Networked Inequality: Preferential Attachment Bias in Graph Neural Network Link Prediction

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Subramonian, Arjun, Sagun, Levent, Sun, Yizhou
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910472191606784
author Subramonian, Arjun
Sagun, Levent
Sun, Yizhou
author_facet Subramonian, Arjun
Sagun, Levent
Sun, Yizhou
contents Graph neural network (GNN) link prediction is increasingly deployed in citation, collaboration, and online social networks to recommend academic literature, collaborators, and friends. While prior research has investigated the dyadic fairness of GNN link prediction, the within-group (e.g., queer women) fairness and "rich get richer" dynamics of link prediction remain underexplored. However, these aspects have significant consequences for degree and power imbalances in networks. In this paper, we shed light on how degree bias in networks affects Graph Convolutional Network (GCN) link prediction. In particular, we theoretically uncover that GCNs with a symmetric normalized graph filter have a within-group preferential attachment bias. We validate our theoretical analysis on real-world citation, collaboration, and online social networks. We further bridge GCN's preferential attachment bias with unfairness in link prediction and propose a new within-group fairness metric. This metric quantifies disparities in link prediction scores within social groups, towards combating the amplification of degree and power disparities. Finally, we propose a simple training-time strategy to alleviate within-group unfairness, and we show that it is effective on citation, social, and credit networks.
format Preprint
id arxiv_https___arxiv_org_abs_2309_17417
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Networked Inequality: Preferential Attachment Bias in Graph Neural Network Link Prediction
Subramonian, Arjun
Sagun, Levent
Sun, Yizhou
Machine Learning
Computers and Society
Social and Information Networks
Graph neural network (GNN) link prediction is increasingly deployed in citation, collaboration, and online social networks to recommend academic literature, collaborators, and friends. While prior research has investigated the dyadic fairness of GNN link prediction, the within-group (e.g., queer women) fairness and "rich get richer" dynamics of link prediction remain underexplored. However, these aspects have significant consequences for degree and power imbalances in networks. In this paper, we shed light on how degree bias in networks affects Graph Convolutional Network (GCN) link prediction. In particular, we theoretically uncover that GCNs with a symmetric normalized graph filter have a within-group preferential attachment bias. We validate our theoretical analysis on real-world citation, collaboration, and online social networks. We further bridge GCN's preferential attachment bias with unfairness in link prediction and propose a new within-group fairness metric. This metric quantifies disparities in link prediction scores within social groups, towards combating the amplification of degree and power disparities. Finally, we propose a simple training-time strategy to alleviate within-group unfairness, and we show that it is effective on citation, social, and credit networks.
title Networked Inequality: Preferential Attachment Bias in Graph Neural Network Link Prediction
topic Machine Learning
Computers and Society
Social and Information Networks
url https://arxiv.org/abs/2309.17417