CoVoSwitch: Machine Translation of Synthetic Code-Switched Text Based on Intonation Units

TL;DR

This paper introduces CoVoSwitch, a synthetic code-switching dataset for 13 languages, and evaluates how multilingual translation models perform on code-switched text, revealing strengths and limitations in translating into English and non-English languages.

Contribution

The paper presents CoVoSwitch, a novel synthetic dataset for code-switching translation, and provides comprehensive analysis of model performance on code-switched speech in multiple languages.

Findings

Models perform better on code-switching into English than non-English.

Low-resource languages benefit most when translating into English.

Models struggle with non-English tokens and hallucinate words absent in source sentences.

Abstract

Multilingual code-switching research is often hindered by the lack and linguistically biased status of available datasets. To expand language representation, we synthesize code-switching data by replacing intonation units detected through PSST, a speech segmentation model fine-tuned from OpenAI's Whisper, using a speech-to-text translation dataset, CoVoST 2. With our dataset, CoVoSwitch, spanning 13 languages, we evaluate the code-switching translation performance of two multilingual translation models, M2M-100 418M and NLLB-200 600M. We reveal that the inclusion of code-switching units results in higher translation performance than monolingual settings and that models are better at code-switching translation into English than non-English. Further, low-resource languages gain most from integration of code-switched units when translating into English but much less when translating into non-English. Translations into low-resource languages also perform worse than even raw code-switched inputs. We find that systems excel at copying English tokens but struggle with non-English tokens, that the off-target problem in monolingual settings is also relevant in code-switching settings, and that models hallucinate in code-switching translation by introducing words absent in both of the original source sentences. CoVoSwitch and code are available at https://github.com/sophiayk20/covoswitch.