WiC: the Word-in-Context Dataset for Evaluating Context-Sensitive Meaning Representations

TL;DR

The paper introduces WiC, a large-scale dataset designed to evaluate the ability of models to understand words' meanings in different contexts, addressing limitations of existing benchmarks.

Contribution

It presents WiC, a new dataset for evaluating context-sensitive word representations, filling a gap left by the shortcomings of previous benchmarks.

Findings

Existing models surpass the Stanford Contextual Word Similarity performance ceiling.

WiC provides a more suitable benchmark for dynamic semantics evaluation.

WiC is curated by experts and publicly available.

Abstract

By design, word embeddings are unable to model the dynamic nature of words' semantics, i.e., the property of words to correspond to potentially different meanings. To address this limitation, dozens of specialized meaning representation techniques such as sense or contextualized embeddings have been proposed. However, despite the popularity of research on this topic, very few evaluation benchmarks exist that specifically focus on the dynamic semantics of words. In this paper we show that existing models have surpassed the performance ceiling of the standard evaluation dataset for the purpose, i.e., Stanford Contextual Word Similarity, and highlight its shortcomings. To address the lack of a suitable benchmark, we put forward a large-scale Word in Context dataset, called WiC, based on annotations curated by experts, for generic evaluation of context-sensitive representations. WiC is released in https://pilehvar.github.io/wic/.