In-Person Poster presentation / poster accept

DCI-ES: An Extended Disentanglement Framework with Connections to Identifiability

Cian Eastwood · Andrei L Nicolicioiu · Julius von Kügelgen · Armin Kekic · Frederik Träuble · Andrea Dittadi · Bernhard Schoelkopf

MH1-2-3-4 #74

Keywords: [ Deep Learning and representational learning ] [ identifiability ] [ representation learning ] [ disentanglement ]

[ Abstract ]
[ Poster [ OpenReview
Wed 3 May 2:30 a.m. PDT — 4:30 a.m. PDT


In representation learning, a common approach is to seek representations which disentangle the underlying factors of variation. Eastwood & Williams (2018) proposed three metrics for quantifying the quality of such disentangled representations: disentanglement (D), completeness (C) and informativeness (I). In this work, we first connect this DCI framework to two common notions of linear and nonlinear identifiability, thereby establishing a formal link between disentanglement and the closely-related field of independent component analysis. We then propose an extended DCI-ES framework with two new measures of representation quality—explicitness (E) and size (S)—and point out how D and C can be computed for black-box predictors. Our main idea is that the functional capacity required to use a representation is an important but thus-far neglected aspect of representation quality, which we quantify using explicitness or ease-of-use (E). We illustrate the relevance of our extensions on the MPI3D and Cars3D datasets.

Chat is not available.