Using BERT for Word Sense Disambiguation

TL;DR

This paper leverages BERT to improve Word Sense Disambiguation by creating better sense representations and training a unified classifier, achieving state-of-the-art results on standard benchmarks.

Contribution

It introduces a novel approach combining BERT with sense definitions for a unified WSD classifier capable of disambiguating unseen polysemes.

Findings

Achieved state-of-the-art performance on English All-word WSD

Demonstrated effectiveness of sense definitions in training

Unified classifier handles unseen polysemes

Abstract

Word Sense Disambiguation (WSD), which aims to identify the correct sense of a given polyseme, is a long-standing problem in NLP. In this paper, we propose to use BERT to extract better polyseme representations for WSD and explore several ways of combining BERT and the classifier. We also utilize sense definitions to train a unified classifier for all words, which enables the model to disambiguate unseen polysemes. Experiments show that our model achieves the state-of-the-art results on the standard English All-word WSD evaluation.