Modelling Commonsense Commonalities with Multi-Facet Concept Embeddings

TL;DR

This paper tackles the limitation of standard concept embeddings in capturing diverse commonsense commonalities by introducing multi-facet concept embeddings learned via facet masking. It presents a tri-encoder bi-encoder framework (Con, Prop, Facet) and trains on ChatGPT- and ConceptNet-derived data to optimize both concept-property alignment and facet-consistency via L1 and L2 losses. Across tasks including predicting commonsense properties, outlier detection, ontology completion, and ultra-fine entity typing, facet-aware representations consistently outperform baselines, with the strongest gains when combining data sources and using larger models. The work demonstrates that dynamic, facet-based representations improve inductive generalisation and open-domain applicability, suggesting broad utility for structured knowledge tasks and downstream recognition systems.

Abstract

Concept embeddings offer a practical and efficient mechanism for injecting commonsense knowledge into downstream tasks. Their core purpose is often not to predict the commonsense properties of concepts themselves, but rather to identify commonalities, i.e.\ sets of concepts which share some property of interest. Such commonalities are the basis for inductive generalisation, hence high-quality concept embeddings can make learning easier and more robust. Unfortunately, standard embeddings primarily reflect basic taxonomic categories, making them unsuitable for finding commonalities that refer to more specific aspects (e.g.\ the colour of objects or the materials they are made of). In this paper, we address this limitation by explicitly modelling the different facets of interest when learning concept embeddings. We show that this leads to embeddings which capture a more diverse range of commonsense properties, and consistently improves results in downstream tasks such as ultra-fine entity typing and ontology completion.