Multi-Paragraph Reasoning with Knowledge-enhanced Graph Neural Network

Abstract

Multi-paragraph reasoning is indispensable for open-domain question answering (OpenQA), which receives less attention in the current OpenQA systems. In this work, we propose a knowledge-enhanced graph neural network (KGNN), which performs reasoning over multiple paragraphs with entities. To explicitly capture the entities' relatedness, KGNN utilizes relational facts in knowledge graph to build the entity graph. The experimental results show that KGNN outperforms in both distractor and full wiki settings than baselines methods on HotpotQA dataset. And our further analysis illustrates KGNN is effective and robust with more retrieved paragraphs.

References in corpus (3)

• SearchQA: A New Q&A Dataset Augmented with Context from a Search Engine
• Quasar: Datasets for Question Answering by Search and Reading
• HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering

Cited by in corpus (6)

• A Survey on Explainability in Machine Reading Comprehension
• Towards Interpretable and Reliable Reading Comprehension: A Pipeline Model with Unanswerability Prediction
• Generating Followup Questions for Interpretable Multi-hop Question Answering
• SRLGRN: Semantic Role Labeling Graph Reasoning Network
• Parameter-Efficient Neural Question Answering Models via Graph-Enriched Document Representations
• Machine Reading Comprehension using Case-based Reasoning