Improving Generalization for Abstract Reasoning Tasks Using Disentangled Feature Representations
This addresses generalization challenges in AI for abstract reasoning tasks, but it is incremental as it builds on existing disentangled VAE methods.
The paper tackled generalization in abstract reasoning tasks by using disentangled VAE representations, showing that unsupervised learning with the right objective outperforms supervised learning, with significant gains in generalization performance.
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.