{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,7]],"date-time":"2026-08-07T19:47:56Z","timestamp":1786132076626,"version":"build-2736575974"},"reference-count":63,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","issue":"9","license":[{"start":{"date-parts":[[2026,9,1]],"date-time":"2026-09-01T00:00:00Z","timestamp":1788220800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2026,9,1]],"date-time":"2026-09-01T00:00:00Z","timestamp":1788220800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2026,9,1]],"date-time":"2026-09-01T00:00:00Z","timestamp":1788220800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"crossref","award":["62506139"],"award-info":[{"award-number":["62506139"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"crossref"}]},{"name":"Research Grants Council of Hong Kong SAR","award":["12200725"],"award-info":[{"award-number":["12200725"]}]},{"DOI":"10.13039\/501100021171","name":"Basic and Applied Basic Research Foundation of Guangdong Province","doi-asserted-by":"crossref","award":["2026B1515020071"],"award-info":[{"award-number":["2026B1515020071"]}],"id":[{"id":"10.13039\/501100021171","id-type":"DOI","asserted-by":"crossref"}]},{"DOI":"10.13039\/501100017607","name":"Shenzhen Municipal Fundamental Research Program","doi-asserted-by":"crossref","award":["JCYJ20241202124430041"],"award-info":[{"award-number":["JCYJ20241202124430041"]}],"id":[{"id":"10.13039\/501100017607","id-type":"DOI","asserted-by":"crossref"}]},{"name":"CCF\u2013DiDi GAIA Collaborative Research Funds","award":["CCF-DiDi GAIA 202508"],"award-info":[{"award-number":["CCF-DiDi GAIA 202508"]}]},{"name":"Hefei Municipal Natural Science Foundation","award":["HZR2431"],"award-info":[{"award-number":["HZR2431"]}]},{"name":"CAAI\u2013MindSpore Open Fund"},{"name":"Australian Government through the ARC Discovery Projects","award":["DP240102088"],"award-info":[{"award-number":["DP240102088"]}]},{"name":"Australian Research Council Future Fellowship","award":["FT220100318"],"award-info":[{"award-number":["FT220100318"]}]},{"name":"Science and Technology Development Fund of Macao SAR\u2014International Collaborative Research","award":["0001\/2025\/AIJ"],"award-info":[{"award-number":["0001\/2025\/AIJ"]}]},{"name":"Science and Technology Development Fund of Macao","award":["0054\/2025\/RIB2"],"award-info":[{"award-number":["0054\/2025\/RIB2"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. Pattern Anal. Mach. Intell."],"published-print":{"date-parts":[[2026,9]]},"DOI":"10.1109\/tpami.2026.3684741","type":"journal-article","created":{"date-parts":[[2026,4,16]],"date-time":"2026-04-16T19:52:59Z","timestamp":1776369179000},"page":"10235-10247","source":"Crossref","is-referenced-by-count":0,"title":["Strength-Adaptive Adversarial Training"],"prefix":"10.1109","volume":"48","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-8182-6570","authenticated-orcid":false,"given":"Chaojian","family":"Yu","sequence":"first","affiliation":[{"name":"The University of Sydney","place":["Sydney, NSW, Australia"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-0694-3603","authenticated-orcid":false,"given":"Dawei","family":"Zhou","sequence":"additional","affiliation":[{"name":"City University of Macau, Macao Special Administrative Region","place":["Macao, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5659-3464","authenticated-orcid":false,"given":"Li","family":"Shen","sequence":"additional","affiliation":[{"name":"Sun Yat-sen University","place":["Shenzhen, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3197-8103","authenticated-orcid":false,"given":"Jun","family":"Yu","sequence":"additional","affiliation":[{"name":"University of Science and Technology of China","place":["Hefei, China"],"department":["Department of Automation"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6338-0958","authenticated-orcid":false,"given":"Bo","family":"Han","sequence":"additional","affiliation":[{"name":"Hong Kong Baptist University","department":["Department of Computer Science"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7147-5589","authenticated-orcid":false,"given":"Mingming","family":"Gong","sequence":"additional","affiliation":[{"name":"University of Melbourne","place":["Melbourne, VIC, Australia"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4695-6134","authenticated-orcid":false,"given":"Nannan","family":"Wang","sequence":"additional","affiliation":[{"name":"Xidian University","place":["Xi\u2019an, China"]}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-9640-6472","authenticated-orcid":false,"given":"Tongliang","family":"Liu","sequence":"additional","affiliation":[{"name":"The University of Sydney","place":["Sydney, NSW, Australia"]}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.90"},{"key":"ref2","article-title":"Residual convolutional ctc networks for automatic speech recognition","author":"Wang","year":"2017"},{"key":"ref3","article-title":"BERT: Pre-training of deep bidirectional transformers for language understanding","author":"Devlin","year":"2018"},{"key":"ref4","article-title":"Intriguing properties of neural networks","author":"Szegedy","year":"2013"},{"key":"ref5","article-title":"Explaining and harnessing adversarial examples","author":"Goodfellow","year":"2014"},{"key":"ref6","article-title":"Towards deep learning models resistant to adversarial attacks","author":"Madry","year":"2017"},{"key":"ref7","first-page":"274","article-title":"Obfuscated gradients give a false sense of security: Circumventing defenses to adversarial examples","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Athalye"},{"key":"ref8","first-page":"125","article-title":"Adversarial examples are not bugs, they are features","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Ilyas"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1017\/CBO9780511569920.003"},{"key":"ref10","first-page":"7472","article-title":"Theoretically principled trade-off between robustness and accuracy","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Zhang"},{"key":"ref11","first-page":"8909","article-title":"Improving adversarial robustness requires revisiting misclassified examples","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Wang"},{"key":"ref12","article-title":"Unlabeled data improves adversarial robustness","author":"Carmon","year":"2019"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.24963\/ijcai.2018\/520"},{"key":"ref14","article-title":"On the convergence and robustness of adversarial training","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Wang"},{"key":"ref15","first-page":"12835","article-title":"Towards defending against adversarial examples via attack-invariant features","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Zhou"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV48922.2021.00778"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.1007\/s11263-025-02467-7"},{"key":"ref18","article-title":"Geometry-aware instance-reweighted adversarial training","author":"Zhang","year":"2020"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP49357.2023.10095507"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2025.3540200"},{"key":"ref21","first-page":"11278","article-title":"Attacks which do not kill training make adversarial learning stronger","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Zhang"},{"key":"ref22","article-title":"MMA training: Direct input space margin maximization through adversarial training","author":"Ding","year":"2018"},{"key":"ref23","article-title":"Instance adaptive adversarial training: Improved accuracy tradeoffs in neural nets","author":"Balaji","year":"2019"},{"key":"ref24","article-title":"Improving adversarial robustness through progressive hardening","author":"Sitawarin","year":"2020"},{"key":"ref25","article-title":"CAT: Customized adversarial training for improved robustness","author":"Cheng","year":"2020"},{"key":"ref26","article-title":"Improving adversarial training using vulnerability-aware perturbation budget","author":"Fakorede","year":"2024"},{"key":"ref27","first-page":"4072","article-title":"On certifying non-uniform bounds against adversarial attacks","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Liu"},{"key":"ref28","article-title":"Robustness may be at odds with accuracy","author":"Tsipras","year":"2018"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-01258-8_39"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2019.00714"},{"key":"ref31","first-page":"8588","article-title":"A closer look at accuracy vs. robustness","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Yang"},{"key":"ref32","article-title":"Understanding and mitigating the tradeoff between robustness and accuracy","author":"Raghunathan","year":"2020"},{"key":"ref33","article-title":"Adversarial robustness may be at odds with simplicity","author":"Nakkiran","year":"2019"},{"key":"ref34","first-page":"8093","article-title":"Overfitting in adversarially robust deep learning","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Rice"},{"key":"ref35","first-page":"5014","article-title":"Adversarially robust generalization requires more data","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Schmidt"},{"key":"ref36","first-page":"12214","article-title":"Are labels required for improving adversarial robustness?","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Alayrac"},{"key":"ref37","article-title":"Adversarially robust generalization just requires more unlabeled data","author":"Zhai","year":"2019"},{"key":"ref38","first-page":"23258","article-title":"Probabilistic margins for instance reweighting in adversarial training","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Liu"},{"key":"ref39","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV48922.2021.01613"},{"key":"ref40","article-title":"Averaging weights leads to wider optima and better generalization","author":"Izmailov","year":"2018"},{"key":"ref41","first-page":"7910","article-title":"Exploring memorization in adversarial training","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Dong"},{"key":"ref42","article-title":"Distilling the knowledge in a neural network","author":"Hinton","year":"2015"},{"key":"ref43","first-page":"7229","article-title":"Robust overfitting may be mitigated by properly learned smoothening","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Chen"},{"key":"ref44","first-page":"6391","article-title":"Visualizing the loss landscape of neural nets","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Li"},{"key":"ref45","article-title":"Adversarial weight perturbation helps robust generalization","author":"Wu","year":"2020"},{"key":"ref46","article-title":"Standard-deviation-inspired regularization for improving adversarial robustness","author":"Fakorede","year":"2024","journal-title":"Trans. Mach. Learn. Res."},{"key":"ref47","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v36i8.20817"},{"key":"ref48","article-title":"Data augmentation alone can improve adversarial training","author":"Li","year":"2023"},{"key":"ref49","first-page":"29935","article-title":"Data augmentation can improve robustness","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Rebuffi"},{"key":"ref50","first-page":"228","article-title":"Pac-learning in the presence of adversaries","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Cullina"},{"key":"ref51","first-page":"7085","article-title":"Rademacher complexity for adversarially robust generalization","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Yin"},{"issue":"356","key":"ref52","first-page":"1","article-title":"On the impact of hard adversarial instances on overfitting in adversarial training","volume":"25","author":"Liu","year":"2024","journal-title":"J. Mach. Learn. Res."},{"key":"ref53","doi-asserted-by":"publisher","DOI":"10.52202\/068431-1276"},{"key":"ref54","first-page":"21476","article-title":"On the loss landscape of adversarial training: Identifying challenges and how to overcome them","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Liu"},{"key":"ref55","doi-asserted-by":"publisher","DOI":"10.1016\/j.patcog.2022.109229"},{"key":"ref56","doi-asserted-by":"publisher","DOI":"10.24963\/ijcai.2022\/512"},{"key":"ref57","first-page":"25595","article-title":"Understanding robust overfitting of adversarial training and beyond","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Yu"},{"key":"ref58","doi-asserted-by":"publisher","DOI":"10.52202\/075280-0694"},{"key":"ref59","article-title":"Learning multiple layers of features from tiny images","author":"Krizhevsky","year":"2009"},{"key":"ref60","doi-asserted-by":"publisher","DOI":"10.5244\/C.30.87"},{"key":"ref61","first-page":"2206","article-title":"Reliable evaluation of adversarial robustness with an ensemble of diverse parameter-free attacks","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Croce"},{"key":"ref62","first-page":"2196","article-title":"Minimally distorted adversarial examples with a fast adaptive boundary attack","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Croce"},{"key":"ref63","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-58592-1_29"}],"container-title":["IEEE Transactions on Pattern Analysis and Machine Intelligence"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/34\/11643729\/11482524.pdf?arnumber=11482524","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,8,7]],"date-time":"2026-08-07T19:11:31Z","timestamp":1786129891000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11482524\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9]]},"references-count":63,"journal-issue":{"issue":"9"},"URL":"https:\/\/doi.org\/10.1109\/tpami.2026.3684741","relation":{},"ISSN":["0162-8828","2160-9292","1939-3539"],"issn-type":[{"value":"0162-8828","type":"print"},{"value":"2160-9292","type":"electronic"},{"value":"1939-3539","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9]]}}}