Commonsense Knowledge in Word Associations and ConceptNet

TL;DR

This paper compares ConceptNet and SWOW, two large-scale commonsense knowledge resources, analyzing their structures and demonstrating their utility in improving reasoning tasks.

Contribution

It provides an in-depth comparison of ConceptNet and SWOW, highlighting their differences, overlaps, and their combined effectiveness in commonsense reasoning.

Findings

Both resources improve performance on reasoning benchmarks.

SWOW, derived from crowd-sourced associations, complements curated knowledge graphs.

The resources encode different aspects of commonsense knowledge.

Abstract

Humans use countless basic, shared facts about the world to efficiently navigate in their environment. This commonsense knowledge is rarely communicated explicitly, however, understanding how commonsense knowledge is represented in different paradigms is important for both deeper understanding of human cognition and for augmenting automatic reasoning systems. This paper presents an in-depth comparison of two large-scale resources of general knowledge: ConcpetNet, an engineered relational database, and SWOW a knowledge graph derived from crowd-sourced word associations. We examine the structure, overlap and differences between the two graphs, as well as the extent to which they encode situational commonsense knowledge. We finally show empirically that both resources improve downstream task performance on commonsense reasoning benchmarks over text-only baselines, suggesting that large-scale word association data, which have been obtained for several languages through crowd-sourcing, can be a valuable complement to curated knowledge graphs