Learning Sparse Sentence Encoding without Supervision: An Exploration of Sparsity in Variational Autoencoders

TL;DR

This paper explores inducing sparsity in variational autoencoders for text, proposing a hierarchical sparse VAE model to improve stability and examining how sparsity affects text classification performance.

Contribution

It introduces a hierarchical sparse VAE model for text, addressing stability issues and analyzing the impact of sparsity on downstream NLP tasks.

Findings

Hierarchical sparse VAE improves stability over state-of-the-art methods.

Sparsity in latent representations correlates with better text classification.

Sparse representations effectively encode task-related information.

Abstract

It has been long known that sparsity is an effective inductive bias for learning efficient representation of data in vectors with fixed dimensionality, and it has been explored in many areas of representation learning. Of particular interest to this work is the investigation of the sparsity within the VAE framework which has been explored a lot in the image domain, but has been lacking even a basic level of exploration in NLP. Additionally, NLP is also lagging behind in terms of learning sparse representations of large units of text e.g., sentences. We use the VAEs that induce sparse latent representations of large units of text to address the aforementioned shortcomings. First, we move in this direction by measuring the success of unsupervised state-of-the-art (SOTA) and other strong VAE-based sparsification baselines for text and propose a hierarchical sparse VAE model to address the stability issue of SOTA. Then, we look at the implications of sparsity on text classification across 3 datasets, and highlight a link between performance of sparse latent representations on downstream tasks and its ability to encode task-related information.