DeepAI AI Chat
Log In Sign Up

Correcting Flaws in Common Disentanglement Metrics

by   Louis Mahon, et al.

Recent years have seen growing interest in learning disentangled representations, in which distinct features, such as size or shape, are represented by distinct neurons. Quantifying the extent to which a given representation is disentangled is not straightforward; multiple metrics have been proposed. In this paper, we identify two failings of existing metrics, which mean they can assign a high score to a model which is still entangled, and we propose two new metrics, which redress these problems. We then consider the task of compositional generalization. Unlike prior works, we treat this as a classification problem, which allows us to use it to measure the disentanglement ability of the encoder, without depending on the decoder. We show that performance on this task is (a) generally quite poor, (b) correlated with most disentanglement metrics, and (c) most strongly correlated with our newly proposed metrics.


page 4

page 12

page 13

page 14


Evaluating Disentangled Representations

There is no generally agreed upon definition of disentangled representat...

Linear Disentangled Representations and Unsupervised Action Estimation

Disentangled representation learning has seen a surge in interest over r...

Disentangled Sequence to Sequence Learning for Compositional Generalization

There is mounting evidence that existing neural network models, in parti...

The Impact of Correlated Metrics on Defect Models

Defect models are analytical models that are used to build empirical the...

Enriching Disentanglement: Definitions to Metrics

Disentangled representation learning is a challenging task that involves...

Measuring Disentanglement: A Review of Metrics

Learning to disentangle and represent factors of variation in data is an...