Approches d'analyse distributionnelle pour am\'eliorer la d\'esambigu\"isation s\'emantique

TL;DR

This paper introduces two distributional analysis methods to improve word sense disambiguation by reducing complexity and enhancing accuracy, comparing different neighbor selection strategies to optimize semantic context analysis.

Contribution

It proposes novel distributional analysis techniques for WSD that reduce computational complexity while maintaining coherence, and compares neighbor selection methods for better disambiguation.

Findings

Distributional neighbors outperform linearly nearest neighbors in WSD accuracy.

The proposed methods reduce exponential complexity without losing semantic coherence.

Results demonstrate improved disambiguation performance using distributional analysis.

Abstract

Word sense disambiguation (WSD) improves many Natural Language Processing (NLP) applications such as Information Retrieval, Machine Translation or Lexical Simplification. WSD is the ability of determining a word sense among different ones within a polysemic lexical unit taking into account the context. The most straightforward approach uses a semantic proximity measure between the word sense candidates of the target word and those of its context. Such a method very easily entails a combinatorial explosion. In this paper, we propose two methods based on distributional analysis which enable to reduce the exponential complexity without losing the coherence. We present a comparison between the selection of distributional neighbors and the linearly nearest neighbors. The figures obtained show that selecting distributional neighbors leads to better results.