Michael Y. Li

I am a first-year Computer Science PhD student at Stanford University, where I'm advised by Noah Goodman and Emily Fox.

Previously, I graduated summa cum laude and Phi Beta Kappa from Princeton University, where I worked with Tom Griffiths and Ryan Adams.

Email  /  GitHub  /  LinkedIn

Research

I'm interested in understanding large language models. I'm also interested in probabilistic machine learning.

NAS-X: Neural Adaptive Smoothing via Twisting

We present a novel importance-sampling based method for approximate Bayesian inference that uses smoothing Sequential Monte Carlo to estimate gradients. We apply it to inference and model learning in discrete latent variable models and nonlinear dynamical systems.

Why think step-by-step? Reasoning emerges from the locality of experience

We show that reasoning emerges in large language models when training data is local.

Gaussian Process Surrogate Models for Neural Networks

We propose a framework that uses Gaussian processes to approximate neural networks. We use this framework to analyze neural network training dynamics and identify influential data points.

Learning to Learn Functions

We propose hierarchical Bayesian models of how people learn to learn functions and validate our model in behavioral experiments.

Design and source code from Jon Barron's website