A Bayesian Unification of Self-Supervised Clustering and Energy-Based Models

arXiv:2401.00873v2 Announce Type: replace
Abstract: Self-supervised learning is a popular and powerful method for utilizing large amounts of unlabeled data, for which a wide variety of training objectives have been proposed in the literature. In this study, we perform a Bayesian analysis of state-of-the-art self-supervised learning objectives, elucidating the underlying probabilistic graphical models in each class and presenting a standardized methodology for their derivation from first principles. The analysis also indicates a natural means of integrating self-supervised learning with likelihood-based …

abstract analysis art arxiv bayesian clustering cs.cv cs.lg data energy literature popular self-supervised learning state study supervised learning training type unification