{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,14]],"date-time":"2026-04-14T16:30:31Z","timestamp":1776184231216,"version":"3.50.1"},"reference-count":70,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2024,10,24]],"date-time":"2024-10-24T00:00:00Z","timestamp":1729728000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Big Data"],"abstract":"<jats:p>Link prediction is a crucial task in network analysis, but it has been shown to be prone to biased predictions, particularly when links are unfairly predicted between nodes from different sensitive groups. In this paper, we study the fair link prediction problem, which aims to ensure that the predicted link probability is independent of the sensitive attributes of the connected nodes. Existing methods typically incorporate debiasing techniques within graph embeddings to mitigate this issue. However, training on large real-world graphs is already challenging, and adding fairness constraints can further complicate the process. To overcome this challenge, we propose <jats:monospace>FairLink<\/jats:monospace>, a method that learns a fairness-enhanced graph to bypass the need for debiasing during the link predictor's training. <jats:monospace>FairLink<\/jats:monospace> maintains link prediction accuracy by ensuring that the enhanced graph follows a training trajectory similar to that of the original input graph. Meanwhile, it enhances fairness by minimizing the absolute difference in link probabilities between node pairs within the same sensitive group and those between node pairs from different sensitive groups. Our extensive experiments on multiple large-scale graphs demonstrate that <jats:monospace>FairLink<\/jats:monospace> not only promotes fairness but also often achieves link prediction accuracy comparable to baseline methods. Most importantly, the enhanced graph exhibits strong generalizability across different GNN architectures. <jats:monospace>FairLink<\/jats:monospace> is highly scalable, making it suitable for deployment in real-world large-scale graphs, where maintaining both fairness and accuracy is critical.<\/jats:p>","DOI":"10.3389\/fdata.2024.1489306","type":"journal-article","created":{"date-parts":[[2024,10,24]],"date-time":"2024-10-24T04:44:42Z","timestamp":1729745082000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":5,"title":["Promoting fairness in link prediction with graph enhancement"],"prefix":"10.3389","volume":"7","author":[{"given":"Yezi","family":"Liu","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Hanning","family":"Chen","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Mohsen","family":"Imani","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"1965","published-online":{"date-parts":[[2024,10,24]]},"reference":[{"key":"B1","first-page":"798","article-title":"\u201cLink prediction using supervised learning,\u201d","author":"Al Hasan","year":"2006","journal-title":"SDM06: workshop on link analysis, counter-terrorism and security"},{"key":"B2","doi-asserted-by":"crossref","first-page":"254","DOI":"10.1201\/9781003278290-37","article-title":"\u201cMachine bias,\u201d","volume-title":"Ethics of Data and Analytics","author":"Angwin","year":"2022"},{"key":"B3","first-page":"715","article-title":"\u201cCompositional fairness constraints for graph embeddings,\u201d","volume-title":"ICML","author":"Bose","year":"2019"},{"key":"B4","first-page":"1220","article-title":"\u201cDebayes: a bayesian method for debiasing network embeddings,\u201d","volume-title":"ICML","author":"Buyl","year":"2020"},{"key":"B5","doi-asserted-by":"crossref","first-page":"13","DOI":"10.1109\/ICDMW.2009.83","article-title":"\u201cBuilding classifiers with independency constraints,\u201d","volume-title":"2009 IEEE International Conference on Data Mining Workshops","author":"Calders","year":"2009"},{"key":"B6","article-title":"Fair mixup: fairness via interpolation","author":"Chuang","year":"2021","journal-title":"arXiv preprint arXiv:2103.06503"},{"key":"B7","doi-asserted-by":"publisher","first-page":"450","DOI":"10.1063\/1.5066450","article-title":"Mathematical model of gender bias and homophily in professional hierarchies","volume":"29","author":"Clifton","year":"2019","journal-title":"Chaos"},{"key":"B8","doi-asserted-by":"publisher","first-page":"833","DOI":"10.1109\/TKDE.2018.2849727","article-title":"A survey on network embedding","volume":"31","author":"Cui","year":"2018","journal-title":"IEEE Trans. Knowl. Data Eng"},{"key":"B9","doi-asserted-by":"publisher","DOI":"10.1145\/3551624.3555287","article-title":"\u201cFairegm: fair link prediction and recommendation via emulated graph modification,\u201d","author":"Current","year":"2022","journal-title":"Proceedings of the 2nd ACM Conference on Equity and Access in Algorithms, Mechanisms, and Optimization"},{"key":"B10","article-title":"A comprehensive survey on trustworthy graph neural networks: privacy, robustness, fairness, and explainability","author":"Dai","year":"2022","journal-title":"arXiv preprint arXiv:2204.08570"},{"key":"B11","doi-asserted-by":"publisher","DOI":"10.1145\/3447548.3467266","article-title":"\u201cIndividual fairness for graph neural networks: a ranking based approach,\u201d","author":"Dong","year":"2021","journal-title":"Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery &Data Mining"},{"key":"B12","doi-asserted-by":"publisher","first-page":"10583","DOI":"10.1109\/TKDE.2023.3265598","article-title":"Fairness in graph mining: a survey","volume":"35","author":"Dong","year":"2023","journal-title":"IEEE Trans. Knowl. Data Eng"},{"key":"B13","article-title":"A benchmark for fairness-aware graph learning","author":"Dong","year":"2024","journal-title":"arXiv preprint arXiv:2407.12112"},{"key":"B14","doi-asserted-by":"publisher","DOI":"10.1145\/2090236.2090255","article-title":"\u201cFairness through awareness,\u201d","author":"Dwork","year":"2012","journal-title":"Proceedings of the 3rd Innovations in Theoretical Computer Science Conference"},{"key":"B15","doi-asserted-by":"publisher","DOI":"10.1145\/2783258.2783311","article-title":"\u201cCertifying and removing disparate impact,\u201d","author":"Feldman","year":"2015","journal-title":"KDD"},{"key":"B16","article-title":"TF-GNN: graph neural networks in tensorflow","author":"Ferludin","year":"2022","journal-title":"arXiv preprint arXiv:2207.03522"},{"key":"B17","article-title":"\u201cSatisfying real-world goals with dataset constraints,\u201d","author":"Goh","year":"2016","journal-title":"Advances in Neural Information Processing Systems"},{"key":"B18","doi-asserted-by":"publisher","DOI":"10.1145\/2939672.2939754","article-title":"\u201cnode2vec: scalable feature learning for networks,\u201d","author":"Grover","year":"2016","journal-title":"Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining"},{"key":"B19","article-title":"Network representation learning: Consolidation and renewed bearing","author":"Gurukar","year":"2019","journal-title":"arXiv preprint arXiv:1905.00987"},{"key":"B20","article-title":"\u201cInductive representation learning on large graphs,\u201d","author":"Hamilton","year":"2017","journal-title":"NeurIPS"},{"key":"B21","article-title":"Mlpinit: embarrassingly simple gnn training acceleration with mlp initialization","author":"Han","year":"2022","journal-title":"arXiv preprint arXiv:2210.00102"},{"key":"B22","first-page":"3315","article-title":"\u201cEquality of opportunity in supervised learning,\u201d","author":"Hardt","year":"2016","journal-title":"NeurIPS"},{"key":"B23","doi-asserted-by":"publisher","DOI":"10.1007\/978-1-4419-8462-3_9","article-title":"\u201cA survey of link prediction in social networks,\u201d","author":"Hasan","year":"2011","journal-title":"Social Network Data Analytics"},{"key":"B24","article-title":"Graph-mlp: node classification without message passing in graph","author":"Hu","year":"2021","journal-title":"arXiv preprint arXiv:2106.04051"},{"key":"B25","doi-asserted-by":"publisher","DOI":"10.1145\/3534678.3539429","article-title":"\u201cCondensing graphs via one-step gradient matching,\u201d","author":"Jin","year":"","journal-title":"KDD"},{"key":"B26","article-title":"Graph condensation for graph neural networks","author":"Jin","year":"2023","journal-title":"arXiv preprint arXiv:2110.07580"},{"key":"B27","article-title":"Empowering graph representation learning with test-time graph transformation","author":"Jin","year":"","journal-title":"arXiv preprint arXiv:2210.03561"},{"key":"B28","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1109\/IC4.2009.4909197","article-title":"\u201cClassifying without discriminating,\u201d","volume-title":"2009 2nd International Conference on Computer, Control and Communication","author":"Kamiran","year":"2009"},{"key":"B29","doi-asserted-by":"publisher","DOI":"10.1145\/3394486.3403080","article-title":"\u201cInform: individual fairness on graph mining,\u201d","author":"Kang","year":"2020","journal-title":"Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery &Data Mining"},{"key":"B30","doi-asserted-by":"publisher","first-page":"11077","DOI":"10.1038\/s41598-018-29405-7","article-title":"Homophily influences ranking of minorities in social networks","volume":"8","author":"Karimi","year":"2018","journal-title":"Sci. Rep"},{"key":"B31","article-title":"Adam: a method for stochastic optimization","author":"Kingma","year":"2014","journal-title":"arXiv preprint arXiv:1412.6980"},{"key":"B32","article-title":"Semi-supervised classification with graph convolutional networks","author":"Kipf","year":"","journal-title":"CoRR, abs\/1609.02907"},{"key":"B33","article-title":"Variational graph auto-encoders","author":"Kipf","year":"","journal-title":"arXiv preprint arXiv:1611.07308"},{"key":"B34","article-title":"\u201cCounterfactual fairness,\u201d","author":"Kusner","year":"2017","journal-title":"Advances in Neural Information Processing Systems"},{"key":"B35","doi-asserted-by":"publisher","first-page":"1078","DOI":"10.1038\/s41562-019-0677-4","article-title":"Homophily and minority-group size explain perception biases in social networks","volume":"3","author":"Lee","year":"2019","journal-title":"Nat. Hum. Behav"},{"key":"B36","article-title":"\u201cLearning to discover social circles in ego networks,\u201d","author":"Leskovec","year":"2012","journal-title":"NeurIPS"},{"key":"B37","article-title":"\u201cEvaluating graph neural networks for link prediction: current pitfalls and new benchmarking,\u201d","author":"Li","year":"2024","journal-title":"NeurIPS"},{"key":"B38","article-title":"\u201cOn dyadic fairness: exploring and mitigating bias in graph connections,\u201d","author":"Li","year":"2021","journal-title":"ICLR"},{"key":"B39","doi-asserted-by":"publisher","DOI":"10.1145\/956863.956972","article-title":"\u201cThe link prediction problem for social networks,\u201d","author":"Liben-Nowell","year":"2003","journal-title":"CIKM"},{"key":"B40","doi-asserted-by":"publisher","DOI":"10.1145\/3583780.3615176","article-title":"\u201cFairgraph: automated graph debiasing with gradient matching,\u201d","author":"Liu","year":"2023","journal-title":"Proceedings of the 32nd ACM International Conference on Information and Knowledge Management"},{"key":"B41","first-page":"2037","article-title":"\u201cTinyData: joint dataset condensation with dimensionality reduction,\u201d","author":"Liu","year":"","journal-title":"2024 32nd European Signal Processing Conference (EUSIPCO)"},{"key":"B42","article-title":"Tinygraph: joint feature and node condensation for graph neural networks","author":"Liu","year":"","journal-title":"arXiv preprint arXiv:2407.08064"},{"key":"B43","doi-asserted-by":"crossref","first-page":"1604","DOI":"10.23919\/EUSIPCO58844.2023.10289852","article-title":"\u201cError detection on knowledge graphs with triple embedding,\u201d","volume-title":"2023 31st European Signal Processing Conference (EUSIPCO)","author":"Liu","year":"2023"},{"key":"B44","doi-asserted-by":"publisher","DOI":"10.1145\/3488560.3498391","article-title":"\u201cLearning fair node representations with graph counterfactual fairness,\u201d","author":"Ma","year":"2022","journal-title":"Proceedings of the Fifteenth ACM International Conference on Web Search and Data Mining"},{"key":"B45","doi-asserted-by":"crossref","first-page":"138","DOI":"10.1109\/DSAA49011.2020.00026","article-title":"\u201cBenchmarking network embedding models for link prediction: are we making progress?\u201d","volume-title":"2020 IEEE 7th International conference on data science and advanced analytics (DSAA)","author":"Mara","year":"2020"},{"key":"B46","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v34i01.5429","article-title":"\u201cBursting the filter bubble: Fairness-aware network link prediction,\u201d","author":"Masrour","year":"2020","journal-title":"AAAI"},{"key":"B47","doi-asserted-by":"crossref","first-page":"437","DOI":"10.1007\/978-3-642-23783-6_28","article-title":"\u201cLink prediction via matrix factorization,\u201d","volume-title":"Machine Learning and Knowledge Discovery in Databases: European Conference, ECML PKDD 2011, Athens, Greece, September 5-9, 2011, Proceedings, Part II 22","author":"Menon","year":"2011"},{"key":"B48","doi-asserted-by":"publisher","DOI":"10.1145\/1183614.1183678","article-title":"\u201cOn the structural properties of massive telecom call graphs: findings and implications,\u201d","author":"Nanavati","year":"2006","journal-title":"CIKM"},{"key":"B49","doi-asserted-by":"publisher","first-page":"025102","DOI":"10.1103\/PhysRevE.64.025102","article-title":"Clustering and preferential attachment in growing networks","volume":"64","author":"Newman","year":"2001","journal-title":"Phys. Rev. E"},{"key":"B50","doi-asserted-by":"publisher","DOI":"10.1145\/3637528.3671616","article-title":"\u201cAddressing shortcomings in fair graph learning datasets: Towards a new benchmark,\u201d","author":"Qian","year":"2024","journal-title":"Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining"},{"key":"B51","doi-asserted-by":"publisher","DOI":"10.24963\/ijcai.2019\/456","article-title":"\u201cFairwalk: towards fair graph embedding,\u201d","author":"Rahman","year":"2019","journal-title":"International Joint Conference on Artificial Intelligence"},{"key":"B52","doi-asserted-by":"publisher","DOI":"10.1145\/1401890.1402008","article-title":"\u201cArnetminer: extraction and mining of academic social networks,\u201d","author":"Tang","year":"2008","journal-title":"Proceedings of the 14th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining"},{"key":"B53","first-page":"2071","article-title":"\u201cComplex embeddings for simple link prediction,\u201d","volume-title":"ICML","author":"Trouillon","year":"2016"},{"key":"B54","doi-asserted-by":"publisher","first-page":"3815","DOI":"10.1145\/3442381.3450065","article-title":"\u201cFairness-aware pagerank,\u201d","volume":"2021","author":"Tsioutsiouliklis","year":"2021","journal-title":"Proceedings of the Web Conference"},{"key":"B55","article-title":"\u201cGraph attention networks,\u201d","author":"Velickovic","year":"2018","journal-title":"ICLR"},{"key":"B56","article-title":"Equivariant and stable positional encoding for more powerful graph neural networks","author":"Wang","year":"2022","journal-title":"arXiv preprint arXiv:2203.00199"},{"key":"B57","article-title":"Explaining dynamic graph neural networks via relevance back-propagation","author":"Xie","year":"2022","journal-title":"arXiv preprint arXiv:2207.11175"},{"key":"B58","first-page":"40","article-title":"\u201cRevisiting semi-supervised learning with graph embeddings,\u201d","volume-title":"International Conference on Machine Learning","author":"Yang","year":"2016"},{"key":"B59","doi-asserted-by":"publisher","DOI":"10.1145\/3038912.3052660","article-title":"\u201cFairness beyond disparate treatment &disparate impact: Learning classification without disparate mistreatment,\u201d","author":"Zafar","year":"2017","journal-title":"Proceedings of the 26th International Conference on World Wide Web"},{"key":"B60","article-title":"Fairness constraints: mechanisms for fair classification","author":"Zafar","year":"2015","journal-title":"arXiv preprint arXiv:1507.05259"},{"key":"B61","first-page":"325","article-title":"\u201cLearning fair representations,\u201d","volume-title":"ICML","author":"Zemel","year":"2013"},{"key":"B62","article-title":"Data-centric AI: perspectives and challenges","author":"Zha","year":"","journal-title":"arXiv preprint arXiv:2301.04819"},{"key":"B63","article-title":"Data-centric artificial intelligence: a survey","author":"Zha","year":"","journal-title":"arXiv preprint arXiv:2303.10158"},{"key":"B64","first-page":"9061","article-title":"Labeling trick: a theory of using graph neural networks for multi-node representation learning","volume":"34","author":"Zhang","year":"2021","journal-title":"NeurIPS"},{"key":"B65","doi-asserted-by":"publisher","DOI":"10.1145\/3511808.3557264","article-title":"\u201cContrastive knowledge graph error detection,\u201d","author":"Zhang","year":"2022","journal-title":"Proceedings of the 31st ACM International Conference on Information &Knowledge Management"},{"key":"B66","article-title":"Graph-less neural networks: Teaching old mlps new tricks via distillation","author":"Zhang","year":"2021","journal-title":"arXiv preprint arXiv:2110.08727"},{"key":"B67","article-title":"Dataset condensation with gradient matching","author":"Zhao","year":"2020","journal-title":"arXiv preprint arXiv:2006.05929"},{"key":"B68","doi-asserted-by":"publisher","first-page":"623","DOI":"10.1140\/epjb\/e2009-00335-8","article-title":"Predicting missing links via local information","volume":"71","author":"Zhou","year":"2009","journal-title":"Eur. Phys. J. B"},{"key":"B69","first-page":"29476","article-title":"\u201cNeural bellman-ford networks: a general graph neural network framework for link prediction,\u201d","author":"Zhu","year":"2021","journal-title":"Advances in Neural Information Processing Systems"},{"key":"B70","doi-asserted-by":"publisher","DOI":"10.1155\/2022\/7438464","article-title":"\u201cCounterfactual fairness with partially known causal graph,\u201d","author":"Zuo","year":"2022","journal-title":"Advances in Neural Information Processing Systems"}],"container-title":["Frontiers in Big Data"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fdata.2024.1489306\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,10,24]],"date-time":"2024-10-24T04:44:53Z","timestamp":1729745093000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fdata.2024.1489306\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,10,24]]},"references-count":70,"alternative-id":["10.3389\/fdata.2024.1489306"],"URL":"https:\/\/doi.org\/10.3389\/fdata.2024.1489306","relation":{},"ISSN":["2624-909X"],"issn-type":[{"value":"2624-909X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2024,10,24]]},"article-number":"1489306"}}