Diverse Visual Experience Promotes Integrated and Human-Aligned Face Representations in Deep Neural Networks

Diverse Visual Experience Promotes Integrated and Human-Aligned Face Representations in Deep Neural Networks | bioRxiv Skip to main content Home About Submit ALERTS / RSS Search for this keyword Advanced Search New Results Diverse Visual Experience Promotes Integrated and Human-Aligned Face Representations in Deep Neural Networks View ORCID Profile Elaheh Akbari , View ORCID Profile Katharina Dobs doi: https://doi.org/10.64898/2025.12.19.695398 Elaheh Akbari 1 Department of Psychology, Justus Liebig University Giessen , Giessen, Germany , 35394 2 Center for Mind , Brain and Behavior, Universities of Marburg, Giessen, and Darmstadt, Marburg, Germany , 35032 Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Elaheh Akbari For correspondence: elahakbri{at}gmail.com Katharina Dobs 2 Center for Mind , Brain and Behavior, Universities of Marburg, Giessen, and Darmstadt, Marburg, Germany , 35032 3 Department of Computer Science, Justus Liebig University Giessen , Giessen, Germany , 35390 4 Center for Applied Computer Science and Data Science, Justus Liebig University Giessen , Giessen, Germany , 35390 Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Katharina Dobs Abstract Full Text Info/History Metrics Data/Code Preview PDF Abstract Humans are experts at recognizing faces, yet this expertise is not uniform: people perceive faces from familiar facial groups more accurately than those from unfamiliar ones—a phenomenon known as the Other-Race Effect (ORE). Diverse facial exposure mitigates this bias, but how it reorganizes face representations to support cross-group recognition remains unclear. To address this question, we combined controlled training variation in deep convolutional neural networks (CNNs) with analyses of representational geometry, lesioning, and human behavioral data. CNNs trained exclusively on either Asian or White faces reproduced ORE-like biases, whereas a CNN trained on both groups showed reduced bias and balanced recognition performance. Critically, representational similarity analyses and lesioning revealed largely overlapping feature spaces across groups, indicating an integrated representational organization. Trial-by-trial comparisons with human choices further showed that the dual-trained CNN best captured human face-matching behavior across both Asian and White participants, outperforming single-trained networks on cross-group trials. Together, these findings show that diverse visual experience promotes an integrated representational geometry that supports cross-group generalization, providing a computational account of how exposure diversity may shape human face representations. Competing Interest Statement The authors have declared no competing interest. Footnotes In this revision, we added additional commentary regarding low-level image statistics and the proportion of lesioned units. We also made rearrangements to improve clarity and added a short paragraph to the Conclusion discussing an additional question addressed by the study. https://github.com/VCCN-lab/ORE_in_CNNs_and_humans https://osf.io/dbks3 https://github.com/martinezjulio/sdnn Funder Information Declared Deutsche Forschungsgemeinschaft, https://ror.org/018mejw64 , Project No. 222641018, TP C9 , Project No. 533717223 European Research Council, https://ror.org/0472cxd90 , ERC-2023-STG-101117441 Hessisches Ministerium für Wissenschaft und Kunst , LOEWE Start Professorship Copyright The copyright holder for this preprint is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC 4.0 International license . Back to top Previous Next Posted March 08, 2026. Download PDF Data/Code Email Thank you for your interest in spreading the word about bioRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following Diverse Visual Experience Promotes Integrated and Human-Aligned Face Representations in Deep Neural Networks Message Subject (Your Name) has forwarded a page to you from bioRxiv Message Body (Your Name) thought you would like to see this page from the bioRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share Diverse Visual Experience Promotes Integrated and Human-Aligned Face Representations in Deep Neural Networks Elaheh Akbari , Katharina Dobs bioRxiv 2025.12.19.695398; doi: https://doi.org/10.64898/2025.12.19.695398 Share This Article: Copy Citation Tools Diverse Visual Experience Promotes Integrated and Human-Aligned Face Representations in Deep Neural Networks Elaheh Akbari , Katharina Dobs bioRxiv 2025.12.19.695398; doi: https://doi.org/10.64898/2025.12.19.695398 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Area Neuroscience Subject Areas All Articles Animal Behavior and Cognition (8164) Biochemistry (19027) Bioengineering (15171) Bioinformatics (45592) Biophysics (22958) Cancer Biology (20204) Cell Biology (27430) Clinical Trials (138) Developmental Biology (14199) Ecology (21485) Epidemiology (2067) Evolutionary Biology (25889) Genetics (16403) Genomics (23920) Immunology (19115) Microbiology (43383) Molecular Biology (18434) Neuroscience (95525) Paleontology (716) Pathology (3049) Pharmacology and Toxicology (5228) Physiology (8314) Plant Biology (16341) Scientific Communication and Education (2121) Synthetic Biology (4664) Systems Biology (10430) Zoology (2435)

Related documents

• Experience Alone Can Generate Human Face Specialization: Evidence From Deep Learning Models.#382516
• FuseReg: Regularizing Layer Fusion Mitigates the Reconstruction-Generation Gap in Representation Autoencoders#1106319
• Early Abstraction, Continuous Generalization: The Dynamics of Real-Life Face Learning#663374
• Early Abstraction, Continuous Generalization: The Dynamics of Real-Life Face Learning#663248
• Classification and Reconstruction Processes in Deep Predictive Coding Networks: Antagonists or Allies?#1141382