Regex in a Time of Deep Learning: The Role of an Old Technology in Age Discrimination Detection in Job Advertisements

TL;DR

This paper examines the continued effectiveness of regex-based methods for detecting age discrimination in Dutch job ads, comparing them to neural embedding approaches and highlighting their respective strengths and limitations.

Contribution

It provides a qualitative analysis of regex advantages in age discrimination detection and explores how neural embeddings can overcome regex limitations.

Findings

Regex approaches are still strong performers in age discrimination detection.

Neural embeddings can address limitations of regex methods.

A comparative analysis of regex and neural embedding techniques.

Abstract

Deep learning holds great promise for detecting discriminatory language in the public sphere. However, for the detection of illegal age discrimination in job advertisements, regex approaches are still strong performers. In this paper, we investigate job advertisements in the Netherlands. We present a qualitative analysis of the benefits of the 'old' approach based on regexes and investigate how neural embeddings could address its limitations.