Kuan-Hsun Tu

I am a research assistant working with Tsung-Wei Ke at National Taiwan University. My research studies why powerful learning systems fall short of their theoretical capabilities, and how we can address these failures with simple and general mechanisms. I currently work on flow matching models, behavioral foundation models, and reinforcement learning for robotics. I am seeking PhD opportunities for Fall 2027.

Kuan-Hsun Tu playing an acoustic guitar

Publications

Task Inference Beyond Least Squares in Behavioral Foundation Model

Kuan-Hsun Tu*, Chien-Sheng Chiang*, Hsin-Wei Chen, Ping-Chun Hsieh, Tsung-Wei Ke

In submission, 2026.

CNP-Flow: Unified Temporal Flow Matching via Conditional Noise Prediction

Kuan-Hsun Tu*, Hsuan-Chi Liu†*, Jia-Wei Liao*, Chien-Sheng Chiang, Cheng-Fu Chou, Tsung-Wei Ke

In submission, 2026.

COT-FM comparison of transport paths on 5-Gaussians and Two Moons distributions

COT-FM: Cluster-wise Optimal Transport Flow Matching

Chien-Sheng Chiang*, Kuan-Hsun Tu†*, Jia-Wei Liao*, Cheng-Fu Chou, Tsung-Wei Ke

CVPR 2026

RobotArena translates real videos into simulated environments for robot policy evaluation

RobotArena ∞: Scalable Robot Benchmarking via Real-to-Sim Translation

Yash Jangir, Yidi Zhang*, Kashu Yamazaki*, Chenyu Zhang*, Kuan-Hsun Tu, Tsung-Wei Ke, Lei Ke, Yonatan Bisk, Katerina Fragkiadaki

ICLR 2026

DexMan pairs human demonstrations with bimanual robot manipulation of everyday objects

DexMan: Learning Bimanual Dexterous Manipulation from Human and Generated Videos

Jhen Hsieh, Kuan-Hsun Tu, Kuo-Han Hung, Tsung-Wei Ke

arXiv preprint, 2025.

* Equal contribution. † Corresponding author.

Honors and Awards

Work and Teaching Experience

Spring 2026

Advanced Network Administration and System Administration

TA

Spring & Fall 2025

Network Administration and System Administration

Lead TA

Fall 2025

Parallel Programming

TA