CORI: CJKV Benchmark with Romanization Integration -- A step towards Cross-lingual Transfer Beyond Textual Scripts

TL;DR

This paper emphasizes the importance of language contact in cross-lingual transfer, introduces a CJKV benchmark dataset, and proposes Romanization integration with contrastive learning to improve transfer performance.

Contribution

It highlights the significance of source language contact, creates a novel CJKV benchmark, and introduces Romanization with contrastive learning for better cross-lingual representations.

Findings

Romanization integration improves transfer accuracy.

Contact-aware source language selection enhances performance.

The CJKV benchmark facilitates in-depth language contact studies.

Abstract

Naively assuming English as a source language may hinder cross-lingual transfer for many languages by failing to consider the importance of language contact. Some languages are more well-connected than others, and target languages can benefit from transferring from closely related languages; for many languages, the set of closely related languages does not include English. In this work, we study the impact of source language for cross-lingual transfer, demonstrating the importance of selecting source languages that have high contact with the target language. We also construct a novel benchmark dataset for close contact Chinese-Japanese-Korean-Vietnamese (CJKV) languages to further encourage in-depth studies of language contact. To comprehensively capture contact between these languages, we propose to integrate Romanized transcription beyond textual scripts via Contrastive Learning objectives, leading to enhanced cross-lingual representations and effective zero-shot cross-lingual transfer.