Investigating Lexical Replacements for Arabic-English Code-Switched Data Augmentation

TL;DR

This paper explores lexical replacement techniques for augmenting Arabic-English code-switched data, demonstrating that predictive models produce more natural sentences and significantly improve performance across NLP tasks.

Contribution

It introduces sequence-to-sequence based lexical replacements for CS data augmentation and compares them with dictionary-based methods, showing their effectiveness in NLP tasks.

Findings

Predictive model yields more natural code-switched sentences.

Both augmentation approaches outperform dictionary-based replacements.

Data augmentation improves MT, ASR, and ST performance significantly.

Abstract

Data sparsity is a main problem hindering the development of code-switching (CS) NLP systems. In this paper, we investigate data augmentation techniques for synthesizing dialectal Arabic-English CS text. We perform lexical replacements using word-aligned parallel corpora where CS points are either randomly chosen or learnt using a sequence-to-sequence model. We compare these approaches against dictionary-based replacements. We assess the quality of the generated sentences through human evaluation and evaluate the effectiveness of data augmentation on machine translation (MT), automatic speech recognition (ASR), and speech translation (ST) tasks. Results show that using a predictive model results in more natural CS sentences compared to the random approach, as reported in human judgements. In the downstream tasks, despite the random approach generating more data, both approaches perform equally (outperforming dictionary-based replacements). Overall, data augmentation achieves 34% improvement in perplexity, 5.2% relative improvement on WER for ASR task, +4.0-5.1 BLEU points on MT task, and +2.1-2.2 BLEU points on ST over a baseline trained on available data without augmentation.