Showing 1–1 of 1 results for author: Steenbrugge, X

Improving Generalization for Abstract Reasoning Tasks Using Disentangled Feature Representations

Authors: Xander Steenbrugge, Sam Leroux, Tim Verbelen, Bart Dhoedt

Abstract: In this work we explore the generalization characteristics of unsupervised representation learning by leveraging disentangled VAE's to learn a useful latent space on a set of relational reasoning problems derived from Raven Progressive Matrices. We show that the latent representations, learned by unsupervised training using the right objective function, significantly outperform the same architectu… ▽ More In this work we explore the generalization characteristics of unsupervised representation learning by leveraging disentangled VAE's to learn a useful latent space on a set of relational reasoning problems derived from Raven Progressive Matrices. We show that the latent representations, learned by unsupervised training using the right objective function, significantly outperform the same architectures trained with purely supervised learning, especially when it comes to generalization. △ Less

Submitted 12 November, 2018; originally announced November 2018.