Early Visual Concept Learning with Unsupervised Deep Learning

TL;DR

This paper presents an unsupervised deep learning method inspired by neuroscience to learn disentangled visual concepts from raw images, enabling zero-shot inference and understanding of objectness.

Contribution

It introduces a variational autoencoder framework that enforces redundancy reduction and statistical independence to learn disentangled factors without supervision.

Findings

Effective across various datasets

Enables zero-shot inference

Develops an intuitive understanding of objectness

Abstract

Automated discovery of early visual concepts from raw image data is a major open challenge in AI research. Addressing this problem, we propose an unsupervised approach for learning disentangled representations of the underlying factors of variation. We draw inspiration from neuroscience, and show how this can be achieved in an unsupervised generative model by applying the same learning pressures as have been suggested to act in the ventral visual stream in the brain. By enforcing redundancy reduction, encouraging statistical independence, and exposure to data with transform continuities analogous to those to which human infants are exposed, we obtain a variational autoencoder (VAE) framework capable of learning disentangled factors. Our approach makes few assumptions and works well across a wide variety of datasets. Furthermore, our solution has useful emergent properties, such as zero-shot inference and an intuitive understanding of "objectness".