Enriching Disentanglement: From Logical Definitions to Quantitative Metrics

arXiv:2305.11512v2 Announce Type: replace
Abstract: Disentangling the explanatory factors in complex data is a promising approach for generalizable and data-efficient representation learning. While a variety of quantitative metrics for learning and evaluating disentangled representations have been proposed, it remains unclear what properties these metrics truly quantify. In this work, we establish a theoretical connection between logical definitions of disentanglement and quantitative metrics using topos theory and enriched category theory. We introduce a systematic approach for converting a first-order predicate into …

abstract arxiv cs.lg data definitions metrics quantitative replace representation representation learning type while work