Details
| Original language | English |
|---|---|
| Number of pages | 15 |
| Journal | IEEE Transactions on Computational Social Systems |
| Early online date | 18 Nov 2024 |
| Publication status | E-pub ahead of print - 18 Nov 2024 |
Abstract
In this article, we introduce FairLPG, a framework for ensuring fairness for the task of link prediction in graphs with multiple sensitive attributes. In the context of link prediction in graphs, the fairness notions of demographic parity and equalized odds try to ensure equal average linking probability and true positive rates across different demographic groups consisting of various node pairs. Existing methods for achieving fairness in link prediction only consider a single sensitive attribute, which makes them unsuited for applications where multiple sensitive attributes need to be accounted for. Additionally, considering multiple sensitive attributes in the context of link prediction leads to overlapping and intersectional groups, which further complicates designing such a framework. The proposed framework FairLPG assumes that the link prediction model generates a prediction score for each node pair to form an edge, and formulates a convex optimization problem that minimizes the squared Euclidean distance between the original prediction scores and transformed scores, subject to the fairness constraints. The transformed scores are then utilized for fair link prediction. To the best of our knowledge, this work is the first to handle the case of intersectional sensitive groups in the graph setting. To demonstrate its effectiveness, we deploy FairLPG on several real-world datasets and graph neural network based link prediction models. It either outperforms or performs competitively with existing methods both in terms of fairness and prediction accuracy across all the datasets and link prediction models at the same time being computationally more efficient.
Keywords
- Convex optimization, fairness in graphs, link prediction, score transformation
ASJC Scopus subject areas
- Mathematics(all)
- Modelling and Simulation
- Social Sciences(all)
- Social Sciences (miscellaneous)
- Computer Science(all)
- Human-Computer Interaction
Cite this
- Standard
- Harvard
- Apa
- Vancouver
- BibTeX
- RIS
In: IEEE Transactions on Computational Social Systems, 18.11.2024.
Research output: Contribution to journal › Article › Research › peer review
}
TY - JOUR
T1 - Fair Link Prediction With Overlapping Groups
AU - Pal, Manjish
AU - Sikdar, Sandipan
AU - Ganguly, Niloy
N1 - Publisher Copyright: © 2014 IEEE.
PY - 2024/11/18
Y1 - 2024/11/18
N2 - In this article, we introduce FairLPG, a framework for ensuring fairness for the task of link prediction in graphs with multiple sensitive attributes. In the context of link prediction in graphs, the fairness notions of demographic parity and equalized odds try to ensure equal average linking probability and true positive rates across different demographic groups consisting of various node pairs. Existing methods for achieving fairness in link prediction only consider a single sensitive attribute, which makes them unsuited for applications where multiple sensitive attributes need to be accounted for. Additionally, considering multiple sensitive attributes in the context of link prediction leads to overlapping and intersectional groups, which further complicates designing such a framework. The proposed framework FairLPG assumes that the link prediction model generates a prediction score for each node pair to form an edge, and formulates a convex optimization problem that minimizes the squared Euclidean distance between the original prediction scores and transformed scores, subject to the fairness constraints. The transformed scores are then utilized for fair link prediction. To the best of our knowledge, this work is the first to handle the case of intersectional sensitive groups in the graph setting. To demonstrate its effectiveness, we deploy FairLPG on several real-world datasets and graph neural network based link prediction models. It either outperforms or performs competitively with existing methods both in terms of fairness and prediction accuracy across all the datasets and link prediction models at the same time being computationally more efficient.
AB - In this article, we introduce FairLPG, a framework for ensuring fairness for the task of link prediction in graphs with multiple sensitive attributes. In the context of link prediction in graphs, the fairness notions of demographic parity and equalized odds try to ensure equal average linking probability and true positive rates across different demographic groups consisting of various node pairs. Existing methods for achieving fairness in link prediction only consider a single sensitive attribute, which makes them unsuited for applications where multiple sensitive attributes need to be accounted for. Additionally, considering multiple sensitive attributes in the context of link prediction leads to overlapping and intersectional groups, which further complicates designing such a framework. The proposed framework FairLPG assumes that the link prediction model generates a prediction score for each node pair to form an edge, and formulates a convex optimization problem that minimizes the squared Euclidean distance between the original prediction scores and transformed scores, subject to the fairness constraints. The transformed scores are then utilized for fair link prediction. To the best of our knowledge, this work is the first to handle the case of intersectional sensitive groups in the graph setting. To demonstrate its effectiveness, we deploy FairLPG on several real-world datasets and graph neural network based link prediction models. It either outperforms or performs competitively with existing methods both in terms of fairness and prediction accuracy across all the datasets and link prediction models at the same time being computationally more efficient.
KW - Convex optimization
KW - fairness in graphs
KW - link prediction
KW - score transformation
UR - http://www.scopus.com/inward/record.url?scp=85210395854&partnerID=8YFLogxK
U2 - 10.21203/rs.3.rs-3743069/v1
DO - 10.21203/rs.3.rs-3743069/v1
M3 - Article
AN - SCOPUS:85210395854
JO - IEEE Transactions on Computational Social Systems
JF - IEEE Transactions on Computational Social Systems
ER -