Multilingual Coreference Resolution in Multiparty Dialogue

TL;DR

This paper introduces a large multilingual dataset for multiparty dialogue coreference resolution, leveraging TV transcripts and annotation projection to enable cross-lingual training and evaluation.

Contribution

The creation of the MMC dataset and the demonstration of effective cross-lingual training methods for coreference resolution in multiparty dialogues.

Findings

Existing models perform poorly on MMC, indicating broader coverage.

Silver data improves model performance in zero-shot cross-lingual settings.

MMC enables better evaluation of multiparty coreference across languages.

Abstract

Existing multiparty dialogue datasets for entity coreference resolution are nascent, and many challenges are still unaddressed. We create a large-scale dataset, Multilingual Multiparty Coref (MMC), for this task based on TV transcripts. Due to the availability of gold-quality subtitles in multiple languages, we propose reusing the annotations to create silver coreference resolution data in other languages (Chinese and Farsi) via annotation projection. On the gold (English) data, off-the-shelf models perform relatively poorly on MMC, suggesting that MMC has broader coverage of multiparty coreference than prior datasets. On the silver data, we find success both using it for data augmentation and training from scratch, which effectively simulates the zero-shot cross-lingual setting.