mStyleDistance: Multilingual Style Embeddings and their Evaluation

TL;DR

This paper introduces mStyleDistance, a multilingual style embedding model trained on nine languages, which outperforms existing models in style analysis and verification tasks across multiple languages.

Contribution

The paper presents the first multilingual style embeddings trained with synthetic data and contrastive learning, along with a new benchmark for evaluating multilingual style quality.

Findings

Outperforms existing models on multilingual style benchmarks

Generalizes well to unseen languages and features

Available publicly at Hugging Face

Abstract

Style embeddings are useful for stylistic analysis and style transfer; however, only English style embeddings have been made available. We introduce Multilingual StyleDistance (mStyleDistance), a multilingual style embedding model trained using synthetic data and contrastive learning. We train the model on data from nine languages and create a multilingual STEL-or-Content benchmark (Wegmann et al., 2022) that serves to assess the embeddings' quality. We also employ our embeddings in an authorship verification task involving different languages. Our results show that mStyleDistance embeddings outperform existing models on these multilingual style benchmarks and generalize well to unseen features and languages. We make our model publicly available at https://huggingface.co/StyleDistance/mstyledistance .