Answering Complex Open-Domain Questions with Multi-Hop Dense Retrieval
This addresses the problem of efficiently retrieving information for complex queries in open-domain question answering, with incremental improvements in speed and applicability.
The paper tackled answering complex open-domain questions by proposing a multi-hop dense retrieval approach, achieving state-of-the-art performance on HotpotQA and multi-evidence FEVER datasets with a 10x faster inference time while matching the best accuracy.
We propose a simple and efficient multi-hop dense retrieval approach for answering complex open-domain questions, which achieves state-of-the-art performance on two multi-hop datasets, HotpotQA and multi-evidence FEVER. Contrary to previous work, our method does not require access to any corpus-specific information, such as inter-document hyperlinks or human-annotated entity markers, and can be applied to any unstructured text corpus. Our system also yields a much better efficiency-accuracy trade-off, matching the best published accuracy on HotpotQA while being 10 times faster at inference time.