MULTITAT: Benchmarking Multilingual Table-and-Text Question Answering

TL;DR

This paper introduces MULTITAT, the first multilingual dataset for table-and-text question answering, highlighting significant performance gaps in non-English languages and proposing a baseline model that outperforms others.

Contribution

The paper creates MULTITAT, a multilingual TATQA dataset in 10 languages, and develops a baseline model to evaluate and improve multilingual TATQA performance.

Findings

Performance drops by 19.4% on non-English data

Baseline model outperforms others by 3.3 points on average

Multilingual TATQA presents unique challenges and opportunities

Abstract

Question answering on the hybrid context of tables and text (TATQA) is a critical task, with broad applications in data-intensive domains. However, existing TATQA datasets are limited to English, leading to several drawbacks: (i) They overlook the challenges of multilingual TAT-QA and cannot assess model performance in the multilingual setting. (ii) They do not reflect real-world scenarios where tables and texts frequently appear in non-English languages. To address the limitations, we propose the first multilingual TATQA dataset (MULTITAT). Specifically, we sample data from 3 mainstream TATQA datasets and translate it into 10 diverse languages. To align the model TATQA capabilities in English with other languages, we develop a baseline, Ours. Experimental results reveal that the performance on non-English data in MULTITAT drops by an average of 19.4% compared to English, proving the necessity of MULTITAT. We further analyze the reasons for this performance gap. Furthermore, Ours outperforms other baselines by an average of 3.3, demonstrating its effectiveness.