A Systematic Survey of Semantic Web Technologies for Bias in Artificial Intelligence Solutions

Tracking #: 2867-4081

This paper is currently under review
Paula Reyero Lobo
Enrico Daga
Harith Alani1
Miriam Fernandez1

Responsible editor: 
Dagmar Gromann

Submission type: 
Survey Article
Bias in artificial intelligence (AI) is a critical and timely issue due to its sociological, economic and legal impact, as decisions made for humans by algorithms could lead to unfair treatment of certain individuals or groups of individuals. Multiple surveys have emerged to give a multidisciplinary overview of bias [1-3] or to review bias in applied areas such as social sciences [4-6], business research [7], criminal justice [8], or data mining [9-14]. Due to the capability of Semantic Web (SW) technologies to fulfil data validity gaps in many AI areas [15], we revise the extent to which they can contribute to bringing solutions to this problem. To the best of our knowledge, there exists no previous work to bring together bias and semantics, so we review their intersectionality following a systematic approach [16]. Consequently, we provide in-depth analysis and categorisation of different types and sources of bias addressed with semantic approaches and discuss their advantages to improve frequent limitations in AI systems. We find works in the areas of information retrieval, recommendation systems, machine and deep learning, and natural language processing, and argue through multiple use cases that semantics can help especially dealing with technical, sociological, and psychological challenges.
Full PDF Version: 
Under Review