Improving Sparse Word Representations with Distributional Inference for Semantic Composition

TL;DR

This paper introduces a method to improve sparse word representations by explicitly inferring unobserved co-occurrences using distributional neighborhoods, enhancing interpretability and performance in semantic composition tasks.

Contribution

It proposes a novel distributional inference approach that enhances sparse word vectors, maintaining interpretability and achieving competitive results in semantic composition.

Findings

Improves performance on word similarity benchmarks

Achieves state-of-the-art results in semantic composition tasks

Maintains full interpretability of word representations

Abstract

Distributional models are derived from co-occurrences in a corpus, where only a small proportion of all possible plausible co-occurrences will be observed. This results in a very sparse vector space, requiring a mechanism for inferring missing knowledge. Most methods face this challenge in ways that render the resulting word representations uninterpretable, with the consequence that semantic composition becomes hard to model. In this paper we explore an alternative which involves explicitly inferring unobserved co-occurrences using the distributional neighbourhood. We show that distributional inference improves sparse word representations on several word similarity benchmarks and demonstrate that our model is competitive with the state-of-the-art for adjective-noun, noun-noun and verb-object compositions while being fully interpretable.