M2QA: Multi-domain Multilingual Question Answering

TL;DR

M2QA is a new benchmark for evaluating multilingual question answering across multiple domains, revealing significant challenges in transferability and robustness of current models to diverse language-domain combinations.

Contribution

The paper introduces M2QA, a comprehensive multilingual, multi-domain QA benchmark, and investigates the performance of models across languages and domains, highlighting transferability issues.

Findings

Performance varies significantly across domain-language pairs.

All models show notable drops when transferring to new language-domain combinations.

Current models are far from fully solving multilingual multi-domain QA.

Abstract

Generalization and robustness to input variation are core desiderata of machine learning research. Language varies along several axes, most importantly, language instance (e.g. French) and domain (e.g. news). While adapting NLP models to new languages within a single domain, or to new domains within a single language, is widely studied, research in joint adaptation is hampered by the lack of evaluation datasets. This prevents the transfer of NLP systems from well-resourced languages and domains to non-dominant language-domain combinations. To address this gap, we introduce M2QA, a multi-domain multilingual question answering benchmark. M2QA includes 13,500 SQuAD 2.0-style question-answer instances in German, Turkish, and Chinese for the domains of product reviews, news, and creative writing. We use M2QA to explore cross-lingual cross-domain performance of fine-tuned models and state-of-the-art LLMs and investigate modular approaches to domain and language adaptation. We witness 1) considerable performance variations across domain-language combinations within model classes and 2) considerable performance drops between source and target language-domain combinations across all model sizes. We demonstrate that M2QA is far from solved, and new methods to effectively transfer both linguistic and domain-specific information are necessary. We make M2QA publicly available at https://github.com/UKPLab/m2qa.