End-to-End Synthetic Data Generation for Domain Adaptation of Question Answering Systems

TL;DR

This paper introduces an end-to-end transformer-based method for generating synthetic question-answer pairs to improve domain adaptation in QA systems, eliminating the need for separate filtering models and achieving state-of-the-art results.

Contribution

The authors develop a unified transformer model trained end-to-end for QA data generation, streamlining the process and enhancing domain adaptation performance.

Findings

Significant improvements over existing methods in domain adaptation of QA models.

The end-to-end generator effectively filters generated data without separate models.

Achieves state-of-the-art results on benchmark datasets.

Abstract

We propose an end-to-end approach for synthetic QA data generation. Our model comprises a single transformer-based encoder-decoder network that is trained end-to-end to generate both answers and questions. In a nutshell, we feed a passage to the encoder and ask the decoder to generate a question and an answer token-by-token. The likelihood produced in the generation process is used as a filtering score, which avoids the need for a separate filtering model. Our generator is trained by fine-tuning a pretrained LM using maximum likelihood estimation. The experimental results indicate significant improvements in the domain adaptation of QA models outperforming current state-of-the-art methods.