Split-Brain Autoencoders: Unsupervised Learning by Cross-Channel Prediction

TLDR

The authors introduce split‑brain autoencoders, a modified autoencoder that learns unsupervised representations by predicting one subset of data channels from another. The model splits into two disjoint sub‑networks, each trained to predict a different subset of channels, thereby extracting features from the full input signal. The method achieves state‑of‑the‑art performance on several large‑scale transfer‑learning benchmarks.

Abstract

We propose split-brain autoencoders, a straightforward modification of the traditional autoencoder architecture, for unsupervised representation learning. The method adds a split to the network, resulting in two disjoint sub-networks. Each sub-network is trained to perform a difficult task - predicting one subset of the data channels from another. Together, the sub-networks extract features from the entire input signal. By forcing the network to solve crosschannel prediction tasks, we induce a representation within the network which transfers well to other, unseen tasks. This method achieves state-of-the-art performance on several large-scale transfer learning benchmarks.