Diffusion Model Invented at Stanford by Jascha Sohl-Dickstein

Jascha Sohl-Dickstein, a postdoc at Stanford’s Neural Dynamics and Computation Lab, invented the diffusion model, a massively significant to this day technique in machine learning. Diffusion models are the driving force behind our contemporary text-to-image models. Developed with Eric A. Weiss, Niru Maheswaranathan and the neuroscientist Surya Ganguli, the physics-based model leverages non-equilibrium thermodynamics to reverse the flow of time in a diffusion process. This novel approach transforms noise into data by training a neural network, allowing for highly detailed and accurate data generation. The diffusion model has since become a foundational tool in AI art, significantly impacting fields such as image and text generation, exemplified by its use in advanced models like DALL-E and Stable Diffusion.