Hua-Hsuan Liang

I recently graduated with an M.S. in Computer Science from Columbia University, where I worked as a research assistant in the ROAM Lab mentored by Zhanpeng He.

My research focuses on robotics and machine learning, with a particular emphasis on reinforcement learning, imitation learning, and human-in-the-loop learning for robotic manipulation. A central theme of my work is developing learning methods that operate under limited, noisy, or subjective supervision, which is a common challenge in real-world robotics.

During my graduate studies, I worked on reinforcement learning from human feedback (RLHF) for robotic manipulation and applied these methods in SpikeATac, a multimodal tactile sensing system for dexterous manipulation. In this project, I studied how sparse human preference labels and qualitative feedback can be used to train critic models and fine-tune manipulation policies beyond demonstration-based learning. My work involved designing RLHF training pipelines, integrating learned critics into reinforcement learning loops, and evaluating policy improvement in contact-rich manipulation tasks.

Before joining Columbia, I earned my undergraduate degree in Computer Science and Information Engineering from National Cheng Kung University, Taiwan. For my capstone project, I was advised by Professor Yeim-Kuan Chang. After graduating, I spent a year conducting research at DCNLab, where I was advised by Professor Chuan-Ching Sue.

I am currently preparing to pursue a Ph.D. and am seeking research opportunities focused on robot learning, RLHF, and data-efficient manipulation. Feel free to reach out if you would like to connect or discuss my work.