{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,23]],"date-time":"2026-10-23T20:09:15Z","timestamp":1792786155192,"version":"4.5.0"},"reference-count":47,"publisher":"The Royal Society","issue":"243","license":[{"start":{"date-parts":[[2026,10,21]],"date-time":"2026-10-21T00:00:00Z","timestamp":1792540800000},"content-version":"vor","delay-in-days":14,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"name":"Human Frontiers Science Program","award":["RGP006\/2024"],"award-info":[{"award-number":["RGP006\/2024"]}],"id":[{"id":"https:\/\/ror.org\/02ebx7v45","id-type":"ROR","asserted-by":"crossref"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026,10,7]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>Palaeontology has seen widespread and growing use of machine learning (ML) to classify and analyse large datasets of fossils. However, palaeontology is a challenging field in which to apply ML. Datasets may be small or unlabelled, images may be complex and different from standard datasets and palaeontologists may lack specialist training and access to necessary computational resources. We show how these challenges can be addressed by using recent developments in self-supervised learning (SSL). Using a frozen DINOv3 feature extractor and a simple linear classifier, with reduced data, we can achieve comparable results to literature benchmarks using convolutional neural networks, the previous standard, when classifying fossil tracks, pollen, radiolaria, foraminifera and a dataset of diverse fossil images. In addition, the rich feature vectors generated by the model can be used for few-shot learning, unsupervised clustering, subject to the presence of imaging artefacts and quantification of disparity. Using state-of-the-art self-supervised methods increases accessibility by reducing code, compute and data required. It also maintains accuracy, while increasing reproducibility by reducing parameters.<\/jats:p>","DOI":"10.1098\/rsif.2025.1243","type":"journal-article","created":{"date-parts":[[2026,10,21]],"date-time":"2026-10-21T00:16:59Z","timestamp":1792541819000},"source":"Crossref","is-referenced-by-count":0,"title":["Complexity begets simplicity: self-supervised learning for palaeontological images with few or no labels"],"prefix":"10.1098","volume":"23","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-3936-2709","authenticated-orcid":false,"given":"Niall","family":"Rodgers","sequence":"first","affiliation":[{"name":"UK Centre for Astrobiology, School of Physics and Astronomy, The University of Edinburgh , Edinburgh, Midlothian, UK"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"175","published-online":{"date-parts":[[2026,10,21]]},"reference":[{"key":"2026102020164544200_bib1","doi-asserted-by":"crossref","first-page":"104765","DOI":"10.1016\/j.earscirev.2024.104765","article-title":"Artificial intelligence in paleontology","volume":"252","author":"Yu","year":"2024","journal-title":"Earth-Sci. Rev."},{"key":"2026102020164544200_bib2","doi-asserted-by":"crossref","first-page":"83","DOI":"10.1007\/s10462-024-11080-y","article-title":"Advancing paleontology: a survey on deep learning methodologies in fossil image analysis","volume":"58","author":"Yaqoob","year":"2025","journal-title":"Artif. Intell. Rev."},{"key":"2026102020164544200_bib3","doi-asserted-by":"crossref","first-page":"103329","DOI":"10.1016\/j.ecoinf.2025.103329","article-title":"A review of machine learning applications for identification and classification problems in paleontology","volume":"91","author":"Marques","year":"2025","journal-title":"Ecol. Inform."},{"key":"2026102020164544200_bib4","article-title":"DINOv3","author":"Sim\u00e9oni","year":"2025"},{"key":"2026102020164544200_bib5","first-page":"9630","article-title":"Emerging properties in self-supervised vision transformers","volume-title":"Proc. IEEE Int. Conf. Comput. Vis.","author":"Caron","year":"2021"},{"key":"2026102020164544200_bib6","article-title":"DINOv2: learning robust visual features without supervision","author":"Oquab","year":"2024"},{"key":"2026102020164544200_bib7","doi-asserted-by":"crossref","first-page":"9052","DOI":"10.1109\/TPAMI.2024.3415112","article-title":"A survey on self-supervised learning: algorithms, applications, and future trends","volume":"46","author":"Gui","year":"2024","journal-title":"IEEE Trans. Pattern. Anal. Mach. Intell."},{"key":"2026102020164544200_bib8","article-title":"A cookbook of self-supervised learning","author":"Balestriero","year":"2023"},{"key":"2026102020164544200_bib9","article-title":"HuggingFace's transformers: state-of-the-art natural language processing","author":"Wolf","year":"2020"},{"key":"2026102020164544200_bib10","doi-asserted-by":"crossref","first-page":"1157","DOI":"10.1029\/2019PA003612","article-title":"Endless forams: &gt;34,000 modern planktonic foraminiferal images for taxonomic training and automated species recognition using convolutional neural networks","volume":"34","author":"Hsiang","year":"2019","journal-title":"Paleoceanogr. Paleoclimatol."},{"key":"2026102020164544200_bib11","doi-asserted-by":"crossref","first-page":"183","DOI":"10.5194\/jm-39-183-2020","article-title":"Automated analysis of foraminifera fossil records by image classification using a convolutional neural network","volume":"39","author":"Marchant","year":"2020","journal-title":"J. Micropalaeontol."},{"key":"2026102020164544200_bib12","doi-asserted-by":"crossref","first-page":"165","DOI":"10.5194\/jm-41-165-2022","article-title":"Artificial intelligence applied to the classification of eight middle Eocene species of the genus Podocyrtis (polycystine radiolaria)","volume":"41","author":"Carlsson","year":"2022","journal-title":"J. Micropalaeontol."},{"key":"2026102020164544200_bib13","doi-asserted-by":"crossref","first-page":"20220588","DOI":"10.1098\/rsif.2022.0588","article-title":"A machine learning approach for the discrimination of theropod and ornithischian dinosaur tracks","volume":"19","author":"Lallensack","year":"2022","journal-title":"J. R. Soc. Interface"},{"key":"2026102020164544200_bib14","doi-asserted-by":"crossref","first-page":"e0229751","DOI":"10.1371\/journal.pone.0229751","article-title":"Precise automatic classification of 46 different pollen types with convolutional neural networks","volume":"15","author":"Sevillano","year":"2020","journal-title":"PLoS ONE"},{"key":"2026102020164544200_bib15","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1017\/pab.2022.14","article-title":"Automatic taxonomic identification based on the fossil image dataset (&gt;415,000 images) and deep convolutional neural networks","volume":"49","author":"Liu","year":"2022","journal-title":"Paleobiology"},{"key":"2026102020164544200_bib16","doi-asserted-by":"crossref","first-page":"34","DOI":"10.1007\/978-3-031-09037-0_4","volume-title":"Pattern recognition and artificial intelligence","author":"Karaderi","year":"2022"},{"key":"2026102020164544200_bib17","doi-asserted-by":"crossref","first-page":"7189","DOI":"10.1038\/s41598-025-90988-z","article-title":"Classifying microfossil radiolarians on fractal pre-trained vision transformers","volume":"15","author":"Mimura","year":"2025","journal-title":"Sci. Rep."},{"key":"2026102020164544200_bib18","doi-asserted-by":"crossref","first-page":"044602","DOI":"10.1088\/2633-1357\/acadb9","article-title":"Morphology exploration of pollen using deep learning latent space","volume":"3","author":"Grant-Jacob","year":"2022","journal-title":"IOP SciNotes"},{"key":"2026102020164544200_bib19","doi-asserted-by":"crossref","first-page":"102476","DOI":"10.1016\/j.marmicro.2025.102476","article-title":"Foraminiferal image classification based on convolutional neural network considering data augmentation optimization","volume":"199","author":"Wang","year":"2025","journal-title":"Marine Micropaleontol."},{"key":"2026102020164544200_bib20","first-page":"4171","article-title":"BERT: Pre-training of deep bidirectional transformers for language understanding","volume-title":"Proc. Conf. North Am. Chapter Assoc. Comput. Linguist. Hum. Lang. Technol.","author":"Devlin","year":"2019"},{"key":"2026102020164544200_bib21","first-page":"1877","article-title":"Language models are few-shot learners","volume-title":"Advances in neural information processing systems","author":"Brown","year":"2020"},{"key":"2026102020164544200_bib22","article-title":"Vision transformers need registers","author":"Darcet","year":"2024"},{"key":"2026102020164544200_bib23","article-title":"Does DINOv3 set a new medical vision standard?","author":"Liu","year":"2025"},{"key":"2026102020164544200_bib24","article-title":"Foundation visual encoders are secretly few-shot anomaly detectors","author":"Zhai","year":"2025"},{"key":"2026102020164544200_bib25","article-title":"Brought a gun to a knife fight: modern VFM baselines outgun specialized detectors on in-the-wild AI image detection","author":"Zhou","year":"2025"},{"key":"2026102020164544200_bib26","doi-asserted-by":"crossref","first-page":"6853","DOI":"10.1016\/j.jrmge.2025.01.057","article-title":"DINOv2 rocks geological image analysis: classification, segmentation, and interpretability","volume":"17","author":"Brondolo","year":"2015","journal-title":"J. Rock Mech. Geotech. Eng."},{"key":"2026102020164544200_bib27","first-page":"21969","article-title":"An evaluation of a pre-trained transformer-based self-distillation model (DINOv2) for cross-domain plant species identification","volume-title":"Neural Comput. Appl.","author":"Ong","year":"2025"},{"key":"2026102020164544200_bib28","doi-asserted-by":"crossref","first-page":"18","DOI":"10.1007\/978-3-031-92387-6_2","article-title":"WildFusion: individual animal identification with\u00c2\u00a0calibrated similarity fusion","volume-title":"Comput. Vis. ECCV 2024 Workshops","author":"Cermak","year":"2025"},{"key":"2026102020164544200_bib29","first-page":"e2025JH000601","article-title":"Cross-domain foundation model adaptation: pioneering computer vision models for geophysical data analysis","volume":"2","author":"Guo","year":"2025","journal-title":"J. Geophys. Res. Mach. Learn. Comput."},{"key":"2026102020164544200_bib30","first-page":"1319","article-title":"AnomalyDINO: boosting patch-based few-shot anomaly detection with DINOv2","volume-title":"Proc. IEEE\/CVF Winter Conf. Appl. Comput. Vis.","author":"Damm","year":"2025"},{"key":"2026102020164544200_bib31","first-page":"2798","article-title":"DINOBot: robot manipulation via retrieval and alignment with vision foundation models","volume-title":"Proc. IEEE Int. Conf. Robot. Autom.","author":"Di Palo","year":"2024"},{"key":"2026102020164544200_bib32","article-title":"Evaluating general purpose vision foundation models for medical image analysis: an experimental study of DINOv2 on radiology benchmarks","author":"Baharoon","year":"2024"},{"key":"2026102020164544200_bib33","first-page":"1","article-title":"Low-resource finetuning of foundation models beats state-of-the-art in histopathology","volume-title":"IEEE Int. Symp. Biomed. Imaging","author":"Roth","year":"2024"},{"key":"2026102020164544200_bib34","doi-asserted-by":"crossref","first-page":"e2022GC010689","DOI":"10.1029\/2022GC010689","article-title":"Forabot: automated planktic foraminifera isolation and imaging","volume":"23","author":"Richmond","year":"2022","journal-title":"Geochem. Geophys. Geosyst."},{"key":"2026102020164544200_bib35","doi-asserted-by":"crossref","first-page":"67\u00a0298","DOI":"10.1109\/ACCESS.2023.3291620","article-title":"ForamViT-GAN: exploring new paradigms in deep learning for micropaleontological image analysis","volume":"11","author":"Ferreira-Chacua","year":"2023","journal-title":"IEEE Access"},{"key":"2026102020164544200_bib36","article-title":"Endless Foram, MD022508 and MD9712138 training datasets","author":"Marchant","year":"2020"},{"key":"2026102020164544200_bib37","first-page":"2825","article-title":"Scikit-learn: machine learning in Python","volume":"12","author":"Pedregosa","year":"2011","journal-title":"J. Mach. Learn. Res."},{"key":"2026102020164544200_bib38","doi-asserted-by":"crossref","first-page":"e0008904","DOI":"10.1371\/journal.pntd.0008904","article-title":"Delimiting cryptic morphological variation among human malaria vector species using convolutional neural networks","volume":"14","author":"Couret","year":"2020","journal-title":"PLoS. Negl. Trop. Dis."},{"key":"2026102020164544200_bib39","doi-asserted-by":"crossref","first-page":"861","DOI":"10.21105\/joss.00861","article-title":"UMAP: uniform manifold approximation and projection","volume":"3","author":"Mcinnes","year":"2018","journal-title":"J. Open Source Softw."},{"key":"2026102020164544200_bib40","doi-asserted-by":"crossref","first-page":"82","DOI":"10.1038\/s43586-024-00363-x","article-title":"Uniform manifold approximation and projection","volume":"4","author":"Healy","year":"2024","journal-title":"Nat. Rev. Methods Primers"},{"key":"2026102020164544200_bib41","first-page":"2579","article-title":"Visualizing data using t-SNE","volume":"9","author":"van der Maaten","year":"2008","journal-title":"J. Mach. Learn. Res."},{"key":"2026102020164544200_bib42","doi-asserted-by":"crossref","first-page":"53","DOI":"10.1016\/0377-0427(87)90125-7","article-title":"Silhouettes: a graphical aid to the interpretation and validation of cluster analysis","volume":"20","author":"Rousseeuw","year":"1987","journal-title":"J. Comput. Appl. Math."},{"key":"2026102020164544200_bib43","doi-asserted-by":"crossref","first-page":"205","DOI":"10.21105\/joss.00205","article-title":"HDBSCAN: hierarchical density based clustering","volume":"2","author":"Mcinnes","year":"2017","journal-title":"J. Open Source Softw."},{"key":"2026102020164544200_bib44","article-title":"Reconstructing the diversity dynamics of paleo-grasslands using deep learning on superresolution images of fossil Poaceae pollen","volume-title":"bioRxiv","author":"Adaim\u00e9","year":"2024"},{"key":"2026102020164544200_bib45","doi-asserted-by":"crossref","first-page":"106911","DOI":"10.1016\/j.marpetgeo.2024.106911","article-title":"A machine learning application for the automatic recognition of planktonic foraminifera in thin sections","volume":"166","author":"Piva","year":"2024","journal-title":"Marine and Petroleum Geol."},{"key":"2026102020164544200_bib46","article-title":"Link to Code Repository for: Complexity Begets Simplicity: Self-Supervised Learning for1 Palaeontological Images with Few or No Labels, published by Niall Rodgers (Version v1.0.0) [Computer software]","author":"nrodgers1","year":"2026"},{"key":"2026102020164544200_bib47","article-title":"Supplementary material from: Complexity begets simplicity: self-supervised learning for palaeontological images with few or no labels","author":"Rodgers","year":"2026"}],"container-title":["Journal of the Royal Society Interface"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/royalsocietypublishing.org\/rsif\/article-pdf\/doi\/10.1098\/rsif.2025.1243\/6172717\/rsif.2025.1243.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/royalsocietypublishing.org\/rsif\/article-pdf\/doi\/10.1098\/rsif.2025.1243\/6172717\/rsif.2025.1243.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,10,21]],"date-time":"2026-10-21T00:17:04Z","timestamp":1792541824000},"score":1,"resource":{"primary":{"URL":"https:\/\/royalsocietypublishing.org\/rsif\/article\/23\/243\/20251243\/483813\/Complexity-begets-simplicity-self-supervised"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,10,7]]},"references-count":47,"journal-issue":{"issue":"243","published-print":{"date-parts":[[2026,10,7]]}},"URL":"https:\/\/doi.org\/10.1098\/rsif.2025.1243","relation":{},"ISSN":["1742-5662"],"issn-type":[{"value":"1742-5662","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,10,7]]},"article-number":"20251243"}}