Word Sense Disambiguation as a Game of Neurosymbolic Darts

TL;DR

This paper introduces a neurosymbolic approach to Word Sense Disambiguation that combines sense embeddings with logical relations, significantly surpassing previous F1 score benchmarks and demonstrating strong experimental results.

Contribution

The paper presents a novel neurosymbolic sense embedding method that encodes hypernym relations and enables logical deduction, improving WSD accuracy beyond 90%.

Findings

F1 scores range from 90.1% to 100% across datasets.

The approach encodes hypernym relations via nested balls in embedding space.

The method outperforms traditional deep-learning WSD models.

Abstract

Word Sense Disambiguation (WSD) is one of the hardest tasks in natural language understanding and knowledge engineering. The glass ceiling of 80% F1 score is recently achieved through supervised deep-learning, enriched by a variety of knowledge graphs. Here, we propose a novel neurosymbolic methodology that is able to push the F1 score above 90%. The core of our methodology is a neurosymbolic sense embedding, in terms of a configuration of nested balls in n-dimensional space. The centre point of a ball well-preserves word embedding, which partially fix the locations of balls. Inclusion relations among balls precisely encode symbolic hypernym relations among senses, and enable simple logic deduction among sense embeddings, which cannot be realised before. We trained a Transformer to learn the mapping from a contextualized word embedding to its sense ball embedding, just like playing the game of darts (a game of shooting darts into a dartboard). A series of experiments are conducted by utilizing pre-training n-ball embeddings, which have the coverage of around 70% training data and 75% testing data in the benchmark WSD corpus. The F1 scores in experiments range from 90.1% to 100.0% in all six groups of test data-sets (each group has 4 testing data with different sizes of n-ball embeddings). Our novel neurosymbolic methodology has the potential to break the ceiling of deep-learning approaches for WSD. Limitations and extensions of our current works are listed.