Testing Individual Fairness in Graph Neural Networks
Roya Nasiri
TL;DR
This paper addresses the gap of individual fairness in Graph Neural Networks by proposing a design-science–driven framework that adapts fairness testing to graph-structured data. It conducts a systematic literature review to build a taxonomy of definitions, metrics, testing methods, and mitigations, then develops a GNN-specific test-case generation approach that preserves graph structure and introduces a fairness-neuron coverage adequacy metric. The framework aims to generate natural individual discriminatory instances (IDIs) and evaluate them with test oracles to detect and mitigate biases, with industrial validation planned through collaboration with Deloitte on graph-based LLMs. The work seeks to deliver a practical toolkit for auditing and improving fairness in GNNs without sacrificing performance, thereby enhancing transparency and trust in graph-enabled AI applications.
Abstract
The biases in artificial intelligence (AI) models can lead to automated decision-making processes that discriminate against groups and/or individuals based on sensitive properties such as gender and race. While there are many studies on diagnosing and mitigating biases in various AI models, there is little research on individual fairness in Graph Neural Networks (GNNs). Unlike traditional models, which treat data features independently and overlook their inter-relationships, GNNs are designed to capture graph-based structure where nodes are interconnected. This relational approach enables GNNs to model complex dependencies, but it also means that biases can propagate through these connections, complicating the detection and mitigation of individual fairness violations. This PhD project aims to develop a testing framework to assess and ensure individual fairness in GNNs. It first systematically reviews the literature on individual fairness, categorizing existing approaches to define, measure, test, and mitigate model biases, creating a taxonomy of individual fairness. Next, the project will develop a framework for testing and ensuring fairness in GNNs by adapting and extending current fairness testing and mitigation techniques. The framework will be evaluated through industrial case studies, focusing on graph-based large language models.