About
I'm a researcher at Princeton University working on machine learning for scientific discovery. I work with Prof. Ryan Adams at the Laboratory for Intelligent Probabilistic Systems (LIPS) on diffusion models that can dynamically adapt a sample's cardinality during generation, and with Prof. Sarah-Jane Leslie on understanding the internal representation spaces in language models from an information-theoretic perspective.
Education
Projects
Birth-Death Diffusion
A diffusion model that handles cardinality changes within the generative process
Neuromechanical Functions
Learning soft material morphologies to translate linear actuations into nonlinear deformations
Visuo-Haptic Illusions
Designing physical proxies to enable haptic interactions in virtual reality
Learnable Base Distributions
Can we learn base distributions that facilitate training/sampling in diffusion models?
Yasse-RL Halaby
Squash-playing RL agent training to challenge Yasser El Halaby, Princeton Class of 2006