Min Liu

CV | Google Scholar | Github | Email | LinkedIn | Twitter

Hi! I am a master student in Machine Learning Department at Carnegie Mellon University, with interests in machine learning and robotics. I am fortunate to work with Prof. Ruslan Salakhutdinov, Prof. Deepak Pathak, and Murtaza Dalal.

Prior to CMU, I received my B.S. in Computer Science at Kuang Yaming Honors School, Nanjing University. I was advised by Prof. Alberto Sangiovanni-Vincentelli and Prof. Xiangyu Yue (now at CUHK) at UC Berkeley, Prof. Lin Shao at National University of Singapore, and Prof. Shujian Huang at Nanjing University.

Feel free to reach out if you have shared interests in robot learning. I am always open to collaboration and insightful discussions.


News
  • [Oct, 2024] I am applying for CS/ML/Robotics PhD for 2025 Fall.
  • [Oct, 2024] We released ManipGen, a generalist agent for long-horizon manipulation!
  • [Jun, 2024] SoftMAC is accepted to IROS 2024.
  • [July, 2023] Non-Adversarial Backdoor is accepted to ICCV 2023.
  • [Dec, 2022] Awarded SenseTime Scholarship (30 undergraduates in the field of AI in China).
  • [Nov, 2022] SelectiveKD is accepted to AAAI 2023.
  • [Mar, 2022] Start internship at ByteDance AI Lab.
Research

My research objective is to develop general-purpose embodied agents that can reason, plan, and act in open-world environments, much like humans. Motivated by this vision, I explore how to scale robot learning on heterogeneous data sources, refine sensorimotor policies through online self-practice, and integrate reactive control with semantic reasoning to solve complex, long-horizon tasks in the physical world.

Key areas of focus include:

  • machine learning for decision making and control
  • robot learning systems that scale with broad sources of data
  • data engines for embodied AI and robotics research
  • Publications
    (* denotes equal contribution; representative papers are highlighted)
    Local Policies Enable Zero-shot Long Horizon Manipulation
    Murtaza Dalal*, Min Liu*, Walter Talbott, Chen Chen, Deepak Pathak, Jian Zhang, Ruslan Salakhutdinov
    In submission to ICRA 2025
    CoRL Learning Effective Abstractions for Planning Workshop, 2024 (Oral)
    paper | code | website | video

    profile photo SoftMAC: Differentiable Soft Body Simulation with Forecast-based Contact Model and Two-way Coupling with Articulated Rigid Bodies and Clothes
    Min Liu, Gang Yang, Siyuan Luo, Lin Shao
    International Conference on Intelligent Robots and Systems (IROS), 2024
    paper | code | website | video

    profile photo Beating Backdoor Attack at Its Own Game
    Min Liu, Alberto Sangiovanni-Vincentelli, Xiangyu Yue
    International Conference on Computer Vision (ICCV), 2023
    paper | code | video

    profile photo Selective Knowledge Distillation for Non-Autoregressive Neural Machine Translation
    Min Liu, Yu Bao, Chengqi Zhao, Shujian Huang
    AAAI Conference on Artificial Intelligence (AAAI), 2023
    paper | code


    Work Experience
    profile photo ByteDance, AI-Lab
    Machine Learning (NLP) Intern
    Shanghai, China
    Education
    NJU logo Carnegie Mellon University
    M.S. in Machine Learning (2023 - 2024, Expected)
    NJU logo Nanjing University
    B.S. in Computer Science, Kuang Yaming Honors School (2019 - 2023)
    Overall GPA: 91.2/100, Ranking: 1/15

    Services
    • Conference reviewer: CVPR (2024, 2025), ICRA (2025)
    Miscellanea
    • My Chinese Name (surname first): 刘 闵.
    • I am from Yangzhou, Jiangsu Province in China. I've also lived in Nanjing, Shanghai, Hong Kong, and Pittsburgh.

    This homepage is designed based on Jon Barron's website and deployed on Github Pages.

    Copyright 2023 © Min Liu