{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,11,15]],"date-time":"2026-11-15T12:28:41Z","timestamp":1794745721124,"version":"4.9.1"},"reference-count":77,"publisher":"Springer Science and Business Media LLC","issue":"8089","license":[{"start":{"date-parts":[[2025,11,12]],"date-time":"2025-11-12T00:00:00Z","timestamp":1762905600000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2025,11,12]],"date-time":"2025-11-12T00:00:00Z","timestamp":1762905600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["Nature"],"published-print":{"date-parts":[[2025,11,13]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>\n                    Deep neural networks have achieved success across a wide range of applications, including as models of human behaviour and neural representations in vision tasks\n                    <jats:sup>1,2<\/jats:sup>\n                    . However, neural network training and human learning differ in fundamental ways, and neural networks often fail to generalize as robustly as humans do\n                    <jats:sup>3,4<\/jats:sup>\n                    , raising questions regarding the similarity of their underlying representations. We need to determine what is missing for modern learning systems to exhibit more human-aligned behaviour. Here we highlight a key misalignment between vision models and humans: whereas human conceptual knowledge is hierarchically organized from fine- to coarse-scale distinctions (for example, ref.\n                    <jats:sup>5<\/jats:sup>\n                    ), model representations do not accurately capture all these levels of abstraction. To address this misalignment, we first train a teacher model to imitate human judgements, then transfer human-aligned structure from its representations to refine the representations of pretrained state-of-the-art vision foundation models via fine-tuning. These human-aligned models more accurately approximate human behaviour and uncertainty across a wide range of similarity tasks, including a dataset of human judgements spanning multiple levels of semantic abstractions. They also perform better on a diverse set of machine learning tasks, increasing generalization and out-of-distribution robustness. Thus, infusing neural networks with additional human knowledge yields a best-of-both-worlds representation that is both more consistent with human cognitive judgements and more practically useful, paving the way towards more robust, interpretable and human-aligned artificial intelligence systems.\n                  <\/jats:p>","DOI":"10.1038\/s41586-025-09631-6","type":"journal-article","created":{"date-parts":[[2025,11,12]],"date-time":"2025-11-12T16:02:02Z","timestamp":1762963322000},"page":"349-355","update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":32,"title":["Aligning machine and human visual representations across abstraction levels"],"prefix":"10.1038","volume":"647","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-0804-4687","authenticated-orcid":false,"given":"Lukas","family":"Muttenthaler","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Klaus","family":"Greff","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0002-1214-4864","authenticated-orcid":false,"given":"Frieda","family":"Born","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9752-932X","authenticated-orcid":false,"given":"Bernhard","family":"Spitzer","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Simon","family":"Kornblith","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-9654-0575","authenticated-orcid":false,"given":"Michael C.","family":"Mozer","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Klaus-Robert","family":"M\u00fcller","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Thomas","family":"Unterthiner","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6988-8437","authenticated-orcid":false,"given":"Andrew K.","family":"Lampinen","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2025,11,12]]},"reference":[{"key":"9631_CR1","doi-asserted-by":"publisher","first-page":"e1003915","DOI":"10.1371\/journal.pcbi.1003915","volume":"10","author":"S-M Khaligh-Razavi","year":"2014","unstructured":"Khaligh-Razavi, S.-M. & Kriegeskorte, N. Deep supervised, but not unsupervised, models may explain it cortical representation. PLoS Comput. Biol. 10, e1003915 (2014).","journal-title":"PLoS Comput. Biol."},{"key":"9631_CR2","doi-asserted-by":"crossref","unstructured":"Peterson, J. C., Battleday, R. M., Griffiths, T. L. & Russakovsky, O. Human uncertainty makes classification more robust. In Proc. IEEE\/CVF International Conference on Computer Vision (ICCV) 9616\u20139625 (IEEE, 2019).","DOI":"10.1109\/ICCV.2019.00971"},{"key":"9631_CR3","doi-asserted-by":"publisher","first-page":"e253","DOI":"10.1017\/S0140525X16001837","volume":"40","author":"BM Lake","year":"2017","unstructured":"Lake, B. M., Ullman, T. D., Tenenbaum, J. B. & Gershman, S. J. Building machines that learn and think like people. Behav. Brain Sci. 40, e253 (2017).","journal-title":"Behav. Brain Sci."},{"key":"9631_CR4","unstructured":"Geirhos, R. et al. Generalisation in humans and deep neural networks. In Proc. Advances in Neural Information Processing Systems Vol. 31 (eds Bengio, S. et al.) 7538\u20137550 (Curran Associates, 2018)."},{"key":"9631_CR5","doi-asserted-by":"publisher","first-page":"1173","DOI":"10.1038\/s41562-020-00951-3","volume":"4","author":"MN Hebart","year":"2020","unstructured":"Hebart, M. N., Zheng, C. Y., Pereira, F. & Baker, C. I. Revealing the multidimensional mental representations of natural objects underlying human similarity judgements. Nat. Hum. Behav. 4, 1173\u20131185 (2020).","journal-title":"Nat. Hum. Behav."},{"key":"9631_CR6","unstructured":"Brown, T. et al. Language models are few-shot learners. In Proc. Advances in Neural Information Processing Systems Vol. 33 (eds Larochelle, H. et al.) 1877\u20131901 (Curran Associates, 2020)."},{"key":"9631_CR7","unstructured":"Radford, A. et al. Learning transferable visual models from natural language supervision. In Proc. 38th Interna- tional Conference on Machine Learning: Proc. Machine Learning Research Vol. 139 (eds Meila, M. & Zhang, T.) 8748\u20138763 (PMLR, 2021)."},{"key":"9631_CR8","unstructured":"Dosovitskiy, A. et al. An image is worth 16\u00d716 words: transformers for image recognition at scale. In Proc. The Ninth International Conference on Learning Representations (ICLR, 2021)."},{"key":"9631_CR9","doi-asserted-by":"crossref","unstructured":"Zhai, X., Mustafa, B., Kolesnikov, A. & Beyer, L. Sigmoid Loss for language image pre-training. In Proc. IEEE\/CVF International Conference on Computer Vision 1975\u201311986 (IEEE, 2023).","DOI":"10.1109\/ICCV51070.2023.01100"},{"key":"9631_CR10","doi-asserted-by":"publisher","DOI":"10.1038\/s41467-019-08987-4","volume":"10","author":"S Lapuschkin","year":"2019","unstructured":"Lapuschkin, S. et al. Unmasking clever Hans predictors and assessing what machines really learn. Nat. Commun. 10, 1096 (2019).","journal-title":"Nat. Commun."},{"key":"9631_CR11","doi-asserted-by":"publisher","first-page":"665","DOI":"10.1038\/s42256-020-00257-z","volume":"2","author":"R Geirhos","year":"2020","unstructured":"Geirhos, R. et al. Shortcut learning in deep neural networks. Nat. Mach. Intell. 2, 665\u2013673 (2020).","journal-title":"Nat. Mach. Intell."},{"key":"9631_CR12","doi-asserted-by":"crossref","unstructured":"Bowers, J. S. et al. Deep problems with neural network models of human vision. Behav. Brain Sci. 46, e385 (2022).","DOI":"10.1017\/S0140525X22002813"},{"key":"9631_CR13","unstructured":"Muttenthaler, L. et al. Improving neural network representations using human similarity judgments. In Proc. Advances in Neural Information Processing Systems Vol. 36 (eds Oh, A. et al) 50978\u201351007 (Curran Associates, 2023)."},{"key":"9631_CR14","doi-asserted-by":"publisher","first-page":"3","DOI":"10.1016\/0010-0277(88)90031-5","volume":"28","author":"JA Fodor","year":"1988","unstructured":"Fodor, J. A. & Pylyshyn, Z. W. Connectionism and cognitive architecture: a critical analysis. Cognition 28, 3\u201371 (1988).","journal-title":"Cognition"},{"key":"9631_CR15","doi-asserted-by":"publisher","first-page":"1279","DOI":"10.1126\/science.1192788","volume":"331","author":"JB Tenenbaum","year":"2011","unstructured":"Tenenbaum, J. B., Kemp, C., Griffiths, T. L. & Goodman, N. D. How to grow a mind: statistics, structure, and abstraction. Science 331, 1279\u20131285 (2011).","journal-title":"Science"},{"key":"9631_CR16","doi-asserted-by":"publisher","first-page":"1332","DOI":"10.1126\/science.aab3050","volume":"350","author":"BM Lake","year":"2015","unstructured":"Lake, B. M., Salakhutdinov, R. & Tenenbaum, J. B. Human-level concept learning through probabilistic program induction. Science 350, 1332\u20131338 (2015).","journal-title":"Science"},{"key":"9631_CR17","doi-asserted-by":"publisher","first-page":"1394","DOI":"10.1111\/cogs.12400","volume":"41","author":"BD Roads","year":"2017","unstructured":"Roads, B. D. & Mozer, M. C. Improving human-machine cooperative classification via cognitive theories of similarity. Cogn. Sci. 41, 1394\u20131411 (2017).","journal-title":"Cogn. Sci."},{"key":"9631_CR18","doi-asserted-by":"crossref","unstructured":"Hendrycks, D. et al. The many faces of robustness: a critical analysis of out- of-distribution generalization. In Proc. IEEE\/CVF International Conference on Computer Vision 8340\u20138349 (IEEE, 2021).","DOI":"10.1109\/ICCV48922.2021.00823"},{"key":"9631_CR19","doi-asserted-by":"crossref","unstructured":"Pooch, E. H. P., Ballester, P. & Barros, R. C. Can we trust deep learning based diagnosis? The impact of domain shift in chest radiograph classification. In Proc.Thoracic Image Analysis: Second International Workshop (eds Petersen, J. et al.) 74\u201383 (Springer, 2020).","DOI":"10.1007\/978-3-030-62469-9_7"},{"key":"9631_CR20","unstructured":"Minderer, M. et al. Revisiting the calibration of modern neural networks. In Proc. Advances in Neural Information Processing Systems Vol. 34 (eds Ranzato, M. et al.) 15682\u201315694 (Curran Associates, 2021)."},{"key":"9631_CR21","doi-asserted-by":"publisher","first-page":"12","DOI":"10.1016\/j.neuroimage.2019.03.031","volume":"194","author":"RM Cichy","year":"2019","unstructured":"Cichy, R. M., Kriegeskorte, N., Jozwik, K. M., van den Bosch, J. J. F. & Charest, I. The spatiotemporal neural dynamics underlying perceived similarity for real-world objects. NeuroImage 194, 12\u201324 (2019).","journal-title":"NeuroImage"},{"key":"9631_CR22","doi-asserted-by":"publisher","first-page":"2608","DOI":"10.1523\/JNEUROSCI.5547-11.2012","volume":"32","author":"AC Connolly","year":"2012","unstructured":"Connolly, A. C., et al. The representation of biological classes in the human brain. J. Neurosci. 32, 2608\u20132618 (2012).","journal-title":"J. Neurosci."},{"key":"9631_CR23","doi-asserted-by":"publisher","first-page":"368","DOI":"10.1016\/j.neuroimage.2019.04.079","volume":"197","author":"ML King","year":"2019","unstructured":"King, M. L., Groen, I. I. A., Steel, A., Kravitz, D. J. & Baker, C. I. Similarity judgments and cortical visual responses reflect different properties of object and scene categories in naturalistic images. NeuroImage 197, 368\u2013382 (2019).","journal-title":"NeuroImage"},{"key":"9631_CR24","doi-asserted-by":"publisher","first-page":"2648","DOI":"10.1111\/cogs.12670","volume":"42","author":"JC Peterson","year":"2018","unstructured":"Peterson, J. C., Abbott, J. T. & Griffiths, T. L. Evaluating (and improving) the correspondence between deep neural networks and human representations. Cogn. Sci. 42, 2648\u20132669 (2018).","journal-title":"Cogn. Sci."},{"key":"9631_CR25","unstructured":"Muttenthaler, L., Dippel, J., Linhardt, L., Vandermeulen, R. A. & Kornblith, S. Human alignment of neural network representations. In Proc. The Eleventh International Conference on Learning Representations (ICLR, 2023)."},{"key":"9631_CR26","doi-asserted-by":"publisher","first-page":"e0223792","DOI":"10.1371\/journal.pone.0223792","volume":"14","author":"MN Hebart","year":"2019","unstructured":"Hebart, M. N. et al. THINGS: a database of 1,854 object concepts and more than 26,000 naturalistic object images. PLoS One\u00a014, e0223792 (2019).","journal-title":"PLoS One"},{"key":"9631_CR27","unstructured":"Lee, D.-H. et al. Pseudo-label: the simple and efficient semi-supervised learning method for deep neural networks. In Proc. Workshop on Challenges In Representation Learning (WREPL) 896\u2013902 (ICML, 2013)."},{"key":"9631_CR28","doi-asserted-by":"publisher","first-page":"211","DOI":"10.1007\/s11263-015-0816-y","volume":"115","author":"O Russakovsky","year":"2015","unstructured":"Russakovsky, O. et al. Imagenet large scale visual recognition challenge. Int. J. Comput. Vis.\u00a0115, 211\u2013252 (2015).","journal-title":"Int. J. Comput. Vis"},{"key":"9631_CR29","unstructured":"Oquab, M. et al. DINOv2: learning robust visual features without supervision. Transact. Mach. Learn. Res. https:\/\/openreview.net\/forum?id=a68SUt6zFt (2024)."},{"key":"9631_CR30","unstructured":"Tschannen, M. et al. Image captioners are scalable vision learners too. In Proc. Advances in Neural Information Processing Systems Vol. 36 (eds Oh, A. et al.) 46830\u201346855 (Curran Associates, 2023)."},{"key":"9631_CR31","doi-asserted-by":"publisher","first-page":"59","DOI":"10.1037\/0033-295X.85.2.59","volume":"85","author":"R Ratcliff","year":"1978","unstructured":"Ratcliff, R. A theory of memory retrieval. Psychol. Rev. 85, 59 (1978).","journal-title":"Psychol. Rev."},{"key":"9631_CR32","doi-asserted-by":"publisher","first-page":"1329","DOI":"10.1016\/j.neuron.2014.12.015","volume":"84","author":"R Kiani","year":"2014","unstructured":"Kiani, R., Corthell, L. & Shadlen, M. N. Choice certainty is informed by both evidence and decision time. Neuron 84, 1329\u20131342 (2014).","journal-title":"Neuron"},{"key":"9631_CR33","unstructured":"Gemini Team et al. Gemini 2.5: pushing the frontier with advanced reasoning, multimodality, long context, and next generation agentic capabilities. Preprint at https:\/\/arxiv.org\/abs\/2507.06261 (2025)."},{"key":"9631_CR34","doi-asserted-by":"publisher","first-page":"e2011417118","DOI":"10.1073\/pnas.2011417118","volume":"118","author":"J Mehrer","year":"2021","unstructured":"Mehrer, J., Spoerer, C. J., Jones, E. C., Kriegeskorte, N. & Kietzmann, T. C. An ecologically motivated image dataset for deep learning yields better models of human vision. Proc. Natl Acad. Sci. USA 118, e2011417118 (2021).","journal-title":"Proc. Natl Acad. Sci. USA"},{"key":"9631_CR35","doi-asserted-by":"publisher","first-page":"382","DOI":"10.1016\/0010-0285(76)90013-X","volume":"8","author":"E Rosch","year":"1976","unstructured":"Rosch, E., Mervis, C. B., Gray, W. D., Johnson, D. M. & Boyes-Braem, P. Basic objects in natural categories. Cogn. Psychol. 8, 382\u2013439 (1976).","journal-title":"Cogn. Psychol."},{"key":"9631_CR36","unstructured":"Wah, C., Branson, S., Welinder, P., Perona, P. & Belongie, S. The Caltech-UCSD Birds-200-2011 dataset. Caltech Vision Lab https:\/\/www.vision.caltech.edu\/datasets\/cub_200_2011\/ (2011)."},{"key":"9631_CR37","doi-asserted-by":"crossref","unstructured":"Nilsback, M.-E. & Zisserman, A. Automated flower classification over a large number of classes. In Proc. Sixth Indian Conference on Computer Vision, Graphics & Image Processing 722\u2013729 (IEEE, 2008).","DOI":"10.1109\/ICVGIP.2008.47"},{"key":"9631_CR38","doi-asserted-by":"crossref","unstructured":"Deng, J. et al. ImageNet: a large-scale hierarchical image database. In Proc. 2009 IEEE Conference on Computer Vision and Pattern Recognition 248\u2013255 (IEEE, 2009).","DOI":"10.1109\/CVPR.2009.5206848"},{"key":"9631_CR39","doi-asserted-by":"publisher","first-page":"1452","DOI":"10.1109\/TPAMI.2017.2723009","volume":"40","author":"B Zhou","year":"2018","unstructured":"Zhou, B., Lapedriza, A., Khosla, A., Oliva, A. & Torralba, A. Places: a 10 million image database for scene recognition. IEEE Trans. Pattern Anal. Mach. Intell. 40, 1452\u20131464 (2018).","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"9631_CR40","doi-asserted-by":"crossref","unstructured":"Sugiyama, M. & Kawanabe, M. Machine Learning in Non-stationary Environments: Introduction to Covariate Shift Adaptation (MIT Press, 2012).","DOI":"10.7551\/mitpress\/9780262017091.001.0001"},{"key":"9631_CR41","doi-asserted-by":"crossref","unstructured":"Farahani, A., Voghoei, S., Rasheed, K. & Arabnia, H. R. in Advances in Data Science and Information Engineering: Proceedings from ICDATA 2020 and IKE 2020 (eds Stahlbock, R. et al.) 877\u2013894 (Springer, 2021).","DOI":"10.1007\/978-3-030-71704-9_65"},{"key":"9631_CR42","unstructured":"Santurkar, S., Tsipras, D. & Madry, A. BREEDS: benchmarks for subpopulation shift. In Proc. The Ninth International Conference on Learning Representations (ICLR, 2021)."},{"key":"9631_CR43","doi-asserted-by":"crossref","unstructured":"Hendrycks, D., Zhao, K., Basart, S., Steinhardt, J. & Song, D. Natural adversarial examples. In Proc. IEEE\/CVF Conference on Computer Vision and Pattern Recognition 15262\u201315271 (IEEE, 2021).","DOI":"10.1109\/CVPR46437.2021.01501"},{"key":"9631_CR44","doi-asserted-by":"crossref","unstructured":"Kriegeskorte, N., Mur, M. & Bandettini, P. A. Representational similarity analysis-connecting the branches of systems neuroscience. Front. Syst. Neurosci. 2, 4 (2008).","DOI":"10.3389\/neuro.06.004.2008"},{"key":"9631_CR45","unstructured":"Sucholutsky, I. et al. Getting aligned on representational alignment. Preprint at https:\/\/arxiv.org\/abs\/2310.13018 (2023)."},{"key":"9631_CR46","doi-asserted-by":"publisher","first-page":"243","DOI":"10.1006\/cogp.1998.0694","volume":"37","author":"F Gary","year":"1998","unstructured":"Gary, F. Marcus. Rethinking eliminative connectionism. Cogn. Psychol. 37, 243\u2013282 (1998).","journal-title":"Cogn. Psychol."},{"key":"9631_CR47","unstructured":"Holyoak, K. J. & Hummel, J. E. in Cognitive Dynamics (eds Dietrich, E. & Markman, A. B.) 229\u2013263 (Psychology Press, 2014)."},{"key":"9631_CR48","doi-asserted-by":"publisher","first-page":"308","DOI":"10.1037\/rev0000371","volume":"130","author":"A Geiger","year":"2023","unstructured":"Geiger, A., Carstensen, A., Frank, M. C. & Potts, C. Relational reasoning and generalization using nonsymbolic neural networks. Psychol. Rev. 130, 308 (2023).","journal-title":"Psychol. Rev."},{"key":"9631_CR49","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1145\/3533378","volume":"55","author":"A Paleyes","year":"2022","unstructured":"Paleyes, A., Urma, R.-G. & Lawrence, N. D. Challenges in deploying machine learning: a survey of case studies. ACM Comput. Surv. 55, 1\u201329 (2022).","journal-title":"ACM Comput. Surv."},{"key":"9631_CR50","unstructured":"Amodei, D. et al. Concrete problems in AI safety. Preprint at https:\/\/arxiv.org\/abs\/1606.06565 (2016)."},{"key":"9631_CR51","doi-asserted-by":"publisher","first-page":"e82580","DOI":"10.7554\/eLife.82580","volume":"12","author":"MN Hebart","year":"2023","unstructured":"Hebart, M. N., et al. THINGS-data, a multimodal collection of large-scale datasets for investigating object representations in human brain and behavior. eLife 12, e82580 (2023).","journal-title":"eLife"},{"key":"9631_CR52","doi-asserted-by":"publisher","first-page":"98","DOI":"10.1016\/0093-934X(88)90126-5","volume":"34","author":"K Fukuzawa","year":"1988","unstructured":"Fukuzawa, K., et al. Internal representations and the conceptual operation of color in pure alexia with color naming defects. Brain Lang. 34, 98\u2013126 (1988).","journal-title":"Brain Lang."},{"key":"9631_CR53","doi-asserted-by":"publisher","first-page":"779","DOI":"10.1167\/4.9.9","volume":"4","author":"R Robilotto","year":"2004","unstructured":"Robilotto, R. & Zaidi, Q. Limits of lightness identification for real objects under natural viewing conditions. J. Vis.\u00a04, 779\u2013797 (2004).","journal-title":"J. Vis"},{"key":"9631_CR54","doi-asserted-by":"publisher","first-page":"1173","DOI":"10.1038\/s41562-020-00951-3","volume":"4","author":"MN Hebart","year":"2020","unstructured":"Hebart, M. N., Zheng, C. Y., Pereira, F. & Baker, C. I. Revealing the multidimensional mental representations of natural objects underlying humansimilarity judgements. Nat. Hum. Behav. 4, 1173\u20131185 (2020).","journal-title":"Nat. Hum. Behav."},{"key":"9631_CR55","unstructured":"Muttenthaler, L. et al. VICE: variational interpretable concept embeddings. Adv. Neural Inf. Process. Syst. 35, 33661\u201333675 (2022)."},{"key":"9631_CR56","unstructured":"Kingma, D. P. & Welling, M. Auto-encoding variational Bayes. In Proc. The Second International Conference on Learning Representations (eds Bengio, Y. & LeCun, Y.) (ICLR, 2014)."},{"key":"9631_CR57","unstructured":"Jimenez Rezende, D., Mohamed, S. & Wierstra, D. Stochastic backpropagation and approximate inference in deep generative models. In Proc. 31st International Conference on Machine Learning Vol. 32 (eds Xing, E. P. & Jebara, T.) 1278\u20131286 (PMLR, 2014)."},{"key":"9631_CR58","unstructured":"Graves, A. Practical variational inference for neural networks. In Proc. Advances in Neural Information Processing Systems Vol. 24 (eds Shawe-Taylor, J. et al.) 2348\u20132356 (Curran Associates, 2011)."},{"key":"9631_CR59","unstructured":"Blundell, C., Cornebise, J., Kavukcuoglu, K. & Wierstra, D. Weight uncertainty in neural network. In Proc. 32nd International Conference on Machine Learning Vol. 37 (eds Bach, F. & Blei, D.) 1613\u20131622 (PMLR, 2015)."},{"key":"9631_CR60","doi-asserted-by":"publisher","first-page":"859","DOI":"10.1080\/01621459.2017.1285773","volume":"112","author":"DM Blei","year":"2017","unstructured":"Blei, D. M., Kucukelbir, A. & McAuliffe, J. D. Variational inference: a review for statisticians. J. Am. Stat. Assoc. 112, 859\u2013877 (2017).","journal-title":"J. Am. Stat. Assoc."},{"key":"9631_CR61","unstructured":"Muttenthaler, L. et al. Improving neural network representations using human similarity judgments. In Proc. Advances in Neural Information Processing Systems Vol. 36 (eds Oh, A. et al.) 50978\u201351007 (Curran Associates, 2023)."},{"key":"9631_CR62","unstructured":"Muttenthaler, L., Vandermeulen, R. A., Zhang, Q., Unterthiner, T. & M\u00fcller, K.-R. Set learning for accurate and calibrated models. In Proc. The Twelfth International Conference on Learning Representations (ICLR, 2024)."},{"key":"9631_CR63","doi-asserted-by":"crossref","unstructured":"Zhai, X., Mustafa, B., Kolesnikov, A. & Beyer, L. Sigmoid loss for language image pre-training. In Proc. IEEE\/CVF International Conference on Computer Vision 11975\u201311986 (IEEE, 2023).","DOI":"10.1109\/ICCV51070.2023.01100"},{"key":"9631_CR64","unstructured":"Radford, A. et al. Learning transferable visual models from natural language supervision. In Proc 38th International Conference on Machine Learning Vol. 139 (eds Meila, M. & Zhang, T.) 8748\u20138763 (PMLR, 2021)."},{"key":"9631_CR65","unstructured":"Jia, C. et al. Scaling up visual and vision-language representation learning with noisy text supervision. In Proc. 38th International Conference on Machine Learning Vol. 139 (eds Meila, M. & Zhang, T.) 4904\u20134916 (PMLR, 2021)."},{"key":"9631_CR66","unstructured":"Alabdulmohsin, I. M., Zhai, X., Kolesnikov, A. & Beyer, L. Getting ViT in shape: scaling laws for compute-optimal model design. In Proc. Advances in Neural Information Processing Systems Vol. 36 (eds Oh, A. et al.) 16406\u201316425 (Curran Associates, 2023)."},{"key":"9631_CR67","unstructured":"Chen, X. et al. Pali-3 vision language models: smaller, faster, stronger. Preprint at https:\/\/arxiv.org\/abs\/2310.09199 (2023)."},{"key":"9631_CR68","unstructured":"Beyer, L. et al. PaliGemma: aversatile 3B VLM for transfer. Preprint at https:\/\/arxiv.org\/abs\/2407.07726 (2024)."},{"key":"9631_CR69","doi-asserted-by":"crossref","unstructured":"Caron, M. et al. Emerging properties in self-supervised vision transformers. In Proc. IEEE\/CVF International Conference on Computer Vision 9650\u20139660 (IEEE, 2021).","DOI":"10.1109\/ICCV48922.2021.00951"},{"key":"9631_CR70","unstructured":"Tschannen, M. et al. SigLIP 2: multilingual vision-language encoders with improved semantic understanding, localization, and dense features. Preprint at https:\/\/arxiv.org\/abs\/2502.14786 (2025)."},{"key":"9631_CR71","doi-asserted-by":"publisher","first-page":"12","DOI":"10.1016\/j.neuroimage.2019.03.031","volume":"194","author":"RM Cichy","year":"2019","unstructured":"Cichy, R. M., Kriegeskorte, N., Jozwik, K. M., van den Bosch, J. J. F. & Charest, I. The spatiotemporal neural dynamics underlying perceivedsimilarity for real-world objects. NeuroImage 194, 12\u201324 (2019).","journal-title":"NeuroImage"},{"key":"9631_CR72","doi-asserted-by":"crossref","unstructured":"Peterson, J. C., Abbott, J. T. & Griffiths, T. L. Adapting deep network features to capture psychological representations. In Proc. 38th Annual Meeting of the Cognitive Science Society, Recognizing and Representing Events (eds Papafragou, A. et al.) (Cognitive Science Society, 2016).","DOI":"10.24963\/ijcai.2017\/697"},{"key":"9631_CR73","doi-asserted-by":"crossref","unstructured":"Peterson, J. C., Battleday, R. M., Griffiths, T. L. & Russakovsky, O. Human uncertainty makes classification more robust. In Proc. IEEE\/CVF International Conference on Computer Vision 9616\u20139625 (IEEE, 2019).","DOI":"10.1109\/ICCV.2019.00971"},{"key":"9631_CR74","doi-asserted-by":"publisher","first-page":"e106953","DOI":"10.1371\/journal.pone.0106953","volume":"9","author":"MB Brodeur","year":"2014","unstructured":"Brodeur, M. B., Gu\u00e9rard, K. & Bouras, M. Bank of standardized stimuli (BOSS) phase II: 930 new normative photos. PLoS One\u00a09,\u00a0e106953 (2014).","journal-title":"PLoS One"},{"key":"9631_CR75","doi-asserted-by":"crossref","unstructured":"Grant, D. A. The Latin square principle in the design and analysis of psychological experiments. Psychol. Bull. 45, 427 (1948).","DOI":"10.1037\/h0053912"},{"key":"9631_CR76","doi-asserted-by":"publisher","unstructured":"Born, F. Frieda-Josefine\/HumanEval_experiment: HumanEval experiment. Zenodo https:\/\/doi.org\/10.5281\/zenodo.13749102 (2025).","DOI":"10.5281\/zenodo.13749102"},{"key":"9631_CR77","doi-asserted-by":"publisher","unstructured":"Born, F. AligNet Human-Validation-Experiment. Zenodo https:\/\/doi.org\/10.5281\/zenodo.15554173 (2025).","DOI":"10.5281\/zenodo.15554173"}],"container-title":["Nature"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.nature.com\/articles\/s41586-025-09631-6.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/www.nature.com\/articles\/s41586-025-09631-6","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/www.nature.com\/articles\/s41586-025-09631-6.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,11,12]],"date-time":"2025-11-12T16:02:14Z","timestamp":1762963334000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.nature.com\/articles\/s41586-025-09631-6"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,11,12]]},"references-count":77,"journal-issue":{"issue":"8089","published-print":{"date-parts":[[2025,11,13]]}},"alternative-id":["9631"],"URL":"https:\/\/doi.org\/10.1038\/s41586-025-09631-6","relation":{},"ISSN":["0028-0836","1476-4687"],"issn-type":[{"value":"0028-0836","type":"print"},{"value":"1476-4687","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,11,12]]},"assertion":[{"value":"14 November 2024","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"15 September 2025","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"12 November 2025","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"All authors, aside from B.S. and F.B., are or were employed by Google DeepMind.","order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Competing interests"}}]}}