InfoSync: Information Synchronization across Multilingual Semi-structured Tables

TL;DR

This paper introduces InfoSync, a new dataset and a two-step method for synchronizing semi-structured tables across multiple languages, improving data consistency in multilingual Wikipedia tables.

Contribution

The paper presents a novel dataset, InfoSyncC, and a two-step synchronization method for multilingual semi-structured tables, addressing a key challenge in cross-lingual data consistency.

Findings

Information alignment F1 score of 87.91 for en <-> non-en

77.28% acceptance rate in Wikipedia edits

Effective cross-lingual table synchronization demonstrated

Abstract

Information Synchronization of semi-structured data across languages is challenging. For instance, Wikipedia tables in one language should be synchronized across languages. To address this problem, we introduce a new dataset InfoSyncC and a two-step method for tabular synchronization. InfoSync contains 100K entity-centric tables (Wikipedia Infoboxes) across 14 languages, of which a subset (3.5K pairs) are manually annotated. The proposed method includes 1) Information Alignment to map rows and 2) Information Update for updating missing/outdated information for aligned tables across multilingual tables. When evaluated on InfoSync, information alignment achieves an F1 score of 87.91 (en <-> non-en). To evaluate information updation, we perform human-assisted Wikipedia edits on Infoboxes for 603 table pairs. Our approach obtains an acceptance rate of 77.28% on Wikipedia, showing the effectiveness of the proposed method.