{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,20]],"date-time":"2026-07-20T04:03:14Z","timestamp":1784520194367,"version":"3.55.0"},"reference-count":49,"publisher":"Springer Science and Business Media LLC","issue":"13","license":[{"start":{"date-parts":[[2026,6,29]],"date-time":"2026-06-29T00:00:00Z","timestamp":1782691200000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2026,6,29]],"date-time":"2026-06-29T00:00:00Z","timestamp":1782691200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"name":"Tailor","award":["952215"],"award-info":[{"award-number":["952215"]}]},{"name":"Synergies","award":["101146542"],"award-info":[{"award-number":["101146542"]}]},{"name":"Surf Snellius","award":["EINF-4569"],"award-info":[{"award-number":["EINF-4569"]}]},{"name":"Turkish MoNE"},{"name":"ELLIOT","award":["101214398"],"award-info":[{"award-number":["101214398"]}]}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["Neural Comput &amp; Applic"],"published-print":{"date-parts":[[2026,7]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>\n                    Continual Learning (CL) addresses the challenge of enabling models to adapt to evolving data and tasks while retaining previously acquired knowledge. The main challenge in this paradigm is catastrophic forgetting, where models lose prior knowledge upon learning new tasks. While much of the CL literature has focused on model-centric innovations, we argue for the substantial potential of a data-centric approach, specifically by revisiting the \u2018learn-it-all\u2019 assumption prevalent in current CL paradigms. This paper presents the first empirical study systematically evaluating the impact of different coreset methods for training samples in combination with CL methods. Unlike conventional uses of coreset selection, which restrict its role to populating a small rehearsal buffer, we present the first empirical study showing that coreset methods can be applied directly to the full training set, substantially reducing the amount of data needed for learning. Our results reveal that training on carefully selected coreset substantially enhances incremental accuracy while reducing computational overhead. We demonstrate that this performance improvement is primarily driven by an improved stability-plasticity trade-off, largely attributable to the enhanced retention of prior knowledge. This study not only highlights the significant benefits of data-centric strategies in CL but also advocates for a shift in research focus towards these approaches to stimulate and guide future advancements in the field. Code is available at\n                    <jats:ext-link xmlns:xlink=\"http:\/\/www.w3.org\/1999\/xlink\" xlink:href=\"https:\/\/github.com\/ElifCerenGokYildirim\/Coreset-CL\" ext-link-type=\"uri\">https:\/\/github.com\/ElifCerenGokYildirim\/Coreset-CL<\/jats:ext-link>\n                    .\n                  <\/jats:p>","DOI":"10.1007\/s00521-026-12273-y","type":"journal-article","created":{"date-parts":[[2026,6,29]],"date-time":"2026-06-29T05:19:00Z","timestamp":1782710340000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Coresets are more than replay: a data-centric view of continual learning"],"prefix":"10.1007","volume":"38","author":[{"given":"Elif Ceren Gok","family":"Yildirim","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Murat Onur","family":"Yildirim","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Joaquin","family":"Vanschoren","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2026,6,29]]},"reference":[{"key":"12273_CR1","unstructured":"Ven GM, Tolias AS (2019) Three scenarios for continual learning. arXiv preprint: arXiv:1904.07734"},{"key":"12273_CR2","doi-asserted-by":"crossref","unstructured":"McCloskey M, Cohen NJ (1989) Catastrophic interference in connectionist networks: the sequential learning problem. Psychology of Learning and Motivation","DOI":"10.1016\/S0079-7421(08)60536-8"},{"key":"12273_CR3","doi-asserted-by":"publisher","first-page":"118782","DOI":"10.1016\/j.neuroimage.2021.118782","volume":"246","author":"MF Pagnotta","year":"2022","unstructured":"Pagnotta MF, Pascucci D, Plomp G (2022) Selective attention involves a feature-specific sequential release from inhibitory gating. Neuroimage 246:118782, 118782","journal-title":"Neuroimage"},{"key":"12273_CR4","doi-asserted-by":"publisher","first-page":"33","DOI":"10.1016\/j.bandc.2015.12.005","volume":"102","author":"L Jones","year":"2016","unstructured":"Jones L, Hills P, Dick K, Jones S, Bright P (2016) Cognitive mechanisms associated with auditory sensory gating. Brain Cogn 102:33\u201345","journal-title":"Brain Cogn"},{"key":"12273_CR5","doi-asserted-by":"publisher","DOI":"10.1146\/annurev.ne.13.030190.000325","author":"MI Posner","year":"1990","unstructured":"Posner MI, Petersen SE (1990) The attention system of the human brain. Annu Rev Neurosci. https:\/\/doi.org\/10.1146\/annurev.ne.13.030190.000325","journal-title":"Annu Rev Neurosci"},{"key":"12273_CR6","unstructured":"Toneva M, Sordoni A, Combes RTD, Trischler A, Bengio Y, Gordon GJ (2018) An empirical study of example forgetting during deep neural network learning. arXiv preprint: arXiv:1812.05159"},{"key":"12273_CR7","doi-asserted-by":"crossref","unstructured":"Guo C, Zhao B, Bai Y (2022) Deepcore: A comprehensive library for coreset selection in deep learning. International Conference on Database and Expert Systems Applications","DOI":"10.1007\/978-3-031-12423-5_14"},{"key":"12273_CR8","unstructured":"Coleman C, Yeh C, Mussmann S, Mirzasoleiman B, Bailis P, Liang P, Leskovec J, Zaharia M (2019) Selection via proxy: efficient data selection for deep learning. arXiv preprint: arXiv:1906.11829"},{"key":"12273_CR9","unstructured":"Paul M, Ganguli S, Dziugaite GK (2021) Deep learning on a data diet: finding important examples early in training. NeurIPS"},{"key":"12273_CR10","doi-asserted-by":"crossref","unstructured":"Welling M (2009) Herding dynamical weights to learn. In: ICML","DOI":"10.1145\/1553374.1553517"},{"key":"12273_CR11","unstructured":"Coleman C, Yeh C, Mussmann S, Mirzasoleiman B, Bailis P, Liang P, Leskovec J, Zaharia M (2019) Selection via proxy: efficient data selection for deep learning. arXiv preprint: arXiv:1906.11829"},{"key":"12273_CR12","unstructured":"Iyer R, Khargoankar N, Bilmes J, Asanani H (2021) Submodular combinatorial information measures with applications in machine learning. In: Algorithmic Learning Theory"},{"key":"12273_CR13","unstructured":"Mirzasoleiman B, Bilmes J, Leskovec J (2020) Coresets for data-efficient training of machine learning models. In: ICML"},{"key":"12273_CR14","unstructured":"Aljundi R, Lin M, Goujaud B, Bengio Y (2019) Gradient based sample selection for online continual learning. NeurIPS"},{"key":"12273_CR15","unstructured":"Borsos Z, Mutny M, Krause A (2020) Coresets via bilevel optimization for continual learning and streaming. NeurIPS"},{"issue":"13","key":"12273_CR16","doi-asserted-by":"publisher","first-page":"3521","DOI":"10.1073\/pnas.1611835114","volume":"114","author":"J Kirkpatrick","year":"2017","unstructured":"Kirkpatrick J, Pascanu R, Rabinowitz N, Veness J, Desjardins G, Rusu AA, Milan K, Quan J, Ramalho T, Grabska-Barwinska A et al (2017) Overcoming catastrophic forgetting in neural networks. PNAS 114(13):3521\u20133526","journal-title":"PNAS"},{"key":"12273_CR17","unstructured":"Lee S-W, Kim J-H, Jun J, Ha J-W, Zhang B-T (2017) Overcoming catastrophic forgetting by incremental moment matching. NeurIPS"},{"issue":"12","key":"12273_CR18","doi-asserted-by":"publisher","first-page":"2935","DOI":"10.1109\/TPAMI.2017.2773081","volume":"40","author":"Z Li","year":"2017","unstructured":"Li Z, Hoiem D (2017) Learning without forgetting. IEEE Trans Pattern Anal Mach Intell 40(12):2935\u20132947","journal-title":"IEEE Trans Pattern Anal Mach Intell"},{"key":"12273_CR19","unstructured":"Chaudhry A, Ranzato M, Rohrbach M, Elhoseiny M (2018) Efficient lifelong learning with a-gem. arXiv preprint: arXiv:1812.00420"},{"key":"12273_CR20","unstructured":"Lopez-Paz D, Ranzato M (2017) Gradient episodic memory for continual learning. NeurIPS, 30"},{"key":"12273_CR21","unstructured":"Rolnick D, Ahuja A, Schwarz J, Lillicrap T, Wayne G (2019) Experience replay for continual learning. NeurIPS, 32"},{"key":"12273_CR22","doi-asserted-by":"crossref","unstructured":"Rebuffi S-A, Kolesnikov A, Sperl G, Lampert CH (2017) icarl: incremental classifier and representation learning. In: CVPR","DOI":"10.1109\/CVPR.2017.587"},{"key":"12273_CR23","unstructured":"Yoon J, Kim S, Yang E, Hwang SJ (2019) Scalable and order-robust continual learning with additive parameter decomposition. arXiv preprint: arXiv:1902.09432"},{"key":"12273_CR24","unstructured":"Yildirim MO, Yildirim ECG, Mocanu DC, Vanschoren J (2025) Self-regulated neurogenesis for online data-incremental learning. arXiv preprint: arXiv:2403.14684"},{"key":"12273_CR25","unstructured":"Zhou D-W, Wang Q-W, Ye H-J, Zhan D-C (2022) A model or 603 exemplars: towards memory-efficient class-incremental learning. arXiv preprint: arXiv:2205.13218"},{"key":"12273_CR26","doi-asserted-by":"crossref","unstructured":"Yan S et al. (2021) Der: Dynamically expandable representation for class incremental learning. In: CVPR","DOI":"10.1109\/CVPR46437.2021.00303"},{"key":"12273_CR27","doi-asserted-by":"crossref","unstructured":"Wang F-Y, Zhou D-W, Ye H-J, Zhan D-C (2022) Foster: feature boosting and compression for class-incremental learning. In: ECCV","DOI":"10.1007\/978-3-031-19806-9_23"},{"key":"12273_CR28","doi-asserted-by":"crossref","unstructured":"Wang Z et al. (2022) Dualprompt: Complementary prompting for rehearsal-free continual learning. In: ECCV","DOI":"10.1007\/978-3-031-19809-0_36"},{"key":"12273_CR29","doi-asserted-by":"crossref","unstructured":"Wang Z, Zhang Z, Lee C-Y, Zhang H, Sun R, Ren X, Su G, Perot V, Dy J, Pfister T (2022) Learning to prompt for continual learning. In: CVPR","DOI":"10.1109\/CVPR52688.2022.00024"},{"key":"12273_CR30","doi-asserted-by":"crossref","unstructured":"Smith JS, Karlinsky L, Gutta V, Cascante-Bonilla P, Kim D, Arbelle A, Panda R, Feris R, Kira Z (2023) Coda-prompt: continual decomposed attention-based prompting for rehearsal-free continual learning. In: CVPR","DOI":"10.1109\/CVPR52729.2023.01146"},{"issue":"3","key":"12273_CR31","doi-asserted-by":"publisher","first-page":"1012","DOI":"10.1007\/s11263-024-02218-0","volume":"133","author":"D-W Zhou","year":"2025","unstructured":"Zhou D-W, Cai Z-W, Ye H-J, Zhan D-C, Liu Z (2025) Revisiting class-incremental learning with pre-trained models: generalizability and adaptivity are all you need. Int J Comput Vision 133(3):1012\u20131032","journal-title":"Int J Comput Vision"},{"key":"12273_CR32","unstructured":"Yildirim MO et al. (2025) Sculpting [cls] features for pre-trained model-based class-incremental learning. arXiv preprint: arXiv:2502.14762"},{"issue":"5","key":"12273_CR33","doi-asserted-by":"publisher","first-page":"6893","DOI":"10.1109\/TNNLS.2022.3213169","volume":"35","author":"A Maalouf","year":"2022","unstructured":"Maalouf A, Eini G, Mussay B, Feldman D, Osadchy M (2022) A unified approach to coreset learning. IEEE Trans Neural Netw Learn Syst 35(5):6893\u20136905","journal-title":"IEEE Trans Neural Netw Learn Syst"},{"key":"12273_CR34","doi-asserted-by":"publisher","DOI":"10.1007\/s13218-017-0519-3","author":"A Munteanu","year":"2018","unstructured":"Munteanu A, Schwiegelshohn C (2018) Coresets-methods and history: a theoreticians design pattern for approximation and streaming algorithms. KI - K\u00fcnstliche Intelligenz. https:\/\/doi.org\/10.1007\/s13218-017-0519-3","journal-title":"KI - K\u00fcnstliche Intelligenz"},{"key":"12273_CR35","unstructured":"Wei K, Iyer R, Bilmes J (2015) Submodularity in data subset selection and active learning. In: ICML"},{"key":"12273_CR36","unstructured":"Sener O, Savarese S (2017) Active learning for convolutional neural networks: a core-set approach. arXiv preprint: arXiv:1708.00489"},{"key":"12273_CR37","unstructured":"Yoon J, Madaan D, Yang E, Hwang SJ (2022) Online coreset selection for rehearsal-based continual learning. In: ICLR"},{"key":"12273_CR38","volume-title":"Learning multiple layers of features from tiny images","author":"A Krizhevsky","year":"2009","unstructured":"Krizhevsky A, Hinton G et al (2009) Learning multiple layers of features from tiny images. ON, Canada, Toronto"},{"issue":"3","key":"12273_CR39","doi-asserted-by":"publisher","first-page":"211","DOI":"10.1007\/s11263-015-0816-y","volume":"115","author":"Oea Russakovsky","year":"2015","unstructured":"Russakovsky Oea et al (2015) Imagenet large scale visual recognition challenge. International Journal of Computer Vision 115(3):211\u2013252","journal-title":"International Journal of Computer Vision"},{"key":"12273_CR40","doi-asserted-by":"publisher","DOI":"10.1007\/s11432-022-3600-y","author":"D-W Zhou","year":"2023","unstructured":"Zhou D-W, Wang F-Y, Ye H-J, Zhan D-C (2023) Pycil: a Python toolbox for class-incremental learning. Sci China Inf Sci. https:\/\/doi.org\/10.1007\/s11432-022-3600-y","journal-title":"Sci China Inf Sci"},{"key":"12273_CR41","doi-asserted-by":"crossref","unstructured":"Nguyen T, Bui M, Ninh H, Nguyen T, Tran HT (2022) Efficient heuristic algorithm to speed up graphcut in gpu for image stitching. In: IEEE PAAP","DOI":"10.1109\/PAAP56126.2022.10010453"},{"key":"12273_CR42","doi-asserted-by":"crossref","unstructured":"Ekstr\u00f6m Sea et al (2020) Fast graph-cut based optimization for practical dense deformable registration of volume images. Comput Med Imaging Graph 84:101745, 101745","DOI":"10.1016\/j.compmedimag.2020.101745"},{"key":"12273_CR43","doi-asserted-by":"crossref","unstructured":"Feldman D, Langberg M (2011) A unified framework for approximating and clustering data. In: ACM Symposium on Theory of Computing","DOI":"10.1145\/1993636.1993712"},{"issue":"5","key":"12273_CR44","first-page":"6893","volume":"35","author":"A Maalouf","year":"2022","unstructured":"Maalouf A, Eini G, Mussay B, Feldman D, Osadchy M (2022) A unified approach to coreset learning. IEEE TNNLS 35(5):6893\u20136905","journal-title":"IEEE TNNLS"},{"key":"12273_CR45","doi-asserted-by":"crossref","unstructured":"He K et al. (2016) Deep residual learning for image recognition. CVPR,","DOI":"10.1109\/CVPR.2016.90"},{"key":"12273_CR46","unstructured":"Dosovitskiy A, Beyer L, et al. An image is worth 16x16 words: transformers for image recognition at scale. arXiv preprint: arXiv:2010.11929 (2021)"},{"key":"12273_CR47","doi-asserted-by":"crossref","unstructured":"Sun H-L, Zhou D-W, Zhao H, Gan L, Zhan D-C, Ye H-J (2025) Mos: model surgery for pre-trained model-based class-incremental learning. AAAI","DOI":"10.1609\/aaai.v39i19.34281"},{"key":"12273_CR48","doi-asserted-by":"crossref","unstructured":"Zhang G, Wang L, Kang G, Chen L, Wei Y (2023) Slca: slow learner with classifier alignment for continual learning on a pre-trained model. In: ICCV","DOI":"10.1109\/ICCV51070.2023.01754"},{"key":"12273_CR49","first-page":"69054","volume":"36","author":"L Wang","year":"2023","unstructured":"Wang L (2023) Hierarchical decomposition of prompt-based continual learning: rethinking obscured sub-optimality. NeurIPS 36:69054\u201369076","journal-title":"NeurIPS"}],"container-title":["Neural Computing and Applications"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s00521-026-12273-y.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1007\/s00521-026-12273-y","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s00521-026-12273-y.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,20]],"date-time":"2026-07-20T03:38:36Z","timestamp":1784518716000},"score":1,"resource":{"primary":{"URL":"https:\/\/link.springer.com\/10.1007\/s00521-026-12273-y"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,6,29]]},"references-count":49,"journal-issue":{"issue":"13","published-print":{"date-parts":[[2026,7]]}},"alternative-id":["12273"],"URL":"https:\/\/doi.org\/10.1007\/s00521-026-12273-y","relation":{},"ISSN":["0941-0643","1433-3058"],"issn-type":[{"value":"0941-0643","type":"print"},{"value":"1433-3058","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,6,29]]},"assertion":[{"value":"29 September 2025","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"14 May 2026","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"29 June 2026","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"The authors declare that they have no financial or non-financial competing interests.","order":1,"name":"Ethics","label":"Conflicts of interests","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"This study did not involve human participants or animals; therefore, ethics approval, informed consent, and animal welfare statements are not applicable.","order":2,"name":"Ethics","label":"Ethics approval and consent to participate","group":{"name":"EthicsHeading","label":"Declarations"}}],"article-number":"559"}}