S2M: Converting Single-Turn to Multi-Turn Datasets for Conversational Question Answering
This addresses the data scarcity issue in CQA by effectively utilizing abundant single-turn datasets, though it is an incremental improvement in data augmentation.
The paper tackles the problem of improving conversational question answering (CQA) by converting single-turn datasets into multi-turn datasets, resulting in a method that achieved first place on the QuAC leaderboard as of August 2022.
Supplying data augmentation to conversational question answering (CQA) can effectively improve model performance. However, there is less improvement from single-turn datasets in CQA due to the distribution gap between single-turn and multi-turn datasets. On the other hand, while numerous single-turn datasets are available, we have not utilized them effectively. To solve this problem, we propose a novel method to convert single-turn datasets to multi-turn datasets. The proposed method consists of three parts, namely, a QA pair Generator, a QA pair Reassembler, and a question Rewriter. Given a sample consisting of context and single-turn QA pairs, the Generator obtains candidate QA pairs and a knowledge graph based on the context. The Reassembler utilizes the knowledge graph to get sequential QA pairs, and the Rewriter rewrites questions from a conversational perspective to obtain a multi-turn dataset S2M. Our experiments show that our method can synthesize effective training resources for CQA. Notably, S2M ranks 1st place on the QuAC leaderboard at the time of submission (Aug 24th, 2022).