Grounded control
I study how robots may learn and use information about objects, their relationships, and their history to guide manipulation. SlotVLA focuses on object relations; LIBERO-Mem looks at what a robot needs to remember across a task.
PhD Student · A*STAR/NUS · Robot Safety
Nhat Chung
I'm a PhD student at
National University of Singapore
and a graduate researcher at
A*STAR Singapore
. I work with
Prof. Ivor Tsang,
Prof. Jin-Song Dong,
and Prof. Qing Guo
in making robots safer whenever they share space with humans, particularly through
learning semantic-action representations and stress-testing.
Email · nhatcm [at] u.nus.edu
News
| Oct 2026 |
New preprint
Our survey, Towards Human-Centered Safety in Vision-Language-Action Robot Control, is available online! It examines how safety-relevant understanding guides robot actions and reviews methods for design, deployment, and validation.
|
|---|---|
| Feb 2026 |
Paper accepted
LIBERO-Mem @ AAAI 2026 studies whether a robot can remember earlier object interactions across a long manipulation task. It introduces an object-centric benchmark and a slot-based model for tracking that history.
|
| Jan 2026 |
Paper accepted
SlotVLA @ ICRA 2026 studies how compact object and relation representations support robot manipulation. It introduces the object-annotated LIBERO+ benchmark and a slot-based model that uses fewer visual tokens while maintaining competitive generalization.
|
| Jan 2026 |
Paper accepted
UNO @ WACV 2026 unifies box- and pixel-level video scene graph generation in one model. Its object and relation slots capture interactions across frames without a separate tracker.
|
| Aug 2025 |
Started my PhD at NUS School of Computing and joined A*STAR's CFAR as a graduate researcher.
|
| Jul 2025 |
Paper accepted
BiMa @ ACM Multimedia 2025 tackles bias in text-video retrieval. It uses scene elements and disentangled text features to focus on relevant content, improving retrieval on out-of-distribution videos.
|
About
I earned my B.Eng. and M.Eng. at International University, Vietnam National University HCMC, where I graduated top of my class. I worked with Dr. Tuan-Anh Vu on competitive programming and computer vision.
Before starting my PhD, I was an AI Resident at FPT Software. There, I worked with Prof. Ngan Le, Prof. Tung Kieu, and Prof. Anh Nguyen on robotics and multimodal systems.
My work has followed three related questions:
Stress testing
I test how vision and robot models behave when scenes are cluttered, ambiguous, or deliberately altered. OBEYED-VLA studies object and geometry grounding; DepthVanish, MAGIC, and work on typographic attacks probe failures in perception and reasoning.
Robust perception
My earlier work covers camouflaged object segmentation, robust background subtraction, and video scene graphs. These projects shaped my interest in the visual information robots rely on.
I'm now bringing those threads together to study when a robot should trust what it sees and remembers before it acts around people.
Experience
Sep 2019 – Nov 2023
Research, Teaching, and Administrative Assistant
Computer Vision and Image Processing Lab, Vietnam National University, Ho Chi Minh City, Vietnam
Professional Service
- Conference reviewer: CoRL 2026, NeurIPS 2026, WACV 2026, CVPR 2026, ECCV 2025, ICCV 2025.
- Journal reviewer: IEEE Transactions on Artificial Intelligence, IEEE Transactions on Image Processing, IEEE Robotics and Automation Letters, International Journal of Social Robotics.
Teaching
- Introduction to AI and Machine Learning, Head TA to Prof. Conghui Hu and Prof. Rizki Maulana @ National University of Singapore, Semester 1, 2026-2027.
- Introduction to AI and Machine Learning, TA to Prof. Conghui Hu, Prof. Patrick Rebentrost, and Prof. Rizki Maulana @ National University of Singapore, Semester 2, 2025-2026.
- Introduction to Machine Learning, TA to Prof. Vincent Ng @ Fulbright University Vietnam, 2023.
Awards and Honors
- Outstanding Reviewer, WACV 2026.
- Second Place, Robustness of Foundation Models Workshop, CVPR 2024.
- First Place, 7th AI City Challenge Workshop, Track 2, CVPR 2023.
- Second Place, 7th AI City Challenge Workshop, Track 1, CVPR 2023.
- Second Place, 5th AI City Challenge Workshop, Track 1, CVPR 2021.
- Gold Medal, International Collegiate Programming Contest, Asia Can Tho Regional Contest 2020.
- Second Prize, Vietnam National Olympiad of Informatics, Individual Contest 2018.