Bridging the Knowledge Gap: Enhancing Question Answering with World and Domain Knowledge
This addresses the knowledge gap in question answering for tasks involving world, causal, and domain-specific knowledge, representing a novel method for a known bottleneck.
The paper tackled the problem of enhancing question answering by injecting task-agnostic knowledge from an ontology into neural network pretraining, resulting in accuracy improvements of 33.3%, 18.6%, and 4% on three tasks and achieving new state-of-the-art results on two of them.
In this paper we present OSCAR (Ontology-based Semantic Composition Augmented Regularization), a method for injecting task-agnostic knowledge from an Ontology or knowledge graph into a neural network during pretraining. We evaluated the impact of including OSCAR when pretraining BERT with Wikipedia articles by measuring the performance when fine-tuning on two question answering tasks involving world knowledge and causal reasoning and one requiring domain (healthcare) knowledge and obtained 33:3%, 18:6%, and 4% improved accuracy compared to pretraining BERT without OSCAR and obtaining new state-of-the-art results on two of the tasks.