{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,2,12]],"date-time":"2026-02-12T17:08:15Z","timestamp":1770916095850,"version":"3.50.1"},"reference-count":120,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","issue":"7","license":[{"start":{"date-parts":[[2025,7,1]],"date-time":"2025-07-01T00:00:00Z","timestamp":1751328000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2025,7,1]],"date-time":"2025-07-01T00:00:00Z","timestamp":1751328000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2025,7,1]],"date-time":"2025-07-01T00:00:00Z","timestamp":1751328000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. Neural Netw. Learning Syst."],"published-print":{"date-parts":[[2025,7]]},"DOI":"10.1109\/tnnls.2024.3443088","type":"journal-article","created":{"date-parts":[[2024,8,27]],"date-time":"2024-08-27T13:23:59Z","timestamp":1724765039000},"page":"13512-13526","source":"Crossref","is-referenced-by-count":5,"title":["Kernel Masked Image Modeling Through the Lens of Theoretical Understanding"],"prefix":"10.1109","volume":"36","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-0010-7508","authenticated-orcid":false,"given":"Yurui","family":"Qian","sequence":"first","affiliation":[{"name":"School of Cyber Science and Technology, University of Science and Technology of China, Hefei, China"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-4219-781X","authenticated-orcid":false,"given":"Yu","family":"Wang","sequence":"additional","affiliation":[{"name":"resides in Saratoga, Saratoga, CA, USA"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4749-105X","authenticated-orcid":false,"given":"Jingjing","family":"Zou","sequence":"additional","affiliation":[{"name":"Herbert Wertheim School of Public Health and Human Longevity Science, University of California at San Diego, La Jolla, CA, USA"}]},{"ORCID":"https:\/\/orcid.org\/0009-0000-3223-3827","authenticated-orcid":false,"given":"Jingyang","family":"Lin","sequence":"additional","affiliation":[{"name":"Department of Computer Science, University of Rochester, Rochester, NY, USA"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4344-8898","authenticated-orcid":false,"given":"Yingwei","family":"Pan","sequence":"additional","affiliation":[{"name":"HiDream.ai, Beijing, China"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7587-101X","authenticated-orcid":false,"given":"Ting","family":"Yao","sequence":"additional","affiliation":[{"name":"HiDream.ai, Beijing, China"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6789-7460","authenticated-orcid":false,"given":"Qibin","family":"Sun","sequence":"additional","affiliation":[{"name":"School of Cyber Science and Technology, University of Science and Technology of China, Hefei, China"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2497-7732","authenticated-orcid":false,"given":"Tao","family":"Mei","sequence":"additional","affiliation":[{"name":"HiDream.ai, Beijing, China"}]}],"member":"263","reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.90"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2017.322"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.1109\/TNNLS.2015.2444417"},{"key":"ref4","article-title":"An image is worth 16\u00d716 words: Transformers for image recognition at scale","volume-title":"Proc. ICLR","author":"Dosovitskiy"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.1109\/TNNLS.2022.3152527"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1109\/tnnls.2022.3152990"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2022.3164083"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2023.3268446"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR42600.2020.00975"},{"key":"ref10","article-title":"BEiT v2: Masked image modeling with vector-quantized visual tokenizers","author":"Peng","year":"2022","journal-title":"arXiv:2208.06366"},{"key":"ref11","first-page":"1597","article-title":"A simple framework for contrastive learning of visual representations","volume-title":"Proc. ICML","author":"Chen"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1109\/tnnls.2023.3265607"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1109\/TNNLS.2022.3191086"},{"key":"ref14","first-page":"12638","article-title":"Joint contrastive learning with infinite possibilities","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Cai"},{"key":"ref15","first-page":"27617","article-title":"Improving self-supervised learning with automated unsupervised outlier arbitration","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Wang"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2022.3180995"},{"key":"ref17","first-page":"4171","article-title":"BERT: Pre-training of deep bidirectional transformers for language understanding","volume-title":"Proc. NAACL","author":"Devlin"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.1109\/WACV56688.2023.00201"},{"key":"ref19","first-page":"10078","article-title":"VideoMAE: Masked autoencoders are data-efficient learners for self-supervised video pre-training","volume-title":"Proc. NeurIPS","author":"Tong"},{"key":"ref20","article-title":"iBOT: Image BERT pre-training with online tokenizer","volume-title":"Proc. ICLR","author":"Zhou"},{"key":"ref21","article-title":"BEiT: BERT pre-training of image transformers","volume-title":"Proc. ICLR","author":"Bao"},{"key":"ref22","first-page":"5628","article-title":"A theoretical analysis of contrastive unsupervised representation learning","volume-title":"Proc. ICML","author":"Saunshi"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v37i1.25130"},{"key":"ref24","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-20056-4_15"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52688.2022.01553"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52688.2022.00943"},{"key":"ref27","first-page":"536","article-title":"Kernel PCA and de-noising in feature spaces","volume-title":"Proc. NeurIPS","author":"Mika"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1162\/089976698300017467"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1090\/s0002-9947-1950-0051437-7"},{"key":"ref30","article-title":"Unsupervised representation learning by predicting image rotations","volume-title":"Proc. ICLR","author":"Gidaris"},{"key":"ref31","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-46466-4_5"},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.278"},{"key":"ref33","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2017.76"},{"key":"ref34","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2017.96"},{"key":"ref35","article-title":"Representation learning with contrastive predictive coding","author":"van den Oord","year":"2018","journal-title":"arXiv:1807.03748"},{"key":"ref36","doi-asserted-by":"publisher","DOI":"10.1109\/tnnls.2024.3388424"},{"key":"ref37","first-page":"12310","article-title":"Barlow twins: Self-supervised learning via redundancy reduction","volume-title":"Proc. ICML","author":"Zbontar"},{"key":"ref38","doi-asserted-by":"publisher","DOI":"10.5555\/3495724.3497510"},{"key":"ref39","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR46437.2021.01549"},{"key":"ref40","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV48922.2021.00951"},{"key":"ref41","article-title":"VICReg: Variance-invariance-covariance regularization for self-supervised learning","volume-title":"Proc. ICLR","author":"Bardes"},{"key":"ref42","article-title":"Understanding dimensional collapse in contrastive self-supervised learning","volume-title":"Proc. ICLR","author":"Jing"},{"key":"ref43","first-page":"10268","article-title":"Understanding self-supervised learning dynamics without contrastive pairs","volume-title":"Proc. ICML","author":"Tian"},{"key":"ref44","doi-asserted-by":"publisher","DOI":"10.1109\/TNNLS.2022.3219936"},{"key":"ref45","doi-asserted-by":"publisher","DOI":"10.1109\/TNNLS.2022.3172423"},{"key":"ref46","doi-asserted-by":"publisher","DOI":"10.1109\/tnnls.2023.3348833"},{"key":"ref47","doi-asserted-by":"publisher","DOI":"10.1109\/TNNLS.2022.3159538"},{"key":"ref48","doi-asserted-by":"publisher","DOI":"10.1109\/TNNLS.2022.3204597"},{"key":"ref49","doi-asserted-by":"publisher","DOI":"10.48550\/ARXIV.1907.11692"},{"key":"ref50","article-title":"ELECTRA: Pre-training text encoders as discriminators rather than generators","volume-title":"Proc. ICLR","author":"Clark"},{"key":"ref51","article-title":"Masked feature prediction for self-supervised visual pre-training","author":"Wei","year":"2021","journal-title":"arXiv:2112.09133"},{"key":"ref52","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-20056-4_7"},{"key":"ref53","first-page":"29273","article-title":"RevColV2: Exploring disentangled representations in masked image modeling","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"36","author":"Han"},{"key":"ref54","article-title":"Masked capsule autoencoders","author":"Everett","year":"2024","journal-title":"arXiv:2403.04724"},{"key":"ref55","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52729.2023.01548"},{"key":"ref56","article-title":"Architecture-agnostic masked image modeling\u2014From ViT back to CNN","author":"Li","year":"2022","journal-title":"arXiv:2205.13943"},{"key":"ref57","article-title":"Masked frequency modeling for self-supervised visual pre-training","author":"Xie","year":"2022","journal-title":"arXiv:2206.07706"},{"key":"ref58","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52729.2023.01394"},{"key":"ref59","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52729.2023.01000"},{"key":"ref60","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v37i2.25269"},{"key":"ref61","article-title":"Layer grafted pre-training: Bridging contrastive learning and masked image modeling for label-efficient representations","volume-title":"Proc. 11th Int. Conf. Learn. Represent.","author":"Jiang"},{"key":"ref62","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v37i3.25419"},{"key":"ref63","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2023.3336525"},{"key":"ref64","article-title":"MaskViT: Masked visual pre-training for video prediction","author":"Gupta","year":"2022","journal-title":"arXiv:2206.11894"},{"key":"ref65","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52688.2022.01103"},{"key":"ref66","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.02554"},{"key":"ref67","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.00574"},{"key":"ref68","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.01525"},{"key":"ref69","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.01113"},{"key":"ref70","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2023.3322604"},{"key":"ref71","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2024.3412004"},{"key":"ref72","article-title":"Understanding masked image modeling via learning occlusion invariant feature","author":"Kong","year":"2022","journal-title":"arXiv:2208.04164"},{"key":"ref73","article-title":"How to understand masked autoencoders","author":"Cao","year":"2022","journal-title":"arXiv:2202.03670"},{"key":"ref74","doi-asserted-by":"publisher","DOI":"10.5555\/3295222.3295349"},{"key":"ref75","article-title":"Towards understanding why mask-reconstruction pretraining helps in downstream tasks","author":"Pan","year":"2022","journal-title":"arXiv:2206.03826"},{"key":"ref76","first-page":"40519","article-title":"On the power of foundation models","volume-title":"Proc. ICML","author":"Yuan"},{"key":"ref77","first-page":"27127","article-title":"How mask matters: Towards theoretical understandings of masked autoencoders","volume-title":"Proc. NeurIPS","author":"Zhang"},{"key":"ref78","first-page":"5000","article-title":"Provable guarantees for self-supervised deep learning with spectral contrastive loss","volume-title":"Proc. NeurIPS","author":"HaoChen"},{"key":"ref79","article-title":"Integrating prior knowledge in contrastive learning with kernel","author":"Dufumier","year":"2022","journal-title":"arXiv:2206.01646"},{"key":"ref80","article-title":"Conditional contrastive learning with kernel","volume-title":"Proc. ICLR","author":"Tsai"},{"key":"ref81","first-page":"15543","article-title":"Self-supervised learning with kernel dependence maximization","volume-title":"Proc. NeurIPS","author":"Li"},{"key":"ref82","article-title":"Joint embedding self-supervised learning in the kernel regime","author":"Kiani","year":"2022","journal-title":"arXiv:2209.14884"},{"key":"ref83","article-title":"Contrastive learning is spectral clustering on similarity graph","author":"Tan","year":"2023","journal-title":"arXiv:2303.15103"},{"key":"ref84","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v38i11.29077"},{"key":"ref85","first-page":"60712","article-title":"Neural harmonics: Bridging spectral embedding and matrix completion in self-supervised learning","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"36","author":"Munkhoeva"},{"key":"ref86","doi-asserted-by":"publisher","DOI":"10.1145\/1390156.1390294"},{"key":"ref87","volume-title":"Graph Kernels by Spectral Transforms","author":"Chapelle","year":"2006"},{"key":"ref88","article-title":"LipsFormer: Introducing Lipschitz continuity to vision transformers","author":"Qi","year":"2023","journal-title":"arXiv:2304.09856"},{"key":"ref89","first-page":"10347","article-title":"Training data-efficient image transformers & distillation through attention","volume-title":"Proc. ICML","author":"Touvron"},{"key":"ref90","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV48922.2021.00950"},{"key":"ref91","article-title":"SupMAE: Supervised masked autoencoders are efficient vision learners","author":"Liang","year":"2022","journal-title":"arXiv:2205.14540"},{"key":"ref92","doi-asserted-by":"publisher","DOI":"10.1007\/s11263-023-01852-4"},{"key":"ref93","article-title":"Progressively compressed auto-encoder for self-supervised representation learning","volume-title":"Proc. 11th Int. Conf. Learn. Represent.","author":"Li"},{"key":"ref94","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2009.5206848"},{"key":"ref95","article-title":"ELEVATER: A benchmark and toolkit for evaluating language-augmented visual models","author":"Li","year":"2022","journal-title":"arXiv:2204.08790"},{"key":"ref96","first-page":"2611","article-title":"The hateful memes challenge: Detecting hate speech in multimodal memes","volume-title":"Proc. NeurIPS","author":"Kiela"},{"key":"ref97","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-00934-2_24"},{"key":"ref98","first-page":"8748","article-title":"Learning transferable visual models from natural language supervision","volume-title":"Proc. ICML","author":"Radford"},{"key":"ref99","doi-asserted-by":"publisher","DOI":"10.1109\/ITSC.2013.6728473"},{"key":"ref100","volume-title":"FER 2013: Kaggle Challenges in Representation Learning Facial Expression Recognition","year":"2013"},{"key":"ref101","article-title":"Learning multiple layers of features from tiny images","author":"Krizhevsky","year":"2009"},{"key":"ref102","doi-asserted-by":"publisher","DOI":"10.1109\/JSTARS.2019.2918242"},{"key":"ref103","doi-asserted-by":"publisher","DOI":"10.1109\/MSP.2012.2211477"},{"key":"ref104","doi-asserted-by":"publisher","DOI":"10.1007\/s11263-009-0275-4"},{"key":"ref105","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2012.6248092"},{"key":"ref106","doi-asserted-by":"publisher","DOI":"10.1109\/IJCNN.2011.6033395"},{"key":"ref107","doi-asserted-by":"publisher","DOI":"10.1109\/JPROC.2017.2675998"},{"key":"ref108","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2014.461"},{"key":"ref109","article-title":"Fine-grained visual classification of aircraft","author":"Maji","year":"2013","journal-title":"arXiv:1306.5151"},{"key":"ref110","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-10599-4_29"},{"key":"ref111","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2004.383"},{"key":"ref112","doi-asserted-by":"publisher","DOI":"10.1109\/ICVGIP.2008.47"},{"key":"ref113","doi-asserted-by":"publisher","DOI":"10.1109\/ICCVW.2013.77"},{"key":"ref114","article-title":"Fixing weight decay regularization in Adam","volume-title":"Proc. ICLR","author":"Loshchilov"},{"key":"ref115","article-title":"Large batch training of convolutional networks","author":"You","year":"2017","journal-title":"arXiv:1708.03888"},{"key":"ref116","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2017.544"},{"key":"ref117","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-01228-1_26"},{"key":"ref118","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-10602-1_48"},{"key":"ref119","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV48922.2021.00986"},{"key":"ref120","article-title":"Self-supervised learning with Swin transformers","author":"Xie","year":"2021","journal-title":"arXiv:2105.04553"}],"container-title":["IEEE Transactions on Neural Networks and Learning Systems"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/5962385\/11073756\/10652984.pdf?arnumber=10652984","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,12,5]],"date-time":"2025-12-05T18:39:43Z","timestamp":1764959983000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/10652984\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,7]]},"references-count":120,"journal-issue":{"issue":"7"},"URL":"https:\/\/doi.org\/10.1109\/tnnls.2024.3443088","relation":{},"ISSN":["2162-237X","2162-2388"],"issn-type":[{"value":"2162-237X","type":"print"},{"value":"2162-2388","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,7]]}}}