Generation-Augmented Retrieval for Open-domain Question Answering

TL;DR

Generation-Augmented Retrieval (GAR) enhances open-domain question answering by generating relevant contexts to improve retrieval accuracy, outperforming existing dense methods and combining well with them for state-of-the-art results.

Contribution

The paper introduces GAR, a novel method that uses text generation to augment queries, improving retrieval performance without external resources.

Findings

GAR achieves state-of-the-art results on Natural Questions and TriviaQA.

Generated diverse contexts improve retrieval accuracy.

Combining GAR with dense retrieval methods yields better performance.

Abstract

We propose Generation-Augmented Retrieval (GAR) for answering open-domain questions, which augments a query through text generation of heuristically discovered relevant contexts without external resources as supervision. We demonstrate that the generated contexts substantially enrich the semantics of the queries and GAR with sparse representations (BM25) achieves comparable or better performance than state-of-the-art dense retrieval methods such as DPR. We show that generating diverse contexts for a query is beneficial as fusing their results consistently yields better retrieval accuracy. Moreover, as sparse and dense representations are often complementary, GAR can be easily combined with DPR to achieve even better performance. GAR achieves state-of-the-art performance on Natural Questions and TriviaQA datasets under the extractive QA setup when equipped with an extractive reader, and consistently outperforms other retrieval methods when the same generative reader is used.