{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,16]],"date-time":"2026-07-16T18:04:06Z","timestamp":1784225046988,"version":"3.55.0"},"reference-count":37,"publisher":"Public Library of Science (PLoS)","issue":"7","license":[{"start":{"date-parts":[[2026,7,16]],"date-time":"2026-07-16T00:00:00Z","timestamp":1784160000000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/100020595","name":"National Science and Technology Council","doi-asserted-by":"publisher","award":["114-2124-M-A49-009-, 114-2622-E-A49-007-, 114-2321-B-182-005-, 113-2923-M-A49-004-MY3, 113-2118-M-A49-007-MY2"],"award-info":[{"award-number":["114-2124-M-A49-009-, 114-2622-E-A49-007-, 114-2321-B-182-005-, 113-2923-M-A49-004-MY3, 113-2118-M-A49-007-MY2"]}],"id":[{"id":"10.13039\/100020595","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100020595","name":"National Science and Technology Council","doi-asserted-by":"publisher","award":["114-2124-M-037-002-"],"award-info":[{"award-number":["114-2124-M-037-002-"]}],"id":[{"id":"10.13039\/100020595","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["www.plosdigitalhealth.org"],"crossmark-restriction":false},"short-container-title":["PLOS Digit Health"],"abstract":"<jats:p>Performance disparities in AI systems can manifest across sensitive groups or across data sources, especially when training data are collected from specific populations. In this work, we propose a feature-selection-based method that improves test-fairness while preserving prediction performance. Built on deep learning models, the proposed approach adopts the influence score (I-score), a statistical measure that captures interaction effects among multiple features. We identify features strongly associated with dataset membership by training an auxiliary model to predict dataset origin and applying I-score-based subset selection; these dataset-associated features are then excluded (masked) from the original prediction model for follow-up inference. We conduct experiments on two skin lesion datasets, ISIC 2019 and ASAN, collected from different populations. The empirical results show that the resulting fair I-score model can maintain high classification performance for skin lesion prediction while reducing cross-dataset subgroup performance disparity under our test-fairness evaluation setting.<\/jats:p>","DOI":"10.1371\/journal.pdig.0001513","type":"journal-article","created":{"date-parts":[[2026,7,16]],"date-time":"2026-07-16T17:23:10Z","timestamp":1784222590000},"page":"e0001513","update-policy":"https:\/\/doi.org\/10.1371\/journal.pdig.corrections_policy","source":"Crossref","is-referenced-by-count":0,"title":["Test-fairness deep learning with influence score"],"prefix":"10.1371","volume":"5","author":[{"given":"Jacky Chung-Hao","family":"Wu","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Chang-Yu","family":"Shih","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Nien-Chen","family":"Wu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Wei-Wen","family":"Chen","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4392-3361","authenticated-orcid":true,"given":"Henry Horng-Shing","family":"Lu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Shaw-Hwa","family":"Lo","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"340","published-online":{"date-parts":[[2026,7,16]]},"reference":[{"issue":"11","key":"pdig.0001513.ref001","doi-asserted-by":"crossref","first-page":"2232","DOI":"10.1038\/s41591-022-01987-w","article-title":"Tackling bias in AI health datasets through the STANDING Together initiative","volume":"28","author":"S Ganapathi","year":"2022","journal-title":"Nat Med"},{"key":"pdig.0001513.ref002","doi-asserted-by":"crossref","first-page":"141","DOI":"10.1146\/annurev-statistics-042720-125902","article-title":"Algorithmic fairness: Choices, assumptions, and definitions","volume":"8","author":"S Mitchell","year":"2021","journal-title":"Annual Review of Statistics and Its Application"},{"issue":"1","key":"pdig.0001513.ref003","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1038\/s41598-022-07939-1","article-title":"A clarification of the nuances in the fairness metrics landscape","volume":"12","author":"A Castelnovo","year":"2022","journal-title":"Sci Rep"},{"issue":"6","key":"pdig.0001513.ref004","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1145\/3457607","article-title":"A survey on bias and fairness in machine learning","volume":"54","author":"N Mehrabi","year":"2021","journal-title":"ACM Comput Surv"},{"key":"pdig.0001513.ref005","unstructured":"Simonyan K, Zisserman A. Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556. 2014."},{"issue":"1","key":"pdig.0001513.ref006","doi-asserted-by":"crossref","first-page":"286","DOI":"10.1038\/s41746-024-01276-5","article-title":"Addressing fairness issues in deep learning-based medical image analysis: a systematic review","volume":"7","author":"Z Xu","year":"2024","journal-title":"NPJ Digit Med"},{"issue":"23","key":"pdig.0001513.ref007","doi-asserted-by":"crossref","DOI":"10.1073\/pnas.1919012117","article-title":"Gender imbalance in medical imaging datasets produces biased classifiers for computer-aided diagnosis","volume":"117","author":"AJ Larrazabal","year":"2020","journal-title":"Proc Natl Acad Sci"},{"issue":"12","key":"pdig.0001513.ref008","doi-asserted-by":"crossref","first-page":"2176","DOI":"10.1038\/s41591-021-01595-0","article-title":"Underdiagnosis bias of artificial intelligence algorithms applied to chest radiographs in under-served patient populations","volume":"27","author":"L Seyyed-Kalantari","year":"2021","journal-title":"Nat Med"},{"issue":"3","key":"pdig.0001513.ref009","doi-asserted-by":"crossref","first-page":"2917","DOI":"10.1007\/s11227-023-05567-8","article-title":"Forecasting the future popularity of the anti-vax narrative on Twitter with machine learning","volume":"80","author":"I Biri","year":"2023","journal-title":"J Supercomput"},{"issue":"10","key":"pdig.0001513.ref010","doi-asserted-by":"crossref","first-page":"1772","DOI":"10.3390\/diagnostics13101772","article-title":"Detection of Monkeypox Disease from Human Skin Images with a Hybrid Deep Learning Model","volume":"13","author":"F Uysal","year":"2023","journal-title":"Diagnostics (Basel)"},{"issue":"4","key":"pdig.0001513.ref011","doi-asserted-by":"crossref","DOI":"10.1214\/09-AOAS265","article-title":"Discovering influential variables: A method of partitions","volume":"3","author":"H Chernoff","year":"2009","journal-title":"Ann Appl Stat"},{"issue":"50","key":"pdig.0001513.ref012","doi-asserted-by":"crossref","first-page":"14277","DOI":"10.1073\/pnas.1616647113","article-title":"Framework for making better predictions by directly estimating variables\u2019 predictivity","volume":"113","author":"A Lo","year":"2016","journal-title":"Proc Natl Acad Sci U S A"},{"issue":"21","key":"pdig.0001513.ref013","doi-asserted-by":"crossref","first-page":"2834","DOI":"10.1093\/bioinformatics\/bts531","article-title":"Interaction-based feature selection and classification for high-dimensional biological data","volume":"28","author":"H Wang","year":"2012","journal-title":"Bioinformatics"},{"key":"pdig.0001513.ref014","doi-asserted-by":"crossref","unstructured":"Selvaraju RR, Cogswell M, Das A, Vedantam R, Parikh D, Batra D. Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization. In: 2017 IEEE International Conference on Computer Vision (ICCV). 2017. p. 618\u201326.","DOI":"10.1109\/ICCV.2017.74"},{"key":"pdig.0001513.ref015","doi-asserted-by":"crossref","unstructured":"Zhou B, Khosla A, Lapedriza A, Oliva A, Torralba A. Learning deep features for discriminative localization. In: Proceedings of the IEEE conference on computer vision and pattern recognition. 2016. p. 2921\u20139.","DOI":"10.1109\/CVPR.2016.319"},{"key":"pdig.0001513.ref016","doi-asserted-by":"crossref","unstructured":"Verma S, Rubin J. Fairness definitions explained. In: 2018 IEEE\/ACM International Workshop on Software Fairness (FairWare). IEEE; 2018. p. 1\u20137.","DOI":"10.1145\/3194770.3194776"},{"issue":"2","key":"pdig.0001513.ref017","doi-asserted-by":"crossref","first-page":"153","DOI":"10.1089\/big.2016.0047","article-title":"Fair Prediction with Disparate Impact: A Study of Bias in Recidivism Prediction Instruments","volume":"5","author":"A Chouldechova","year":"2017","journal-title":"Big Data"},{"key":"pdig.0001513.ref018","unstructured":"Kleinberg JM, Mullainathan S, Raghavan M. Inherent trade-offs in the fair determination of risk scores. In: 8th Innovations in Theoretical Computer Science Conference, ITCS 2017, January 9-11, 2017, Berkeley, CA, USA. 2017."},{"key":"pdig.0001513.ref019","unstructured":"Lundberg SM, Lee S-I. A unified approach to interpreting model predictions. In: NIPS\u201917. Red Hook (NY): Curran Associates Inc; 2017. p. 4768\u201377."},{"key":"pdig.0001513.ref020","unstructured":"Lin M, Chen Q, Yan S. Network in network. arXiv preprint. arXiv:1312.4400. 2013."},{"key":"pdig.0001513.ref021","article-title":"Equality of opportunity in supervised learning","volume":"29","author":"M Hardt","year":"2016","journal-title":"Adv Neural Inform Process Syst"},{"key":"pdig.0001513.ref022","article-title":"Counterfactual fairness","volume":"30","author":"MJ Kusner","year":"2017","journal-title":"Adv Neural Inform Process Syst"},{"key":"pdig.0001513.ref023","unstructured":"Zong Y, Yang Y, Hospedales T. Medfair: benchmarking fairness for medical imaging. arXiv preprint arXiv:2210.0172. 2022."},{"key":"pdig.0001513.ref024","doi-asserted-by":"crossref","unstructured":"Zhang BH, Lemoine B, Mitchell M. Mitigating Unwanted Biases with Adversarial Learning. In: Proceedings of the 2018 AAAI\/ACM Conference on AI, Ethics, and Society. 2018. p. 335\u201340.","DOI":"10.1145\/3278721.3278779"},{"key":"pdig.0001513.ref025","unstructured":"Madras D, Creager E, Pitassi T, Zemel R. Learning adversarially fair and transferable representations. In: International Conference on Machine Learning. 2018. p. 3384\u201393."},{"issue":"1","key":"pdig.0001513.ref026","doi-asserted-by":"crossref","first-page":"55","DOI":"10.1038\/s41746-023-00805-y","article-title":"An adversarial training framework for mitigating algorithmic biases in clinical machine learning","volume":"6","author":"J Yang","year":"2023","journal-title":"NPJ Digit Med"},{"key":"pdig.0001513.ref027","doi-asserted-by":"crossref","unstructured":"Du S, Hers B, Bayasi N, Hamarneh G, Garbi R. Fairdisco: Fairer ai in dermatology via disentanglement contrastive learning. In: European Conference on Computer Vision. Springer; 2022. p. 185\u2013202.","DOI":"10.1007\/978-3-031-25069-9_13"},{"issue":"1","key":"pdig.0001513.ref028","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1007\/s10115-011-0463-8","article-title":"Data preprocessing techniques for classification without discrimination","volume":"33","author":"F Kamiran","year":"2012","journal-title":"Knowl Inf Syst"},{"key":"pdig.0001513.ref029","unstructured":"De Sousa Ribeiro F, Xia T, Monteiro M, Pawlowski N, Glocker B. High fidelity image counterfactuals with probabilistic causal models. In: Proceedings of the 40th International Conference on Machine Learning. 2023. p. 7390\u2013425."},{"key":"pdig.0001513.ref030","doi-asserted-by":"crossref","unstructured":"Kumar A, Fathi N, Mehta R, Nichyporuk B, Falet J-PR, Tsaftaris S, et al. Debiasing counterfactuals in the presence of spurious correlations. In: Workshop on Clinical Image-Based Procedures. Springer; 2023. p. 276\u201386.","DOI":"10.1007\/978-3-031-45249-9_27"},{"issue":"1","key":"pdig.0001513.ref031","article-title":"Characteristics of publicly available skin cancer image datasets: a systematic review","volume":"4","author":"D Wen","year":"2022","journal-title":"Lancet Digit Health"},{"issue":"1","key":"pdig.0001513.ref032","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1038\/sdata.2018.161","article-title":"The ham10000 dataset, a large collection of multi-source dermatoscopic images of common pigmented skin lesions","volume":"5","author":"P Tschandl","year":"2018","journal-title":"Sci Data"},{"key":"pdig.0001513.ref033","doi-asserted-by":"crossref","unstructured":"Codella NCF, Gutman D, Celebi ME, Helba B, Marchetti MA, Dusza SW, et al. Skin lesion analysis toward melanoma detection: A challenge at the 2017 International symposium on biomedical imaging (ISBI), hosted by the international skin imaging collaboration (ISIC). In: 2018 IEEE 15th International Symposium on Biomedical Imaging (ISBI 2018). IEEE; 2018. p. 168\u201372.","DOI":"10.1109\/ISBI.2018.8363547"},{"key":"pdig.0001513.ref034","unstructured":"Combalia M, Codella NC, Rotemberg V, Helba B, Vilaplana V, Reiter O, Carrera C, et al. Bcn20000: Dermoscopic lesions in the wild. arXiv preprint arXiv:1908.02288. 2019."},{"issue":"7","key":"pdig.0001513.ref035","doi-asserted-by":"crossref","first-page":"1529","DOI":"10.1016\/j.jid.2018.01.028","article-title":"Classification of the Clinical Images for Benign and Malignant Cutaneous Tumors Using a Deep Learning Algorithm","volume":"138","author":"SS Han","year":"2018","journal-title":"J Invest Dermatol"},{"key":"pdig.0001513.ref036","doi-asserted-by":"crossref","first-page":"106221","DOI":"10.1016\/j.dib.2020.106221","article-title":"PAD-UFES-20: A skin lesion dataset composed of patient data and clinical images collected from smartphones","volume":"32","author":"AGC Pacheco","year":"2020","journal-title":"Data Brief"},{"key":"pdig.0001513.ref037","unstructured":"Kusner MJ, Loftus JR, Russell C, Silva R. Counterfactual fairness. In: Advances in Neural Information Processing Systems 30: Annual Conference on Neural Information Processing Systems 2017, December 4-9, 2017, Long Beach CA, USA. 2017. p. 4066\u201376."}],"container-title":["PLOS Digital Health"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pdig.0001513","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,16]],"date-time":"2026-07-16T17:23:20Z","timestamp":1784222600000},"score":1,"resource":{"primary":{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pdig.0001513"}},"subtitle":[],"editor":[{"given":"Martin G","family":"Frasch","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"editor"}]}],"short-title":[],"issued":{"date-parts":[[2026,7,16]]},"references-count":37,"journal-issue":{"issue":"7","published-online":{"date-parts":[[2026,7,16]]}},"URL":"https:\/\/doi.org\/10.1371\/journal.pdig.0001513","relation":{},"ISSN":["2767-3170"],"issn-type":[{"value":"2767-3170","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,7,16]]}}}