COMET: Commonsense Transformers for Automatic Knowledge Graph Construction

Abstract

We present the first comprehensive study on automatic knowledge base construction for two prevalent commonsense knowledge graphs: ATOMIC (Sap et al., 2019) and ConceptNet (Speer et al., 2017). Contrary to many conventional KBs that store knowledge with canonical templates, commonsense KBs only store loosely structured open-text descriptions of knowledge. We posit that an important step toward automatic commonsense completion is the development of generative models of commonsense knowledge, and propose COMmonsEnse Transformers (COMET) that learn to generate rich and diverse commonsense descriptions in natural language. Despite the challenges of commonsense modeling, our investigation reveals promising results when implicit knowledge from deep pre-trained language models is transferred to generate explicit knowledge in commonsense knowledge graphs. Empirical results demonstrate that COMET is able to generate novel knowledge that humans rate as high quality, with up to 77.5% (ATOMIC) and 91.7% (ConceptNet) precision at top 1, which approaches human performance for these resources. Our findings suggest that using generative commonsense models for automatic commonsense KB completion could soon be a plausible alternative to extractive methods.

Accepted to ACL 2019

References in corpus (2)

• Sequence to Sequence Learning with Neural Networks
• ConceptNet 5.5: An Open Multilingual Graph of General Knowledge

Cited by in corpus (14)

• Calibrate Before Use: Improving Few-Shot Performance of Language Models
• Abductive Commonsense Reasoning
• Playing Text-Based Games with Common Sense
• Joint Reasoning for Multi-Faceted Commonsense Knowledge
• Critical Thinking for Language Models
• Semantic Hypergraphs
• Natural Language QA Approaches using Reasoning with External Knowledge
• Think you have Solved Direct-Answer Question Answering? Try ARC-DA, the Direct-Answer AI2 Reasoning Challenge
• Modeling Human Mental States with an Entity-based Narrative Graph
• Semantic-Enhanced Explainable Finetuning for Open-Domain Dialogues
• Probing Pre-Trained Language Models for Disease Knowledge
• Modelling General Properties of Nouns by Selectively Averaging Contextualised Embeddings
• Analyzing Commonsense Emergence in Few-shot Knowledge Models
• CHARET: Character-centered Approach to Emotion Tracking in Stories