Han Yang (杨涵)

I am a fourth-year undergraduate student at The Chinese University of Hong Kong (CUHK), majoring in Computer Science. Currently, I am fortunate to be advised by Prof. Chuang Gan at University of Massachusetts Amherst (UMass) and the MIT-IBM Watson AI Lab, as well as Prof. Yilun Du at MIT. Prior to this, I was also honored to be advised by Prof. Chi-Wing Fu at CUHK.

My research experience covers various aspects of deep learning. Under the supervision of Prof. Chi-Wing Fu, I focused on low-level vision, self-supervised learning, and semi-supervised learning on shadow detection task. With Prof. Chuang Gan, I expanded my work to include audio processing, autoencoders, generative models (e.g., diffusion model) and large language models. My most recent research with Prof. Chuang Gan and Yilun Du involves building a lifelong memory system for embodied agents to perform embodied question-answering and lifelong navigation tasks.

My research interests lie in embodied AI and robot learning, with the ultimate goal of building human-like intelligent agents capable of navigating human environments and performing tasks with accuracy and robustness. Toward this goal, I am particularly interested in the following two topics: 1) Robotic perception: The intelligent agent should be able to actively and continuously perceive its surroundings, explore and navigate for given tasks, and build representations to memorize the newly explored environment. 2) Interaction with the physical world: The intelligent agent should possess both high-level decision-making and reasoning abilities, as well as robust low-level motion planning abilities, to accomplish complex tasks.