Knowledge Guided Text Retrieval and Reading for Open Domain Question Answering

TL;DR

This paper presents a knowledge-guided passage graph retrieval and reading method for open-domain question answering, improving coverage and accuracy by leveraging external knowledge bases and passage relations.

Contribution

It introduces a novel graph-based retrieval and reading approach that enhances passage relevance and information fusion without requiring extensive end-to-end training.

Findings

Achieves 2-11% absolute improvement over non-graph baselines.

Matches or exceeds state-of-the-art performance on three QA datasets.

Does not rely on expensive end-to-end training.

Abstract

We introduce an approach for open-domain question answering (QA) that retrieves and reads a passage graph, where vertices are passages of text and edges represent relationships that are derived from an external knowledge base or co-occurrence in the same article. Our goals are to boost coverage by using knowledge-guided retrieval to find more relevant passages than text-matching methods, and to improve accuracy by allowing for better knowledge-guided fusion of information across related passages. Our graph retrieval method expands a set of seed keyword-retrieved passages by traversing the graph structure of the knowledge base. Our reader extends a BERT-based architecture and updates passage representations by propagating information from related passages and their relations, instead of reading each passage in isolation. Experiments on three open-domain QA datasets, WebQuestions, Natural Questions and TriviaQA, show improved performance over non-graph baselines by 2-11% absolute. Our approach also matches or exceeds the state-of-the-art in every case, without using an expensive end-to-end training regime.