Evaluating vector-space models of analogy

TL;DR

This paper evaluates how well modern vector-space models, like word2vec, capture human-like analogies and relational similarities, revealing both strengths and limitations of the parallelogram model in representing semantic relationships.

Contribution

It provides a detailed analysis of the parallelogram model's effectiveness in capturing semantic analogies in word embeddings and discusses its geometric limitations.

Findings

Some semantic relationships are well captured by the model.

The parallelogram model has intrinsic geometric limitations.

Certain relational similarities are poorly modeled.

Abstract

Vector-space representations provide geometric tools for reasoning about the similarity of a set of objects and their relationships. Recent machine learning methods for deriving vector-space embeddings of words (e.g., word2vec) have achieved considerable success in natural language processing. These vector spaces have also been shown to exhibit a surprising capacity to capture verbal analogies, with similar results for natural images, giving new life to a classic model of analogies as parallelograms that was first proposed by cognitive scientists. We evaluate the parallelogram model of analogy as applied to modern word embeddings, providing a detailed analysis of the extent to which this approach captures human relational similarity judgments in a large benchmark dataset. We find that that some semantic relationships are better captured than others. We then provide evidence for deeper limitations of the parallelogram model based on the intrinsic geometric constraints of vector spaces, paralleling classic results for first-order similarity.