Diffusion models, seen through information theory

Meet Yassine Hamdi, an INFORMED AI postdoctoral researcher based at Imperial College London, who explores the theoretical foundations of diffusion models and their applications in conditional image generation, including medical imaging tasks such as MRI reconstruction. In this video, he explains how fundamental links between information and estimation can provide rigorous guarantees for diffusion models and help bridge the gap between theory and real-world AI systems. For more information about the research of INFORMED AI research hub see: https://informed-ai.net/