DeepAI AI Chat
Log In Sign Up

Are Disentangled Representations Helpful for Abstract Visual Reasoning?

by   Sjoerd van Steenkiste, et al.
ETH Zurich

A disentangled representation encodes information about the salient factors of variation in the data independently. Although it is often argued that this representational format is useful in learning to solve many real-world up-stream tasks, there is little empirical evidence that supports this claim. In this paper, we conduct a large-scale study that investigates whether disentangled representations are more suitable for abstract reasoning tasks. Using two new tasks similar to Raven's Progressive Matrices, we evaluate the usefulness of the representations learned by 360 state-of-the-art unsupervised disentanglement models. Based on these representations, we train 3600 abstract reasoning models and observe that disentangled representations do in fact lead to better up-stream performance. In particular, they appear to enable quicker learning using fewer samples.


page 4

page 7

page 9

page 15

page 23

page 24

page 25

page 26


Improving Generalization for Abstract Reasoning Tasks Using Disentangled Feature Representations

In this work we explore the generalization characteristics of unsupervis...

On the Transfer of Disentangled Representations in Realistic Settings

Learning meaningful representations that disentangle the underlying stru...

Weakly-Supervised Disentanglement Without Compromises

Intelligent agents should be able to learn useful representations by obs...

DAReN: A Collaborative Approach Towards Reasoning And Disentangling

Computational learning approaches to solving visual reasoning tests, suc...

Disentangled Representations from Non-Disentangled Models

Constructing disentangled representations is known to be a difficult tas...

Disentangled Representations in Neural Models

Representation learning is the foundation for the recent success of neur...

Learning Deep Disentangled Embeddings with the F-Statistic Loss

Deep-embedding methods aim to discover representations of a domain that ...