Johns Hopkins University · B.S. expected 2029
Zhihan Shi / King
My interests lie at the intersection between AI/ML and neuroscience. I ask questions about how a mind behaves, whether artificial or human. I am currently working on psychiatric intake tools, affective valence reinforcement learning, and EEG analysis.
In one paragraph
My current work shares a common theme. How does a system decide that something mattered? In reinforcement learning, that looks like designing tasks to test an agent's causal inference. In psychiatric intake, it's measuring the little things clinicians do that set them apart from LLMs.
I came to Hopkins to engage in world-class interdisciplinary research. I still have three years left, but I think I've already found it. Currently, I lead a study at the PALS Lab comparing clinicians and large language models on psychiatric intake. I am also researching emotion-modulated deep RL agents in the Honey Lab. Lastly, I've delved into the startup world and shipped EEG-related tools with Elysius Labs.