Detection of Age Discrimination Semi-Automatic Detection of Age Discrimination in Dutch Job Advertisements Through Automatic Regex Generation
Keywords
Loading...
Authors
Issue Date
2020-02-01
Language
en
Document type
Journal Title
Journal ISSN
Volume Title
Publisher
Title
ISSN
Volume
Issue
Startpage
Endpage
DOI
Abstract
The employment context is the most common domain for age discrimination in the Netherlands, which
includes age discrimination in job advertisements. This study proposes a possible architecture for an age
discrimination detection approach using regular expressions and machine learning approaches, such as
active learning and genetic programming. We conducted two pre studies. First, we identified strengths
and weaknesses of regular expression-based age discrimination detection and (2) explored if the latent
space spanned by ALBERT word embeddings trained on a job description captures objective age discrimination
as defined by the Dutch Equal Employment Act. These findings build the foundation for the
proposed age discrimination architecture, which allows the automatic generation of regular expressions
with little annotation. Our experiments show that the approach is able to generate valid regexes, but
does not yet reach enough generalization to be used as the pure source of regexes for age discrimination
detection.
Description
Citation
Supervisor
Faculty
Faculteit der Sociale Wetenschappen