Generator-Retriever-Generator Approach for Open-Domain Question Answering

TL;DR

The paper introduces a Generator-Retriever-Generator (GRG) method that combines document generation and retrieval with large language models to improve open-domain question answering accuracy.

Contribution

It presents a novel hybrid approach that integrates document generation and retrieval, outperforming existing pipelines on multiple QA datasets.

Findings

Outperforms state-of-the-art methods on TriviaQA, NQ, and WebQ datasets.

Improves performance metrics by at least +1.6 to +5.2 points.

Demonstrates effectiveness of combining generated and retrieved documents.

Abstract

Open-domain question answering (QA) tasks usually require the retrieval of relevant information from a large corpus to generate accurate answers. We propose a novel approach called Generator-Retriever-Generator (GRG) that combines document retrieval techniques with a large language model (LLM), by first prompting the model to generate contextual documents based on a given question. In parallel, a dual-encoder network retrieves documents that are relevant to the question from an external corpus. The generated and retrieved documents are then passed to the second LLM, which generates the final answer. By combining document retrieval and LLM generation, our approach addresses the challenges of open-domain QA, such as generating informative and contextually relevant answers. GRG outperforms the state-of-the-art generate-then-read and retrieve-then-read pipelines (GENREAD and RFiD) improving their performance by at least by +5.2, +4.2, and +1.6 on TriviaQA, NQ, and WebQ datasets, respectively. We provide code, datasets, and checkpoints at https://github.com/abdoelsayed2016/GRG.