Episode 103: MatterGen
Materialism: A Materials Science Podcast
00:00
Innovations in Generative Models for Materials Science
This chapter explores the evolution of generative models, particularly focusing on diffusion models and their applications in materials science. It discusses advanced techniques such as conditioned generation and classifier-free guidance, which enhance the discovery of novel materials. Additionally, the chapter addresses the challenges of creating stable materials and the methodologies for evaluating their properties, emphasizing the significance of data quality and representation.
Transcript
Play full episode
