■Bibliographic Information
Manato Yaguchi(*) , Kotaro Sakamoto(*) , Ryosuke Sakamoto(*) , Masato Tanabe(*) , Masatomo Akagawa(*) , Yusuke Hayashi(*) , Masahiro Suzuki, Yutaka Matsuo “The Geometry of Phase Transitions in Diffusion Models: Tubular Neighbourhoods and Singularities”. Transactions on Machine Learning Research (TMLR).
(*) Equal Contribution
https://openreview.net/forum?id=ahVFKFLYk2
■Overview
Diffusion models undergo phase transitions during the generative process where data features suddenly emerge in the final stages.
The current study aims to elucidate this critical phenomenon from the geometrical perspective. Through theoretical and empirical evidence, we demonstrate that phase transitions in the generative process of diffusion models Through theoretical and empirical evidence, we demonstrate that phase transitions in the generative process of diffusion models are closely related to the injectivity radius. Our findings offer a novel perspective on phase transitions in diffusion models, with potential implications for improving performance and sampling efficiency.
［NEWS] Our paper has been accepted to Transactions on Machine Learning Research (Featured Certification).
