SkillMatch: Evaluating Self-supervised Learning of Skill Relatedness

TL;DR

This paper introduces SkillMatch, a benchmark for skill relatedness derived from job ads, and proposes a self-supervised learning method that significantly improves skill relatedness modeling over traditional approaches.

Contribution

It provides the first benchmark for skill relatedness evaluation and presents a scalable self-supervised learning method to enhance skill modeling accuracy.

Findings

Self-supervised model outperforms traditional models

SkillMatch benchmark enables direct evaluation of skill relatedness methods

Public release of SkillMatch facilitates future research

Abstract

Accurately modeling the relationships between skills is a crucial part of human resources processes such as recruitment and employee development. Yet, no benchmarks exist to evaluate such methods directly. We construct and release SkillMatch, a benchmark for the task of skill relatedness, based on expert knowledge mining from millions of job ads. Additionally, we propose a scalable self-supervised learning technique to adapt a Sentence-BERT model based on skill co-occurrence in job ads. This new method greatly surpasses traditional models for skill relatedness as measured on SkillMatch. By releasing SkillMatch publicly, we aim to contribute a foundation for research towards increased accuracy and transparency of skill-based recommendation systems.