# Language Modelling Makes Sense: Propagating Representations through WordNet for Full-Coverage Word Sense Disambiguation

## TL;DR

This paper demonstrates that using sense-level embeddings derived from contextual language models and WordNet can significantly improve Word Sense Disambiguation, surpassing complex neural models with a simple nearest neighbor approach.

## Contribution

It introduces a method to create full-coverage sense embeddings from contextual models without task-specific tuning, enabling effective WSD with simple algorithms.

## Key findings

- Sense embeddings outperform previous neural models in WSD tasks.
- Robustness analysis reveals limitations when ignoring POS and lemma features.
- Sense embeddings facilitate concept-level analysis of language models.

## Abstract

Contextual embeddings represent a new generation of semantic representations learned from Neural Language Modelling (NLM) that addresses the issue of meaning conflation hampering traditional word embeddings. In this work, we show that contextual embeddings can be used to achieve unprecedented gains in Word Sense Disambiguation (WSD) tasks. Our approach focuses on creating sense-level embeddings with full-coverage of WordNet, and without recourse to explicit knowledge of sense distributions or task-specific modelling. As a result, a simple Nearest Neighbors (k-NN) method using our representations is able to consistently surpass the performance of previous systems using powerful neural sequencing models. We also analyse the robustness of our approach when ignoring part-of-speech and lemma features, requiring disambiguation against the full sense inventory, and revealing shortcomings to be improved. Finally, we explore applications of our sense embeddings for concept-level analyses of contextual embeddings and their respective NLMs.