Multi-CrossRE A Multi-Lingual Multi-Domain Dataset for Relation Extraction

TL;DR

This paper introduces Multi-CrossRE, the largest multi-lingual relation extraction dataset covering 26 languages and six domains, enabling research beyond English and demonstrating high-quality translations through baseline experiments.

Contribution

It presents Multi-CrossRE, a novel multi-lingual, multi-domain relation extraction dataset based on machine translation of CrossRE, filling a major resource gap in non-English RE research.

Findings

Baseline results are consistent across original and back-translated data.

High translation quality confirmed by native speaker checks.

Dataset covers 26 languages and 6 domains, broadening RE research scope.

Abstract

Most research in Relation Extraction (RE) involves the English language, mainly due to the lack of multi-lingual resources. We propose Multi-CrossRE, the broadest multi-lingual dataset for RE, including 26 languages in addition to English, and covering six text domains. Multi-CrossRE is a machine translated version of CrossRE (Bassignana and Plank, 2022), with a sub-portion including more than 200 sentences in seven diverse languages checked by native speakers. We run a baseline model over the 26 new datasets and--as sanity check--over the 26 back-translations to English. Results on the back-translated data are consistent with the ones on the original English CrossRE, indicating high quality of the translation and the resulting dataset.