Toward Interpretable Semantic Textual Similarity via Optimal Transport-based Contrastive Sentence Learning

TL;DR

This paper introduces an optimal transport-based method for interpretable semantic textual similarity, improving accuracy and providing human-aligned explanations through a novel contrastive learning framework.

Contribution

It proposes RCMD, a new distance measure based on optimal transport, and CLRCMD, a contrastive learning framework that enhances sentence similarity and interpretability.

Findings

Outperforms baselines on STS benchmarks

Provides human-aligned interpretability of sentence similarity

Enhances sentence similarity quality through contrastive learning

Abstract

Recently, finetuning a pretrained language model to capture the similarity between sentence embeddings has shown the state-of-the-art performance on the semantic textual similarity (STS) task. However, the absence of an interpretation method for the sentence similarity makes it difficult to explain the model output. In this work, we explicitly describe the sentence distance as the weighted sum of contextualized token distances on the basis of a transportation problem, and then present the optimal transport-based distance measure, named RCMD; it identifies and leverages semantically-aligned token pairs. In the end, we propose CLRCMD, a contrastive learning framework that optimizes RCMD of sentence pairs, which enhances the quality of sentence similarity and their interpretation. Extensive experiments demonstrate that our learning framework outperforms other baselines on both STS and interpretable-STS benchmarks, indicating that it computes effective sentence similarity and also provides interpretation consistent with human judgement. The code and checkpoint are publicly available at https://github.com/sh0416/clrcmd.