Jason Chen

I am a third year undergraduate student at the University of Southern California working with Prof. Daniel Seita and Prof. Gaurav S. Sukhatme. I'm majoring in Computer Science.

My current research focuses on a data-driven approach to embodied intelligence that generates synthetic data from previously collected datasets for deployment in offline reinforcement learning, planning, and imitation learning.

I am interested in developing data-efficient algorithms and policies for multi-modal perception that harness vision, touch, audio, and language for fine-grained and effective manipulation.

I am currently also getting an accelerated master's degree in Computer Science at USC.

Email / GitHub / Linkedin / Google Scholar

Research

* indicates equal contribution, † indicates equal advising

ROPA: Synthetic Robot Pose Generation for RGB-D Bimanual Data Augmentation

Jason Chen, I-Chun (Arthur) Liu, Gaurav S. Sukhatme, Daniel Seita
In Review, 2025
arxiv / website

ROPA is a data augmentation method for bimanual imitation learning that generates 3rd person robot poses.

D-CODA: Diffusion for Coordinated Dual-Arm Data Augmentation

I-Chun (Arthur) Liu, Jason Chen, Gaurav S. Sukhatme, Daniel Seita
Conference on Robot Learning (CoRL), 2025
arxiv / website

D-CODA is a method for offline data augmentation tailored to eye-in-hand bimanual imitation learning.

Awards

Undergraduate Research Associates Program Scholarship: University of Southern California, 2025
Center for Undergraduate Research in Viterbi Engineering: University of Southern California, 2024-2025
1st Place in Phase 1 of RCare World Challenge : 2024

Services

RSS Volunteer: 2025
High School Robotics Mentor: 2023, 2024

Inspired by the template here.