Breaking the Dyadic Barrier: Rethinking Fairness in Link Prediction Beyond Demographic Parity

TL;DR

This paper challenges the traditional dyadic fairness approach in link prediction, highlighting its limitations, and introduces a new framework and method that better detect and mitigate systemic biases, improving fairness without sacrificing utility.

Contribution

It formalizes the limitations of existing fairness evaluations in link prediction and proposes a novel framework and a lightweight post-processing method for improved bias mitigation.

Findings

The proposed method achieves state-of-the-art fairness-utility trade-offs.

It reveals systemic biases hidden by dyadic fairness definitions.

The framework enables more expressive fairness assessments.

Abstract

Link prediction is a fundamental task in graph machine learning with applications, ranging from social recommendation to knowledge graph completion. Fairness in this setting is critical, as biased predictions can exacerbate societal inequalities. Prior work adopts a dyadic definition of fairness, enforcing fairness through demographic parity between intra-group and inter-group link predictions. However, we show that this dyadic framing can obscure underlying disparities across subgroups, allowing systemic biases to go undetected. Moreover, we argue that demographic parity does not meet desired properties for fairness assessment in ranking-based tasks such as link prediction. We formalize the limitations of existing fairness evaluations and propose a framework that enables a more expressive assessment. Additionally, we propose a lightweight post-processing method combined with decoupled link predictors that effectively mitigates bias and achieves state-of-the-art fairness-utility trade-offs.