Identifying and Explaining Discriminative Attributes

TL;DR

This paper introduces an explicit word vector model for identifying discriminative attributes, analyzing various data sources and knowledge bases to enhance explainability and achieve competitive performance in natural language inference tasks.

Contribution

The paper presents a novel explicit word vector model and a comprehensive analysis of data sources for improving interpretability in semantic representations.

Findings

Explicit vector spaces support discriminative attribute identification.

Different data sources provide complementary semantic information.

Model achieves F1-score of 0.69, comparable to state-of-the-art methods.

Abstract

Identifying what is at the center of the meaning of a word and what discriminates it from other words is a fundamental natural language inference task. This paper describes an explicit word vector representation model (WVM) to support the identification of discriminative attributes. A core contribution of the paper is a quantitative and qualitative comparative analysis of different types of data sources and Knowledge Bases in the construction of explainable and explicit WVMs: (i) knowledge graphs built from dictionary definitions, (ii) entity-attribute-relationships graphs derived from images and (iii) commonsense knowledge graphs. Using a detailed quantitative and qualitative analysis, we demonstrate that these data sources have complementary semantic aspects, supporting the creation of explicit semantic vector spaces. The explicit vector spaces are evaluated using the task of discriminative attribute identification, showing comparable performance to the state-of-the-art systems in the task (F1-score = 0.69), while delivering full model transparency and explainability.