Jayjun Lee

Welcome! I'm an MS Robotics student at the University of Michigan, where I'm advised by Professor Nima Fazeli. My research focuses on Robot Learning – developing algorithms for robots to perceive and interact with the physical world. These days I'm interested in multi-modal (i.e. vision, tactile, F/T, language, audio) perception and learning representations to acquire contact- and force-rich manipulation skills.


Formal Bio     Github     Google Scholar     Twitter

Email: jayjun [at] umich [dot] edu

News

Jan. 2025

RACER is accepted to ICRA 2025 in Atlanta!

Jan. 2025

COMFORT is accepted to ICLR 2025 in Singapore!

Oct. 2024

RACER won the Best Overall Award at UM AI Symposium 2024!

Sep. 2024

NISP is accepted to CoRL 2024 in Munich, Germany!

Some recent highlights from our research:
* indicates equal contribution

RACER: Rich Language-guided Failure Recovery Policies for Imitation Learning

*, *, , and
★ Best Overall Award @ UM AI Symposium 2024 ★
The 3rd Workshop on Language and Robot Learning (LangRob) @ CoRL 2024
International Conference on Robotics and Automation (ICRA) 2025
Webpage  •   Paper  •   Code

Do Vision-Language Models Represent Space and How? Evaluating Spatial Frame of Reference Under Ambiguities

*, *, *, , , , and
The 1st Pluralistic Alignment Workshop @ NeurIPS 2024
International Conference on Learning Representations (ICLR) 2025 (Oral, 1.8% acceptance rate)
Webpage  •   Paper  •   Code

Visual-auditory Extrinsic Contact Estimation

*, *, and
Under review
Webpage  •   Paper

Neural Inverse Source Problems

, , , and
Conference on Robot Learning (CoRL) 2024
Webpage  •   Paper  •   Code

Naturalistic Robot Arm Trajectory Generation via Representation Learning

and Adam Spiers
Poster @ UK-RAS Towards Autonomous Robotic Systems (TAROS) 2023
PDF