Research notes and technical explorations across learning systems, physics, and chemistry.
MiniMax-H3 is a large diffusion model that turns a text prompt into a five-second video with synchronised audio. This post is about what you can do with very little: one consum...
(★ This Blog is part of a review series on Learning Theory: Training Dynamics collection. ★) This introduction gives a concise outline of the theoretical machinery behind Neura...
This blog is an invited contribution by Dr. Shengchao Liu; see the Wave Intelligence Lab. Flow Matching: A Physicist’s Perspective One of the earliest and most influential sc...