Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Prompting Towards Alleviating Code-Switched Data Scarcity in Under-Resourced Languages with GPT as a Pivot

Michelle Terblanche, Kayode Olaleye, Vukosi Marivate

TL;DR

This work tackles code-switched data scarcity in under-resourced languages by leveraging GPT-3.5 prompting to generate Afrikaans–English and Yoruba–English CS sentences. It enhances prompts with topic-keyword lists, linguistic guidelines, and few-shot examples to boost data diversity, prompt adherence, and acceptability, while systematically comparing Afrikaans–English to Yoruba–English. The findings show higher quality and adherence for Afrikaans–English, with Yoruba–English exhibiting greater challenges due to non-Latin scripts and linguistic complexity, underscoring the need for language-specific prompting and human-in-the-loop evaluation. The paper proposes a linguistically-guided prompting framework and positions GPT as a practical pivot to mitigate CS data scarcity in low-resource languages, emphasizing native-speaker involvement for achieving usable synthetic data.

Abstract

Many multilingual communities, including numerous in Africa, frequently engage in code-switching during conversations. This behaviour stresses the need for natural language processing technologies adept at processing code-switched text. However, data scarcity, particularly in African languages, poses a significant challenge, as many are low-resourced and under-represented. In this study, we prompted GPT 3.5 to generate Afrikaans--English and Yoruba--English code-switched sentences, enhancing diversity using topic-keyword pairs, linguistic guidelines, and few-shot examples. Our findings indicate that the quality of generated sentences for languages using non-Latin scripts, like Yoruba, is considerably lower when compared with the high Afrikaans-English success rate. There is therefore a notable opportunity to refine prompting guidelines to yield sentences suitable for the fine-tuning of language models. We propose a framework for augmenting the diversity of synthetically generated code-switched data using GPT and propose leveraging this technology to mitigate data scarcity in low-resourced languages, underscoring the essential role of native speakers in this process.

Prompting Towards Alleviating Code-Switched Data Scarcity in Under-Resourced Languages with GPT as a Pivot

TL;DR

This work tackles code-switched data scarcity in under-resourced languages by leveraging GPT-3.5 prompting to generate Afrikaans–English and Yoruba–English CS sentences. It enhances prompts with topic-keyword lists, linguistic guidelines, and few-shot examples to boost data diversity, prompt adherence, and acceptability, while systematically comparing Afrikaans–English to Yoruba–English. The findings show higher quality and adherence for Afrikaans–English, with Yoruba–English exhibiting greater challenges due to non-Latin scripts and linguistic complexity, underscoring the need for language-specific prompting and human-in-the-loop evaluation. The paper proposes a linguistically-guided prompting framework and positions GPT as a practical pivot to mitigate CS data scarcity in low-resource languages, emphasizing native-speaker involvement for achieving usable synthetic data.

Abstract

Many multilingual communities, including numerous in Africa, frequently engage in code-switching during conversations. This behaviour stresses the need for natural language processing technologies adept at processing code-switched text. However, data scarcity, particularly in African languages, poses a significant challenge, as many are low-resourced and under-represented. In this study, we prompted GPT 3.5 to generate Afrikaans--English and Yoruba--English code-switched sentences, enhancing diversity using topic-keyword pairs, linguistic guidelines, and few-shot examples. Our findings indicate that the quality of generated sentences for languages using non-Latin scripts, like Yoruba, is considerably lower when compared with the high Afrikaans-English success rate. There is therefore a notable opportunity to refine prompting guidelines to yield sentences suitable for the fine-tuning of language models. We propose a framework for augmenting the diversity of synthetically generated code-switched data using GPT and propose leveraging this technology to mitigate data scarcity in low-resourced languages, underscoring the essential role of native speakers in this process.
Paper Structure (26 sections, 3 figures, 5 tables)

This paper contains 26 sections, 3 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Code-Switching Research
  4. Code-Switched Text Generation via GPT-3.5 Prompting
  5. Prompting for Afrikaans--English CS Sentences
  6. Topic-Keyword Basic Prompting
  7. Linguistic-Based Prompting
  8. Few-Shot Prompting
  9. Prompting for Yoruba--English CS Sentences
  10. Evaluation of Generated Data
  11. Data Diversity
  12. Content Diversity
  13. Linguistic Diversity
  14. Prompt Adherence
  15. Statistical Evaluation of Prompt Adherence
  16. ...and 11 more sections

Figures (3)

  • Figure 1: Distribution of top 10 general CS words across all topics.
  • Figure 2: Distribution of pronouns.
  • Figure 4: Evaluation of manual annotation of sentences.