Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

FairSNA: Algorithmic Fairness in Social Network Analysis

Akrati Saxena, George Fletcher, Mykola Pechenizkiy

TL;DR

This paper addresses the gap in fairness within social network analysis by highlighting structural biases and proposing a comprehensive FairSNA framework. It introduces a taxonomy (group vs. individual, feature-aware vs. feature-blind, pre/in/post-processing) and surveys state-of-the-art fairness constraints and methods across key SNA tasks, including link prediction, centrality ranking, influence maximization, influence blocking, and community detection, plus other topics. It discusses datasets and synthetic network models, underlining gaps such as the predominance of in-processing and feature-aware approaches and the need for broader definitions and evaluation across tasks. The work maps open directions and calls for future research to bridge fairness and SNA in dynamic, diverse networks with standardized metrics and interdisciplinary collaboration.

Abstract

In recent years, designing fairness-aware methods has received much attention in various domains, including machine learning, natural language processing, and information retrieval. However, understanding structural bias and inequalities in social networks and designing fairness-aware methods for various research problems in social network analysis (SNA) have not received much attention. In this work, we highlight how the structural bias of social networks impacts the fairness of different SNA methods. We further discuss fairness aspects that should be considered while proposing network structure-based solutions for different SNA problems, such as link prediction, influence maximization, centrality ranking, and community detection. This paper clearly highlights that very few works have considered fairness and bias while proposing solutions; even these works are mainly focused on some research topics, such as link prediction, influence maximization, and PageRank. However, fairness has not yet been addressed for other research topics, such as influence blocking and community detection. We review state-of-the-art for different research topics in SNA, including the considered fairness constraints, their limitations, and our vision. This paper also covers evaluation metrics, available datasets, and synthetic network generating models used in such studies. Finally, we highlight various open research directions that require researchers' attention to bridge the gap between fairness and SNA.

FairSNA: Algorithmic Fairness in Social Network Analysis

TL;DR

This paper addresses the gap in fairness within social network analysis by highlighting structural biases and proposing a comprehensive FairSNA framework. It introduces a taxonomy (group vs. individual, feature-aware vs. feature-blind, pre/in/post-processing) and surveys state-of-the-art fairness constraints and methods across key SNA tasks, including link prediction, centrality ranking, influence maximization, influence blocking, and community detection, plus other topics. It discusses datasets and synthetic network models, underlining gaps such as the predominance of in-processing and feature-aware approaches and the need for broader definitions and evaluation across tasks. The work maps open directions and calls for future research to bridge fairness and SNA in dynamic, diverse networks with standardized metrics and interdisciplinary collaboration.

Abstract

In recent years, designing fairness-aware methods has received much attention in various domains, including machine learning, natural language processing, and information retrieval. However, understanding structural bias and inequalities in social networks and designing fairness-aware methods for various research problems in social network analysis (SNA) have not received much attention. In this work, we highlight how the structural bias of social networks impacts the fairness of different SNA methods. We further discuss fairness aspects that should be considered while proposing network structure-based solutions for different SNA problems, such as link prediction, influence maximization, centrality ranking, and community detection. This paper clearly highlights that very few works have considered fairness and bias while proposing solutions; even these works are mainly focused on some research topics, such as link prediction, influence maximization, and PageRank. However, fairness has not yet been addressed for other research topics, such as influence blocking and community detection. We review state-of-the-art for different research topics in SNA, including the considered fairness constraints, their limitations, and our vision. This paper also covers evaluation metrics, available datasets, and synthetic network generating models used in such studies. Finally, we highlight various open research directions that require researchers' attention to bridge the gap between fairness and SNA.
Paper Structure (30 sections, 2 equations, 3 figures, 3 tables)

This paper contains 30 sections, 2 equations, 3 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Taxonomy of Fairness-aware SNA (FairSNA)
  3. FairSNA: Fairness Constraints, State-of-the-art, and Future Directions
  4. Fair Link Prediction
  5. Fairness Definition.
  6. State-of-the-Art.
  7. Future Directions.
  8. Fair Centrality Ranking
  9. Fairness Definitions.
  10. State-of-the-Art and Future Directions.
  11. Fair Influence Maximization
  12. Fairness Definitions.
  13. State-of-the-Art.
  14. Future Directions.
  15. Fair Influence Blocking Maximization
  16. ...and 15 more sections

Figures (3)

  • Figure 1: (a). Dutch School Social network having two groups where blue nodes represent girls and pink nodes represent boys, (b). Dashed lines represent around 10% u.a.r. removed links from the network, (c). and (d). Values corresponding to dashed lines show the Jaccard coefficient and Adamic Adar scores for predicting the missing links, respectively.
  • Figure 2: Nodes are ranked based on four different centrality rankings, and nodes' size corresponds to their centrality rank, where (a) Degree Ranking and (b) PageRank are fairness-oblivious centrality ranking, and (c) Locally Fair Pagerank and (d) Fairness sensitive PageRank are fairness-aware PageRank from tsioutsiouliklis2021fairness.
  • Figure 3: Influence propagation spread using Independent Cascade model where top-2 seed nodes are chosen using four different methods, including degree, pagerank, CELF, and fairness sensitive pagerank. Seed nodes are shown in green color and influenced nodes are shown in grey color. A node's size is proportional to the probability of it getting infected.