On Deep Multi-View Representation Learning: Objectives and Optimization

TL;DR

This paper analyzes various deep multi-view representation learning techniques, highlighting the effectiveness of correlation-based methods and introducing the novel DCCAE approach, with empirical validation across multiple data modalities.

Contribution

The paper provides a comprehensive comparison of existing multi-view learning methods and introduces DCCAE, a new variant that improves performance on diverse tasks.

Findings

Correlation-based methods outperform autoencoder-like techniques.

DCCAE achieves the best results on most tasks.

Stochastic optimization enhances minibatch correlation training.

Abstract

We consider learning representations (features) in the setting in which we have access to multiple unlabeled views of the data for learning while only one view is available for downstream tasks. Previous work on this problem has proposed several techniques based on deep neural networks, typically involving either autoencoder-like networks with a reconstruction objective or paired feedforward networks with a batch-style correlation-based objective. We analyze several techniques based on prior work, as well as new variants, and compare them empirically on image, speech, and text tasks. We find an advantage for correlation-based representation learning, while the best results on most tasks are obtained with our new variant, deep canonically correlated autoencoders (DCCAE). We also explore a stochastic optimization procedure for minibatch correlation-based objectives and discuss the time/performance trade-offs for kernel-based and neural network-based implementations.