Matt McWilliams

This project seeks to understand the effects of guidance scale on classifier-free sampling of conditional diffusion models. We trained a conditioned UNET diffusion model on the MNIST dataset and sampled using classifier-free guidance. Classifier-free guidance is essential for modern image generation diffusion models, allowing label-dependent features become more prominant simply by changing a variable, the guidance scale.

LemmaCraft is a visual math proof engine with a custom parser and internal mathematical model. It allows its users to derive Peano-arithmetic results from only a few axioms. It consists of four chapters designed to teach users the software and introduce key lemmas: a tutorial chapter followed by three chapters that each cover a new mathematical domain: addition, multiplication, and exponentiation. In each chapter, players prove progressively more challenging lemmas about basic operations.