Separated by an Un-common Language: Towards Judgment Language Informed Vector Space Modeling

TL;DR

This study investigates how the language of judgment influences human semantic similarity scores and demonstrates that multilingual vector space models can better align with human judgments across different languages.

Contribution

The paper introduces multilingual evaluation of semantic models and shows multilingual VSMs can improve correlation with human judgments across languages.

Findings

Human judgments are significantly affected by judgment language.

Monolingual VSMs do not always align best with judgments in their training language.

Multilingual VSMs can enhance correlation with human semantic judgments.

Abstract

A common evaluation practice in the vector space models (VSMs) literature is to measure the models' ability to predict human judgments about lexical semantic relations between word pairs. Most existing evaluation sets, however, consist of scores collected for English word pairs only, ignoring the potential impact of the judgment language in which word pairs are presented on the human scores. In this paper we translate two prominent evaluation sets, wordsim353 (association) and SimLex999 (similarity), from English to Italian, German and Russian and collect scores for each dataset from crowdworkers fluent in its language. Our analysis reveals that human judgments are strongly impacted by the judgment language. Moreover, we show that the predictions of monolingual VSMs do not necessarily best correlate with human judgments made with the language used for model training, suggesting that models and humans are affected differently by the language they use when making semantic judgments. Finally, we show that in a large number of setups, multilingual VSM combination results in improved correlations with human judgments, suggesting that multilingualism may partially compensate for the judgment language effect on human judgments.