Amber Li

Hi! I am a third year Ph.D. student in the Robotics Institute at Carnegie Mellon University. I am fortunate to be advised by Professors David Held and Maxim Likhachev. I have also been a research intern at Basis Research Institute.

Previously I received my B.S. and M.Eng. in computer science from MIT, where I was a part of the Learning and Intelligent Systems Group and worked with Tom Silver, advised by Leslie Kaelbling.

I have also spent some time as a software engineer at Two Sigma. In my free time, I enjoy running, reading, and playing the violin. I have also dabbled in dance.

Research

I'm broadly interested in creating intelligent embodied agents that can learn and act in dynamic environments over long time horizons. Recently, I've been investigating how we can efficiently plan and adapt robot policies and/or world models based on observed outcomes.

Adapting World Models from VLM Plans for Robotic Tool Use
Amber Li, Sam Ethan Mathew, Maxim Likhachev, David Held
In submission to the International Conference on Robotics and Automation (ICRA) 2027
project page

Long-Horizon Multi-Object Rearrangement via Distilled Tree Search over Continuous Actions
Kallol Saha, Yining Hong, Amber Li, Yejin Choi, Maxim Likhachev, David Held
Conference on Robot Learning (CoRL) 2026 (Oral presentation, Top 1%)
project page

EMPIRIC: Experiment-Driven Learning of Residual World Models for Robot Planning
Yichao Liang, Amber Li, Dat Nguyen*, Emily Bunnapradist*, Michelangelo Naim*, Sreela Kodali*, Matteo Merler, Bowen Li, Kiran Gopinathan, Yiyun Liu, Nikhil Pimpalkhare, Joshua B. Tenenbaum, Adrian Weller, Zenna Tavares, Tom Silver†, Kevin Ellis†
Learning Effective Abstractions for Planning (LEAP) Workshop, Conference on Robot Learning (CoRL) 2026
project page / arXiv / code / workshop

Scouting the Dynamics Gap: Test-Time Policy Adaptation via Action-Outcome Feedback
Yishu Li*, Liyuan Geng*, Xinyi Mao*, Amber Li, David Held
Conference on Robot Learning (CoRL) 2026
project page / arXiv

Embodied Active Learning of Relational State Abstractions for Bilevel Planning
Amber Li*, Tom Silver
Conference on Lifelong Learning Agents (CoLLAs) 2023 (Oral presentation, Top 12)
presentation / arXiv / code

Contact

Email / Resume / Scholar / Twitter / Github

Thanks to Brandon Wang for reminding me about my old website design. I lost the original code, but luckily coding agents can help me recreate it.