{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,2,4]],"date-time":"2026-02-04T17:37:30Z","timestamp":1770226650397,"version":"3.49.0"},"reference-count":36,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2023,5,24]],"date-time":"2023-05-24T00:00:00Z","timestamp":1684886400000},"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. Comput. Sci."],"abstract":"<jats:p>Semi-supervised learning (SSL) methods provide a powerful tool for utilizing abundant unlabeled data to strengthen standard supervised learning. Traditional graph-based SSL methods prevail in classical SSL problems for their intuitional implementation and effective performance. However, they encounter troubles when applying to image classification followed by modern deep learning, since the diffusion algorithms face the curse of dimensionality. In this study, we propose a simple and efficient SSL method, combining a graph-based SSL paradigm with differential privacy. We aim at developing coherent latent feature space of deep neural networks so that the diffusion algorithm in the latent space can give more precise predictions for unlabeled data. Our approach achieves state-of-the-art performance on the Cifar10, Cifar100, and Mini-imagenet benchmark datasets and obtains an error rate of 18.56% on Cifar10 using only 1% of all labels. Furthermore, our approach inherits the benefits of graph-based SSL methods with a simple training process and can be easily combined with any network architecture.<\/jats:p>","DOI":"10.3389\/fcomp.2023.1114186","type":"journal-article","created":{"date-parts":[[2023,5,24]],"date-time":"2023-05-24T05:19:24Z","timestamp":1684905564000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":3,"title":["Robust deep semi-supervised learning with label propagation and differential privacy"],"prefix":"10.3389","volume":"5","author":[{"given":"Zhicong","family":"Yan","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Shenghong","family":"Li","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Zhongli","family":"Duan","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yuanyuan","family":"Zhao","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"1965","published-online":{"date-parts":[[2023,5,24]]},"reference":[{"key":"B1","article-title":"\u201cThere are many consistent explanations of unlabeled data: why you should average,\u201d","volume-title":"Proceedings of the International Conference on Learning Representations (ICLR)","author":"Athiwaratkun","year":"2019"},{"key":"B2","first-page":"2399","article-title":"Manifold regularization: a geometric framework for learning from labeled and unlabeled examples","volume":"7","author":"Belkin","year":"2006","journal-title":"J. Mach. Learn. Res"},{"key":"B3","article-title":"\u201cLarge scale GAN training for high fidelity natural image synthesis,\u201d","volume-title":"Proceedings of the International Conference on Learning Representations (ICLR)","author":"Brock","year":"2018"},{"key":"B4","doi-asserted-by":"publisher","DOI":"10.7551\/mitpress\/9780262033589.001.0001","author":"Chapelle","year":"2006","journal-title":"Semi-Supervised Learning"},{"key":"B5","article-title":"\u201cGood semi-supervised learning that requires a bad GAN,\u201d","volume-title":"Proceedings of the Advances in Neural Information Processing Systems (NIPS)","author":"Dai","year":"2017"},{"key":"B6","doi-asserted-by":"publisher","first-page":"211","DOI":"10.1561\/0400000042","article-title":"The algorithmic foundations of differential privacy","volume":"9","author":"Dwork","year":"2014","journal-title":"Found. Trends Theor. Comput. Sci"},{"key":"B7","first-page":"4367","article-title":"\u201cDynamic few-shot visual learning without forgetting,\u201d","volume-title":"Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition","author":"Gidaris","year":"2018"},{"key":"B8","doi-asserted-by":"publisher","first-page":"2261","DOI":"10.1109\/TNNLS.2014.2376936","article-title":"Deformed graph Laplacian for semisupervised learning","volume":"26","author":"Gong","year":"2015","journal-title":"IEEE Trans. Neural Netw. Learn. Syst"},{"key":"B9","article-title":"\u201cGenerative adversarial nets,\u201d","volume-title":"Proceedings of the Advances in Neural Information Processing Systems (NIPS)","author":"Goodfellow","year":"2014"},{"key":"B10","article-title":"\u201cImproved training of wasserstein GANs,\u201d","volume-title":"Proceedings of the Advances in Neural Information Processing Systems (NIPS)","author":"Gulrajani","year":"2017"},{"key":"B11","doi-asserted-by":"crossref","DOI":"10.1109\/CVPR.2019.00521","article-title":"\u201cLabel propagation for deep semi-supervised learning,\u201d","volume-title":"Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR)","author":"Iscen","year":"2019"},{"key":"B12","first-page":"441","article-title":"\u201cGraph construction and b-matching for semi-supervised learning,\u201d","volume-title":"Proceedings of the International Conference on Machine Learning (ICML)","author":"Jebara","year":"2009"},{"key":"B13","first-page":"2459","article-title":"\u201cSemi-supervised learning via compact latent space clustering,\u201d","volume-title":"International Conference on Machine Learning","author":"Kamnitsas","year":"2018"},{"key":"B14","article-title":"\u201cSemi-supervised learning with deep generative models,\u201d","volume-title":"Proceedings of the Advances in Neural Information Processing Systems (NIPS)","author":"Kingma","year":"2014"},{"key":"B15","article-title":"Auto-encoding variational bayes","author":"Kingma","year":"2013","journal-title":"arXiv preprint arXiv:1312.6114"},{"key":"B16","doi-asserted-by":"publisher","first-page":"3964","DOI":"10.1109\/TPAMI.2020.2992934","article-title":"Normalizing flows: an introduction and review of current methods","volume":"43","author":"Kobyzev","year":"2020","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell"},{"key":"B17","volume-title":"Learning multiple layers of features from tiny images","author":"Krizhevsky","year":"2009"},{"key":"B18","doi-asserted-by":"publisher","first-page":"3267","DOI":"10.1109\/TII.2019.2951766","article-title":"Desvig: decentralized swift vigilance against adversarial attacks in industrial artificial intelligence systems","volume":"16","author":"Li","year":"2020","journal-title":"IEEE TII"},{"key":"B19","doi-asserted-by":"publisher","first-page":"1260","DOI":"10.1109\/TII.2022.3173996","article-title":"Multi-tentacle federated learning over software-defined industrial internet of things against adaptive poisoning attacks","volume":"19","author":"Li","year":"2022","journal-title":"IEEE Trans. Indus. Inform"},{"key":"B20","article-title":"SGDR: Stochastic gradient descent with warm restarts","author":"Loshchilov","year":"2016","journal-title":"arXiv preprint arXiv:1608.03983"},{"key":"B21","doi-asserted-by":"crossref","DOI":"10.1109\/CVPR.2018.00927","article-title":"\u201cSmooth neighbors on teacher graphs for semi-supervised learning,\u201d","volume-title":"Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR)","author":"Luo","year":"2018"},{"key":"B22","first-page":"1445","article-title":"\u201cAuxiliary deep generative models,\u201d","volume-title":"Proceedings of the International Conference on Machine Learning (ICML)","author":"Maal\u00f8e","year":"2016"},{"key":"B23","article-title":"Adversarial autoencoders","author":"Makhzani","year":"2015","journal-title":"arXiv preprint arXiv:1511.05644"},{"key":"B24","doi-asserted-by":"publisher","first-page":"1979","DOI":"10.1109\/TPAMI.2018.2858821","article-title":"Virtual adversarial training: a regularization method for supervised and semi-supervised learning","volume":"41","author":"Miyato","year":"2019","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell"},{"key":"B25","article-title":"Unsupervised representation learning with deep convolutional generative adversarial networks","author":"Radford","year":"2015","journal-title":"arXiv preprint arXiv:1511.06434"},{"key":"B26","article-title":"\u201cSemi-supervised learning with ladder networks,\u201d","volume-title":"Proceedings of the Advances in Neural Information Processing Systems (NIPS)","author":"Rasmus","year":"2015"},{"key":"B27","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1109\/TDSC.2022.3222972","article-title":"Protecting intellectual property with reliable availability of learning models in ai-based cybersecurity services","author":"Ren","year":"2022","journal-title":"IEEE TDSC"},{"key":"B28","first-page":"1163","article-title":"\u201cRegularization with stochastic transformations and perturbations for deep semi-supervised learning,\u201d","volume-title":"Proceedings of the Advances in Neural Information Processing Systems (NIPS)","author":"Sajjadi","year":"2016"},{"key":"B29","article-title":"\u201cImproved techniques for training GANs,\u201d","volume-title":"Proceedings of the Advances in Neural Information Processing Systems (NIPS)","author":"Salimans","year":"2016"},{"key":"B30","first-page":"1195","article-title":"\u201cMean teachers are better role models: weight-averaged consistency targets improve semi-supervised deep learning results,\u201d","volume-title":"Proceedings of the Advances in Neural Information Processing Systems (NIPS)","author":"Tarvainen","year":"2017"},{"key":"B31","doi-asserted-by":"publisher","first-page":"173","DOI":"10.1016\/j.neucom.2015.01.020","article-title":"Posterior distribution learning (PDL): a novel supervised learning framework using unlabeled samples to improve classification performance","volume":"157","author":"Tu","year":"2015","journal-title":"Neurocomputing"},{"key":"B32","first-page":"3635","article-title":"\u201cInterpolation consistency training for semi-supervised learning,\u201d","volume-title":"Proceedings of the International Joint Conference on Artificial Intelligence (IJCAI)","author":"Verma","year":"2019"},{"key":"B33","first-page":"10676","article-title":"\u201cTangent-normal adversarial regularization for semi-supervised learning,\u201d","volume-title":"Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition (CVPR)","author":"Yu","year":"2019"},{"key":"B34","article-title":"\u201cMixup: beyond empirical risk minimization,\u201d","volume-title":"Proceedings of the International Conference on Learning Representations (ICLR)","author":"Zhang","year":"2018"},{"key":"B35","article-title":"\u201cLearning with local and global consistency,\u201d","volume-title":"Proceedings of the Advances in Neural Information Processing Systems (NIPS)","author":"Zhou","year":"2003"},{"key":"B36","first-page":"912","article-title":"\u201cSemi-supervised learning using gaussian fields and harmonic functions,\u201d","volume-title":"Proceedings of the International Conference on Machine Learning (ICML)","author":"Zhu","year":"2003"}],"container-title":["Frontiers in Computer Science"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fcomp.2023.1114186\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,5,24]],"date-time":"2023-05-24T05:19:39Z","timestamp":1684905579000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fcomp.2023.1114186\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,5,24]]},"references-count":36,"alternative-id":["10.3389\/fcomp.2023.1114186"],"URL":"https:\/\/doi.org\/10.3389\/fcomp.2023.1114186","relation":{},"ISSN":["2624-9898"],"issn-type":[{"value":"2624-9898","type":"electronic"}],"subject":[],"published":{"date-parts":[[2023,5,24]]},"article-number":"1114186"}}