Adversarial Learning for Debiasing Knowledge Graph Embeddings

TL;DR

This paper investigates biases in knowledge graph embeddings, particularly popularity and gender biases, and introduces FAN, an adversarial framework to mitigate sensitive attribute information in embeddings.

Contribution

It identifies biases in KG embeddings and proposes a novel adversarial filtering framework, FAN, to reduce sensitive attribute leakage, advancing debiasing methods for knowledge graph representations.

Findings

Prediction accuracy negatively correlates with node degree in node2vec embeddings.

In knowledge graph embeddings, prediction accuracy positively correlates with node degree.

FAN effectively filters out sensitive attribute information from KG embeddings.

Abstract

Knowledge Graphs (KG) are gaining increasing attention in both academia and industry. Despite their diverse benefits, recent research have identified social and cultural biases embedded in the representations learned from KGs. Such biases can have detrimental consequences on different population and minority groups as applications of KG begin to intersect and interact with social spheres. This paper aims at identifying and mitigating such biases in Knowledge Graph (KG) embeddings. As a first step, we explore popularity bias -- the relationship between node popularity and link prediction accuracy. In case of node2vec graph embeddings, we find that prediction accuracy of the embedding is negatively correlated with the degree of the node. However, in case of knowledge-graph embeddings (KGE), we observe an opposite trend. As a second step, we explore gender bias in KGE, and a careful examination of popular KGE algorithms suggest that sensitive attribute like the gender of a person can be predicted from the embedding. This implies that such biases in popular KGs is captured by the structural properties of the embedding. As a preliminary solution to debiasing KGs, we introduce a novel framework to filter out the sensitive attribute information from the KG embeddings, which we call FAN (Filtering Adversarial Network). We also suggest the applicability of FAN for debiasing other network embeddings which could be explored in future work.