DeBayes: a Bayesian Method for Debiasing Network Embeddings

TL;DR

DeBayes introduces a Bayesian approach to debias network embeddings, effectively reducing bias and improving fairness in link prediction tasks across sensitive attributes.

Contribution

It presents a novel Bayesian method that leverages biased priors to learn debiased network embeddings, addressing a gap in fairness techniques for network analysis.

Findings

Embeddings learned with DeBayes are more fair according to demographic parity.

DeBayes improves fairness metrics in link prediction tasks.

The method maintains competitive performance while reducing bias.

Abstract

As machine learning algorithms are increasingly deployed for high-impact automated decision making, ethical and increasingly also legal standards demand that they treat all individuals fairly, without discrimination based on their age, gender, race or other sensitive traits. In recent years much progress has been made on ensuring fairness and reducing bias in standard machine learning settings. Yet, for network embedding, with applications in vulnerable domains ranging from social network analysis to recommender systems, current options remain limited both in number and performance. We thus propose DeBayes: a conceptually elegant Bayesian method that is capable of learning debiased embeddings by using a biased prior. Our experiments show that these representations can then be used to perform link prediction that is significantly more fair in terms of popular metrics such as demographic parity and equalized opportunity.