On generalized degree fairness in graph neural networks

Conventional graph neural networks (GNNs) are often confronted with fairness issues that may stem from their input, including node attributes and neighbors surrounding a node. While several recent approaches have been proposed to eliminate the bias rooted in sensitive attributes, they ignore the oth...

Full description

Saved in:
Bibliographic Details
Main Authors: LIU, Zemin, NGUYEN, Trung Kien, FANG, Yuan
Format: text
Language:English
Published: Institutional Knowledge at Singapore Management University 2023
Subjects:
Online Access:https://ink.library.smu.edu.sg/sis_research/8189
https://ink.library.smu.edu.sg/context/sis_research/article/9192/viewcontent/AAAI23_DegFairGNN.pdf
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Singapore Management University
Language: English
id sg-smu-ink.sis_research-9192
record_format dspace
spelling sg-smu-ink.sis_research-91922023-09-26T10:26:25Z On generalized degree fairness in graph neural networks LIU, Zemin NGUYEN, Trung Kien FANG, Yuan Conventional graph neural networks (GNNs) are often confronted with fairness issues that may stem from their input, including node attributes and neighbors surrounding a node. While several recent approaches have been proposed to eliminate the bias rooted in sensitive attributes, they ignore the other key input of GNNs, namely the neighbors of a node, which can introduce bias since GNNs hinge on neighborhood structures to generate node representations. In particular, the varying neighborhood structures across nodes, manifesting themselves in drastically different node degrees, give rise to the diverse behaviors of nodes and biased outcomes. In this paper, we first define and generalize the degree bias using a generalized definition of node degree as a manifestation and quantification of different multi-hop structures around different nodes. To address the bias in the context of node classification, we propose a novel GNN framework called Generalized Degree Fairness-centric Graph Neural Network (DegFairGNN). Specifically, in each GNN layer, we employ a learnable debiasing function to generate debiasing contexts, which modulate the layer-wise neighborhood aggregation to eliminate the degree bias originating from the diverse degrees among nodes. Extensive experiments on three benchmark datasets demonstrate the effectiveness of our model on both accuracy and fairness metrics. 2023-02-01T08:00:00Z text application/pdf https://ink.library.smu.edu.sg/sis_research/8189 info:doi/10.48550/arXiv.2302.03881 https://ink.library.smu.edu.sg/context/sis_research/article/9192/viewcontent/AAAI23_DegFairGNN.pdf http://creativecommons.org/licenses/by-nc-nd/4.0/ Research Collection School Of Computing and Information Systems eng Institutional Knowledge at Singapore Management University De-biasing Graph neural networks Layer-wise Multi-hops Neighborhood structure Network frameworks Node attribute Node degree Sensitive attribute Databases and Information Systems
institution Singapore Management University
building SMU Libraries
continent Asia
country Singapore
Singapore
content_provider SMU Libraries
collection InK@SMU
language English
topic De-biasing
Graph neural networks
Layer-wise
Multi-hops
Neighborhood structure
Network frameworks
Node attribute
Node degree
Sensitive attribute
Databases and Information Systems
spellingShingle De-biasing
Graph neural networks
Layer-wise
Multi-hops
Neighborhood structure
Network frameworks
Node attribute
Node degree
Sensitive attribute
Databases and Information Systems
LIU, Zemin
NGUYEN, Trung Kien
FANG, Yuan
On generalized degree fairness in graph neural networks
description Conventional graph neural networks (GNNs) are often confronted with fairness issues that may stem from their input, including node attributes and neighbors surrounding a node. While several recent approaches have been proposed to eliminate the bias rooted in sensitive attributes, they ignore the other key input of GNNs, namely the neighbors of a node, which can introduce bias since GNNs hinge on neighborhood structures to generate node representations. In particular, the varying neighborhood structures across nodes, manifesting themselves in drastically different node degrees, give rise to the diverse behaviors of nodes and biased outcomes. In this paper, we first define and generalize the degree bias using a generalized definition of node degree as a manifestation and quantification of different multi-hop structures around different nodes. To address the bias in the context of node classification, we propose a novel GNN framework called Generalized Degree Fairness-centric Graph Neural Network (DegFairGNN). Specifically, in each GNN layer, we employ a learnable debiasing function to generate debiasing contexts, which modulate the layer-wise neighborhood aggregation to eliminate the degree bias originating from the diverse degrees among nodes. Extensive experiments on three benchmark datasets demonstrate the effectiveness of our model on both accuracy and fairness metrics.
format text
author LIU, Zemin
NGUYEN, Trung Kien
FANG, Yuan
author_facet LIU, Zemin
NGUYEN, Trung Kien
FANG, Yuan
author_sort LIU, Zemin
title On generalized degree fairness in graph neural networks
title_short On generalized degree fairness in graph neural networks
title_full On generalized degree fairness in graph neural networks
title_fullStr On generalized degree fairness in graph neural networks
title_full_unstemmed On generalized degree fairness in graph neural networks
title_sort on generalized degree fairness in graph neural networks
publisher Institutional Knowledge at Singapore Management University
publishDate 2023
url https://ink.library.smu.edu.sg/sis_research/8189
https://ink.library.smu.edu.sg/context/sis_research/article/9192/viewcontent/AAAI23_DegFairGNN.pdf
_version_ 1779157219646898176