Commonsense Knowledge Base Completion with Structural and Semantic Context

TL;DR

This paper introduces novel methods for completing commonsense knowledge graphs by combining structural graph information and semantic context from pre-trained language models, improving link prediction performance.

Contribution

It proposes joint models leveraging graph convolutional networks and language model transfer learning, with empirical validation on ATOMIC and ConceptNet datasets.

Findings

Language models significantly boost link prediction accuracy.

Learning from local graph structure improves performance (+1.5 MRR on ConceptNet).

Models reveal types of commonsense knowledge well captured by language models.

Abstract

Automatic KB completion for commonsense knowledge graphs (e.g., ATOMIC and ConceptNet) poses unique challenges compared to the much studied conventional knowledge bases (e.g., Freebase). Commonsense knowledge graphs use free-form text to represent nodes, resulting in orders of magnitude more nodes compared to conventional KBs (18x more nodes in ATOMIC compared to Freebase (FB15K-237)). Importantly, this implies significantly sparser graph structures - a major challenge for existing KB completion methods that assume densely connected graphs over a relatively smaller set of nodes. In this paper, we present novel KB completion models that can address these challenges by exploiting the structural and semantic context of nodes. Specifically, we investigate two key ideas: (1) learning from local graph structure, using graph convolutional networks and automatic graph densification and (2) transfer learning from pre-trained language models to knowledge graphs for enhanced contextual representation of knowledge. We describe our method to incorporate information from both these sources in a joint model and provide the first empirical results for KB completion on ATOMIC and evaluation with ranking metrics on ConceptNet. Our results demonstrate the effectiveness of language model representations in boosting link prediction performance and the advantages of learning from local graph structure (+1.5 points in MRR for ConceptNet) when training on subgraphs for computational efficiency. Further analysis on model predictions shines light on the types of commonsense knowledge that language models capture well.