Training Generative Question-Answering on Synthetic Data Obtained from an Instruct-tuned Model
This addresses the challenge of data scarcity for non-English QA systems, offering a practical solution for resource-constrained languages, though it is incremental as it builds on existing instruct-tuning methods.
The paper tackles the problem of training question-answering systems for non-English languages by proposing a cost-effective method that uses an instruct-tuned model to generate synthetic QA pairs in zero-shot or few-shot settings, achieving comparable performance to models trained on manually curated datasets without human expenses.
This paper presents a simple and cost-effective method for synthesizing data to train question-answering systems. For training, fine-tuning GPT models is a common practice in resource-rich languages like English, however, it becomes challenging for non-English languages due to the scarcity of sufficient question-answer (QA) pairs. Existing approaches use question and answer generators trained on human-authored QA pairs, which involves substantial human expenses. In contrast, we use an instruct-tuned model to generate QA pairs in a zero-shot or few-shot manner. We conduct experiments to compare various strategies for obtaining QA pairs from the instruct-tuned model. The results demonstrate that a model trained on our proposed synthetic data achieves comparable performance to a model trained on manually curated datasets, without incurring human costs.