Toan Nguyen

Toan Nguyen

CS PhD at USC

Email: tientoan [at] usc [dot] edu

I am a first-year Ph.D. student at the Department of Computer Science, University of Southern California. I am co-advised by Prof. Yue Wang and Prof. Daniel Seita in the Physical Superintelligence Lab and SLURM Lab. Prior to joining USC, I was an AI Resident at the FPT AI Center. I graduated with a bachelor's degree in Computer Science at Ho Chi Minh City University of Science.

Nowadays, I'm particularly interested in:

  • In-context robot learning: Enabling robots to generalize to novel tasks and embodiments with minimal human supervision, leveraging episodic memory and in-context learning to self-correct and refine behavior at inference time — without retraining.
  • Humanoid loco-manipulation: Developing whole-body control models that exploit the humanoid form factor to perform dexterous, meaningful tasks in everyday indoor environments.

Under Submission

ICLR: In-Context Imitation Learning with Visual Reasoning

Toan Nguyen, Weiduo Yuan, Songlin Wei, Hui Li, Daniel Seita†, Yue Wang†

Under Submission

Publications

Other Projects

USC CSCI 699 - Robotic Perception

Active Perception with VLMs

We developed an active perception system leveraging vision-language models on a Dexmate Vega humanoid.

3D Policy Learning with 3D Foundation Models

USC CSCI 699 - Robot Learning

3D Policy Learning with 3D Foundation Models

We leveraged 3D foundation models to reconstruct rich 3D scene representations for policy learning on an SO-100 robot arm, eliminating the need for expensive depth sensors.

Professional Services

  • Conference Reviewer: ICRA 2026, WACV 2026, IROS 2025, ECCV 2024, ICRA 2024, IROS 2024
  • Journal Reviewer: RA-L 2025
  • Organizer: UROS, a cross-department robotics reading group and seminar series at USC.