# Textual Data Augmentation for Arabic-English Code-Switching Speech Recognition

arXiv: https://arxiv.org/abs/2201.02550 · HTML: https://mlsanity.com/paper/2201.02550/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- Builds on 11 of its 39 references that are in the map.
- First posted on arXiv 7 Jan 2022.

## Builds on (most cited first, 11 of 11)

- [ESPnet: End-to-End Speech Processing Toolkit](https://mlsanity.com/paper/1804.00015/) (2018)
- [OpenNMT: Open-source Toolkit for Neural Machine Translation](https://mlsanity.com/paper/1709.03815/) (2017)
- [TED-LIUM 3: twice as much data and corpus repartition for experiments on speaker adaptation](https://mlsanity.com/paper/1805.04699/) (2018)
- [Word Alignment by Fine-tuning Embeddings on Parallel Corpora](https://mlsanity.com/paper/2101.08231/) (2021)
- [CoSDA-ML: Multi-Lingual Code-Switching Data Augmentation for Zero-Shot Cross-Lingual NLP](https://mlsanity.com/paper/2006.06402/) (2020)
- [Speech Recognition Challenge in the Wild: Arabic MGB-3](https://mlsanity.com/paper/1709.07276/) (2017)
- [Code-Switched Language Models Using Neural Based Synthetic Data from Parallel Sentences](https://mlsanity.com/paper/1909.08582/) (2019)
- [Towards One Model to Rule All: Multilingual Strategy for Dialectal Code-Switching Arabic ASR](https://mlsanity.com/paper/2105.14779/) (2021)
- [QASR: QCRI Aljazeera Speech Resource -- A Large Scale Annotated Arabic Speech Corpus](https://mlsanity.com/paper/2106.13000/) (2021)
- [Arabic Code-Switching Speech Recognition using Monolingual Data](https://mlsanity.com/paper/2107.01573/) (2021)
- [Investigations on Speech Recognition Systems for Low-Resource Dialectal Arabic-English Code-Switching Speech](https://mlsanity.com/paper/2108.12881/) (2021)

## Built on by (most cited first, 10 of 10)

- [Speech collage: code-switched audio generation by collaging monolingual corpora](https://mlsanity.com/paper/2309.15674/) (2023)
- [Investigating Lexical Replacements for Arabic-English Code-Switched Data Augmentation](https://mlsanity.com/paper/2205.12649/) (2022)
- [A Survey of Code-switched Arabic NLP: Progress, Challenges, and Future Directions](https://mlsanity.com/paper/2501.13419/) (2025)
- [Data Augmentation Techniques for Machine Translation of Code-Switched Texts: A Comparative Study](https://mlsanity.com/paper/2310.15262/) (2023)
- [Simple yet Effective Code-Switching Language Identification with Multitask Pre-Training and Transfer Learning](https://mlsanity.com/paper/2305.19759/) (2023)
- [Improving Code-Switching and Named Entity Recognition in ASR with Speech Editing based Data Augmentation](https://mlsanity.com/paper/2306.08588/) (2023)
- [Gated Low-rank Adaptation for personalized Code-Switching Automatic Speech Recognition on the low-spec devices](https://mlsanity.com/paper/2406.02562/) (2024)
- [Adapting Language Balance in Code-Switching Speech](https://mlsanity.com/paper/2510.18724/) (2025)
- [The Impact of Code-switched Synthetic Data Quality is Task Dependent: Insights from MT and ASR](https://mlsanity.com/paper/2503.23576/) (2025)
- [Bayesian Low-Rank Factorization for Robust Model Adaptation](https://mlsanity.com/paper/2510.18723/) (2025)

## Abstract (arXiv)

The pervasiveness of intra-utterance code-switching (CS) in spoken content requires that speech recognition (ASR) systems handle mixed language. Designing a CS-ASR system has many challenges, mainly due to data scarcity, grammatical structure complexity, and domain mismatch. The most common method for addressing CS is to train an ASR system with the available transcribed CS speech, along with monolingual data. In this work, we propose a zero-shot learning methodology for CS-ASR by augmenting the monolingual data with artificially generating CS text. We based our approach on random lexical replacements and Equivalence Constraint (EC) while exploiting aligned translation pairs to generate random and grammatically valid CS content. Our empirical results show a 65.5% relative reduction in language model perplexity, and 7.7% in ASR WER on two ecologically valid CS test sets. The human evaluation of the generated text using EC suggests that more than 80% is of adequate quality.