Hypernyms under Siege: Linguistically-motivated Artillery for Hypernymy Detection

TL;DR

This paper evaluates various unsupervised, linguistically-motivated measures for hypernymy detection across different semantic models, highlighting their robustness compared to supervised methods which are sensitive to training data distribution.

Contribution

It provides a comprehensive analysis of unsupervised hypernymy detection methods based on linguistic principles, comparing them to supervised approaches and emphasizing their robustness.

Findings

Unsupervised measures are more robust than supervised ones.

Supervised methods outperform unsupervised but are sensitive to training data.

Linguistically-motivated measures remain useful for hypernymy detection.

Abstract

The fundamental role of hypernymy in NLP has motivated the development of many methods for the automatic identification of this relation, most of which rely on word distribution. We investigate an extensive number of such unsupervised measures, using several distributional semantic models that differ by context type and feature weighting. We analyze the performance of the different methods based on their linguistic motivation. Comparison to the state-of-the-art supervised methods shows that while supervised methods generally outperform the unsupervised ones, the former are sensitive to the distribution of training instances, hurting their reliability. Being based on general linguistic hypotheses and independent from training data, unsupervised measures are more robust, and therefore are still useful artillery for hypernymy detection.