Learning Conceptual Space Representations of Interrelated Concepts

TL;DR

This paper introduces a Bayesian approach to learn conceptual space representations of interrelated concepts, enabling better modeling and prediction in knowledge base tasks by incorporating background knowledge.

Contribution

It proposes a novel Bayesian model that constructs informative priors from interrelated concept knowledge, improving conceptual representation learning.

Findings

Enhanced prediction accuracy in knowledge base completion

Better modeling of concepts with limited instance data

Effective incorporation of background knowledge into concept representations

Abstract

Several recently proposed methods aim to learn conceptual space representations from large text collections. These learned representations asso- ciate each object from a given domain of interest with a point in a high-dimensional Euclidean space, but they do not model the concepts from this do- main, and can thus not directly be used for catego- rization and related cognitive tasks. A natural solu- tion is to represent concepts as Gaussians, learned from the representations of their instances, but this can only be reliably done if sufficiently many in- stances are given, which is often not the case. In this paper, we introduce a Bayesian model which addresses this problem by constructing informative priors from background knowledge about how the concepts of interest are interrelated with each other. We show that this leads to substantially better pre- dictions in a knowledge base completion task.