{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,28]],"date-time":"2026-04-28T05:16:35Z","timestamp":1777353395528,"version":"3.51.4"},"reference-count":57,"publisher":"Wiley","issue":"1","license":[{"start":{"date-parts":[[2023,3,21]],"date-time":"2023-03-21T00:00:00Z","timestamp":1679356800000},"content-version":"vor","delay-in-days":79,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100012166","name":"National Key Research and Development Program of China","doi-asserted-by":"publisher","award":["2022YFB2701401"],"award-info":[{"award-number":["2022YFB2701401"]}],"id":[{"id":"10.13039\/501100012166","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["62272124"],"award-info":[{"award-number":["62272124"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100010828","name":"Department of Education of Guizhou Province","doi-asserted-by":"publisher","award":["[2018]140"],"award-info":[{"award-number":["[2018]140"]}],"id":[{"id":"10.13039\/501100010828","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100010828","name":"Department of Education of Guizhou Province","doi-asserted-by":"publisher","award":["[2018]141"],"award-info":[{"award-number":["[2018]141"]}],"id":[{"id":"10.13039\/501100010828","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100003459","name":"Guizhou University","doi-asserted-by":"publisher","award":["[2020]61"],"award-info":[{"award-number":["[2020]61"]}],"id":[{"id":"10.13039\/501100003459","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100003459","name":"Guizhou University","doi-asserted-by":"publisher","award":["[2019]56"],"award-info":[{"award-number":["[2019]56"]}],"id":[{"id":"10.13039\/501100003459","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100002338","name":"Ministry of Education of the People's Republic of China","doi-asserted-by":"publisher","award":["GZUAMT2021KF[01]"],"award-info":[{"award-number":["GZUAMT2021KF[01]"]}],"id":[{"id":"10.13039\/501100002338","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["International Journal of Intelligent Systems"],"published-print":{"date-parts":[[2023,1]]},"abstract":"<jats:p>Shared gradients are widely used to protect the private information of training data in distributed machine learning systems. However, Deep Leakage from Gradients (DLG) research has found that private training data can be recovered from shared gradients. The DLG method still has some issues such as the \u201cExploding Gradient,\u201d low attack success rate, and low fidelity of recovered data. In this study, a Wasserstein DLG method, named WDLG, is proposed; the theoretical analysis shows that under the premise that the output layer of the model has a \u201cbias\u201d term, predicting the \u201clabel\u201d of the data by whether the \u201cbias\u201d is \u201cnegative\u201d or not is independent of the approximation of the shared gradient, and thus, the label of the data can be recovered with 100% accuracy. In the proposed method, the Wasserstein distance is used to calculate the error loss between the shared gradient and the virtual gradient, which improves model training stability, solves the \u201cExploding Gradient\u201d phenomenon, and improves the fidelity of the recovered data. Moreover, a large learning rate strategy is designed to improve model training convergence speed in\u2010depth. Finally, the WDLG method is validated on datasets from MNIST, Fashion MNIST, SVHN, CIFAR\u2010100, and LFW. Experiments results show that the proposed WDLG method provides more stable updates for virtual data, a higher attack success rate, faster model convergence, higher image fidelity during recovery, and support for designing large learning rate strategies.<\/jats:p>","DOI":"10.1155\/2023\/5510329","type":"journal-article","created":{"date-parts":[[2023,3,22]],"date-time":"2023-03-22T02:35:08Z","timestamp":1679452508000},"update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":12,"title":["Fast and Accurate Deep Leakage from Gradients Based on Wasserstein Distance"],"prefix":"10.1155","volume":"2023","author":[{"ORCID":"https:\/\/orcid.org\/0009-0004-9650-0786","authenticated-orcid":false,"given":"Xing","family":"He","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-8733-4596","authenticated-orcid":false,"given":"Changgen","family":"Peng","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6590-5757","authenticated-orcid":false,"given":"Weijie","family":"Tan","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"311","published-online":{"date-parts":[[2023,3,21]]},"reference":[{"key":"e_1_2_9_1_2","first-page":"1273","article-title":"Communication-efficient learning of deep networks from decentralized data","author":"McMahan B.","year":"2017","journal-title":"Artificial intelligence and statistics PMLR"},{"key":"e_1_2_9_2_2","unstructured":"Kone\u010dn\u00fdJ. McMahanH. B. andYuF. X. Federated learning: strategies for improving communication efficiency 2016 https:\/\/arxiv.org\/abs\/1610.05492."},{"key":"e_1_2_9_3_2","doi-asserted-by":"publisher","DOI":"10.1109\/msp.2020.2975749"},{"key":"e_1_2_9_4_2","first-page":"429","article-title":"Federated optimization in heterogeneous networks","volume":"2","author":"Li T.","year":"2020","journal-title":"Proceedings of Machine Learning and Systems"},{"key":"e_1_2_9_5_2","doi-asserted-by":"publisher","DOI":"10.1561\/2200000083"},{"key":"e_1_2_9_6_2","doi-asserted-by":"crossref","unstructured":"NasrM. ShokriR. andHoumansadrA. Comprehensive privacy analysis of deep learning: passive and active white-box inference attacks against centralized and federated learning Proceedings of the 2019 IEEE symposium on security and privacy (SP) May 2019 San Francisco CA USA 739\u2013753.","DOI":"10.1109\/SP.2019.00065"},{"key":"e_1_2_9_7_2","doi-asserted-by":"publisher","DOI":"10.1108\/ijwis-04-2022-0078"},{"key":"e_1_2_9_8_2","first-page":"16937","article-title":"Inverting gradients-how easy is it to break privacy in federated learning?","volume":"33","author":"Geiping J.","year":"2020","journal-title":"Advances in Neural Information Processing Systems"},{"key":"e_1_2_9_9_2","unstructured":"GengJ. MouY. andLiF. Towards general deep leakage in federated learning 2021 https:\/\/arxiv.org\/abs\/2110.09074."},{"key":"e_1_2_9_10_2","doi-asserted-by":"publisher","DOI":"10.1109\/tifs.2017.2787987"},{"key":"e_1_2_9_11_2","doi-asserted-by":"crossref","unstructured":"AbadiM. ChuA. andGoodfellowI. Deep learning with differential privacy Proceedings of the 2016 ACM SIGSAC conference on computer and communications security October 2016 Vienna Austria 308\u2013318.","DOI":"10.1145\/2976749.2978318"},{"key":"e_1_2_9_12_2","first-page":"374","article-title":"Towards federated learning at scale: system design","volume":"1","author":"Bonawitz K.","year":"2019","journal-title":"Proceedings of Machine Learning and Systems"},{"key":"e_1_2_9_13_2","unstructured":"LiD.andFedmdW. J. Heterogenous federated learning via model distillation 1\u20138 Proceedings of the NeurIPS Workshop Feder. Learn. Data Privacy Confidentiality December 2019 Vancouver Canada."},{"key":"e_1_2_9_14_2","doi-asserted-by":"publisher","DOI":"10.1002\/int.22907"},{"key":"e_1_2_9_15_2","unstructured":"MelisL. SongC. andDe CristofaroE. Inference attacks against collaborative learning 2018 https:\/\/arxiv.org\/abs\/1805.04049."},{"key":"e_1_2_9_16_2","doi-asserted-by":"crossref","unstructured":"MelisL. SongC. andDe CristofaroE. Exploiting unintended feature leakage in collaborative learning Proceedings of the 2019 IEEE symposium on security and privacy (SP) May 2019 California CA USA IEEE 691\u2013706.","DOI":"10.1109\/SP.2019.00029"},{"key":"e_1_2_9_17_2","doi-asserted-by":"crossref","unstructured":"FredriksonM. JhaS. andRistenpartT. Model inversion attacks that exploit confidence information and basic countermeasures Proceedings of the 22nd ACM SIGSAC conference on computer and communications security October 2015 Colorado CO USA 1322\u20131333.","DOI":"10.1145\/2810103.2813677"},{"key":"e_1_2_9_18_2","doi-asserted-by":"crossref","unstructured":"ShokriR. StronatiM. andSongC. Membership inference attacks against machine learning models Proceedings of the 2017 IEEE symposium on security and privacy (SP) May 2017 California CA USA IEEE 3\u201318.","DOI":"10.1109\/SP.2017.41"},{"key":"e_1_2_9_19_2","doi-asserted-by":"publisher","DOI":"10.2478\/popets-2019-0008"},{"key":"e_1_2_9_20_2","unstructured":"SalemA. BhattacharyaA. andBackesM. Updates-leak: data set inference and reconstruction attacks in online learning Proceedings of the 29th USENIX security symposium (USENIX Security 20) August 2020 Berkeley CA USA 1291\u20131308."},{"key":"e_1_2_9_21_2","first-page":"318","article-title":"Artificial intelligence security governance challenges and countermeasures","volume":"8","author":"Peng C. G.","year":"2022","journal-title":"Journal of Information Security Research"},{"key":"e_1_2_9_22_2","doi-asserted-by":"publisher","DOI":"10.1007\/s11263-016-0911-8"},{"key":"e_1_2_9_23_2","article-title":"Inceptionism: going deeper into neural networks","author":"Mordvintsev A.","year":"2015","journal-title":"Google Research Blog"},{"key":"e_1_2_9_24_2","article-title":"Deep leakage from gradients","volume":"32","author":"Zhu L.","year":"2019","journal-title":"Advances in Neural Information Processing Systems"},{"key":"e_1_2_9_25_2","unstructured":"ZhaoB. MopuriK. R. andBilenH. Idlg: improved deep leakage from gradients 2020 https:\/\/arxiv.org\/abs\/2001.02610."},{"key":"e_1_2_9_26_2","unstructured":"LeCunY. CortesC. andBurgesC. J. C. The mnist database of handwritten digits 1998 https:\/\/idr.openmicroscopy.org\/."},{"key":"e_1_2_9_27_2","unstructured":"XiaoH. RasulK. andVollgrafR. Fashion-mnist: a novel image dataset for benchmarking machine learning algorithms 2017 https:\/\/arxiv.org\/abs\/1708.07747."},{"key":"e_1_2_9_28_2","article-title":"Reading digits in natural images with unsupervised feature learning","volume":"5","author":"Netzer Y.","year":"2011","journal-title":"In NIPS Workshop"},{"key":"e_1_2_9_29_2","unstructured":"KrizhevskyA.andHintonG. Learning multiple layers of features from tiny images 2009 Master\u2019s Thesis University of Toronto Canada."},{"key":"e_1_2_9_30_2","article-title":"Labeled faces in the wild: a database forstudying face recognition in unconstrained environments","author":"Huang G. B.","year":"2007","journal-title":"Technical Report 07\u201349"},{"key":"e_1_2_9_31_2","article-title":"Hogwild: a lock-free approach to parallelizing stochastic gradient descent","volume":"24","author":"Recht B.","year":"2011","journal-title":"Advances in Neural Information Processing Systems"},{"key":"e_1_2_9_32_2","article-title":"More effective distributed ml via a stale synchronous parallel parameter server","volume":"26","author":"Ho Q.","year":"2013","journal-title":"Advances in Neural Information Processing Systems"},{"key":"e_1_2_9_33_2","article-title":"Can decentralized algorithms outperform centralized algorithms? a case study for decentralized parallel stochastic gradient descent","volume":"30","author":"Lian X.","year":"2017","journal-title":"Advances in Neural Information Processing Systems"},{"key":"e_1_2_9_34_2","unstructured":"SergeevA.andDel BalsoM. Horovod: fast and easy distributed deep learning in TensorFlow 2018 https:\/\/arxiv.org\/abs\/1802.05799."},{"key":"e_1_2_9_35_2","unstructured":"AbadiM. BarhamP. andChenJ. TensorFlow a system for Large-Scale machine learning Proceedings of the 12th USENIX symposium on operating systems design and implementation (OSDI 16) November 2016 Savannah GA USA 265\u2013283."},{"key":"e_1_2_9_36_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.ress.2017.11.021"},{"key":"e_1_2_9_37_2","unstructured":"YuanJ. LiX. andChengC. Oneflow: redesign the distributed deep learning framework from scratch 2021 https:\/\/arxiv.org\/."},{"key":"e_1_2_9_38_2","unstructured":"IandolaF. N. HanS. andMoskewiczM. W. SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and< 0.5 MB model size 2016 https:\/\/arxiv.org\/abs\/1602.07360."},{"key":"e_1_2_9_39_2","article-title":"Terngrad: ternary gradients to reduce communication in distributed deep learning","volume":"30","author":"Wen W.","year":"2017","journal-title":"Advances in Neural Information Processing Systems"},{"key":"e_1_2_9_40_2","doi-asserted-by":"publisher","DOI":"10.1109\/tpds.2021.3062721"},{"key":"e_1_2_9_41_2","doi-asserted-by":"publisher","DOI":"10.1016\/j.jpdc.2008.09.002"},{"key":"e_1_2_9_42_2","unstructured":"JangE. GuS. andPooleB. Categorical reparameterization with gumbel-softmax 2016 https:\/\/arxiv.org\/abs\/1611.01144."},{"key":"e_1_2_9_43_2","doi-asserted-by":"publisher","DOI":"10.1016\/s1352-2310(97)00447-0"},{"key":"e_1_2_9_44_2","doi-asserted-by":"publisher","DOI":"10.1038\/323533a0"},{"key":"e_1_2_9_45_2","doi-asserted-by":"crossref","unstructured":"YeomS. GiacomelliI. andFredriksonM. Privacy risk in machine learning: analyzing the connection to overfitting Proceedings of the 2018 IEEE 31st computer security foundations symposium (CSF) July 2018 Oxford UK 268\u2013282.","DOI":"10.1109\/CSF.2018.00027"},{"key":"e_1_2_9_46_2","doi-asserted-by":"crossref","unstructured":"SalemA. ZhangY. andHumbertM. Ml-leaks: model and data independent membership inference attacks and defenses on machine learning models 2018 https:\/\/arxiv.org\/abs\/1806.01246.","DOI":"10.14722\/ndss.2019.23119"},{"key":"e_1_2_9_47_2","unstructured":"LeinoK.andFredriksonM. Stolen memories: leveraging model memorization for calibrated white-box membership inference Proceedings of the 29th USENIX security symposium (USENIX Security 20) August 2020 Boston MA USA 1605\u20131622."},{"key":"e_1_2_9_48_2","doi-asserted-by":"crossref","unstructured":"SongL. ShokriR. andMittalP. Privacy risks of securing machine learning models against adversarial examples Proceedings of the 2019 ACM SIGSAC Conference on Computer and Communications Security November 2019 London UK 241\u2013257.","DOI":"10.1145\/3319535.3354211"},{"key":"e_1_2_9_49_2","article-title":"Generative adversarial nets","volume":"1050","author":"Goodfellow I. J.","year":"2014","journal-title":"Stat"},{"key":"e_1_2_9_50_2","doi-asserted-by":"crossref","unstructured":"HitajB. AtenieseG. andPerez-CruzF. Deep models under the GAN: information leakage from collaborative deep learning Proceedings of the 2017 ACM SIGSAC conference on computer and communications security November 2017 Dallas TX USA 603\u2013618.","DOI":"10.1145\/3133956.3134012"},{"key":"e_1_2_9_51_2","doi-asserted-by":"publisher","DOI":"10.1109\/tcbb.2019.2940583"},{"key":"e_1_2_9_52_2","unstructured":"PanX. ZhangM. andYanY. Theory-oriented deep leakage from gradients via linear equation solver 2020 https:\/\/arxiv.org\/abs\/2010.13356."},{"key":"e_1_2_9_53_2","doi-asserted-by":"crossref","unstructured":"LiZ. HubchakM. andZhuY. Deep leakage from gradients in multiple-label medical image classification IEEE Proceedings of the 2021 IEEE 9th International Conference on Healthcare Informatics (ICHI) August 2021 Victoria BC Canada 447\u2013448.","DOI":"10.1109\/ICHI52183.2021.00078"},{"key":"e_1_2_9_54_2","doi-asserted-by":"publisher","DOI":"10.1002\/int.22997"},{"key":"e_1_2_9_55_2","unstructured":"ArjovskyM. ChintalaS. andBottouL. Wasserstein generative adversarial networks Proceedings of the International conference on machine learning August 2017 Sydney Australia 214\u2013223."},{"key":"e_1_2_9_56_2","article-title":"Sinkhorn distances: lightspeed computation of optimal transport","volume":"26","author":"Cuturi M.","year":"2013","journal-title":"Advances in Neural Information Processing Systems"},{"key":"e_1_2_9_57_2","article-title":"Automatic differentiation in PyTorch","volume":"3","author":"Paszke A.","year":"2017","journal-title":"In NIPS-W"}],"container-title":["International Journal of Intelligent Systems"],"original-title":[],"language":"en","link":[{"URL":"http:\/\/downloads.hindawi.com\/journals\/ijis\/2023\/5510329.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"http:\/\/downloads.hindawi.com\/journals\/ijis\/2023\/5510329.xml","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1155\/2023\/5510329","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,12,31]],"date-time":"2024-12-31T05:33:53Z","timestamp":1735623233000},"score":1,"resource":{"primary":{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/10.1155\/2023\/5510329"}},"subtitle":[],"editor":[{"given":"Yu-an","family":"Tan","sequence":"additional","affiliation":[],"role":[{"role":"editor","vocabulary":"crossref"}]}],"short-title":[],"issued":{"date-parts":[[2023,1]]},"references-count":57,"journal-issue":{"issue":"1","published-print":{"date-parts":[[2023,1]]}},"alternative-id":["10.1155\/2023\/5510329"],"URL":"https:\/\/doi.org\/10.1155\/2023\/5510329","archive":["Portico"],"relation":{},"ISSN":["0884-8173","1098-111X"],"issn-type":[{"value":"0884-8173","type":"print"},{"value":"1098-111X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2023,1]]},"assertion":[{"value":"2022-12-05","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2023-02-28","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2023-03-21","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"5510329"}}