Cai, Shuangyu and Zhou, Hanchu and Wang, Hang and Soltani, Iman and Zhang, Junshan},
  journal = {IEEE Internet of Things Journal},
  year = {2024},
  publisher = {IEEE},
}
  9. World-Model based Hierarchical Planning with Semantic Communications for Autonomous Driving — research preview
    Preprint 2024

    World-Model based Hierarchical Planning with Semantic Communications for Autonomous Driving

    Dechen Gao, Hang Wang, Shuangyu Cai, Hanchu Zhou, Nejib Ammar, Shatadal Mishra, Iman Soltani, Junshan Zhang

  11. COLM 2026

    Personalized RewardBench: Evaluating Reward Models with Human-Aligned Personalization

    Qiyao Ma, Dechen Gao, Rui Cai, Boqi Zhao, Hanchu Zhou, Junshan Zhang, Zhe Zhao

    Paper Code Dataset
    BibTeX 
    @inproceedings{ma2026personalizedrewardbench,
  title = {Personalized RewardBench: Evaluating Reward Models with Human-Aligned Personalization},
  author = {Ma, Qiyao and Gao, Dechen and Cai, Rui and Zhao, Boqi and Zhou, Hanchu and Zhang, Junshan and Zhao, Zhe},
  booktitle = {Conference on Language Modeling (COLM)},
  year = {2026},
}