IDEAS home Printed from https://ideas.repec.org/a/nat/natcom/v13y2022i1d10.1038_s41467-022-35659-7.html
   My bibliography  Save this article

Efficient neural codes naturally emerge through gradient descent learning

Author

Listed:
  • Ari S. Benjamin

    (University of Pennsylvania)

  • Ling-Qi Zhang

    (University of Pennsylvania)

  • Cheng Qiu

    (University of Pennsylvania)

  • Alan A. Stocker

    (University of Pennsylvania)

  • Konrad P. Kording

    (University of Pennsylvania
    University of Pennsylvania)

Abstract

Human sensory systems are more sensitive to common features in the environment than uncommon features. For example, small deviations from the more frequently encountered horizontal orientations can be more easily detected than small deviations from the less frequent diagonal ones. Here we find that artificial neural networks trained to recognize objects also have patterns of sensitivity that match the statistics of features in images. To interpret these findings, we show mathematically that learning with gradient descent in neural networks preferentially creates representations that are more sensitive to common features, a hallmark of efficient coding. This effect occurs in systems with otherwise unconstrained coding resources, and additionally when learning towards both supervised and unsupervised objectives. This result demonstrates that efficient codes can naturally emerge from gradient-like learning.

Suggested Citation

  • Ari S. Benjamin & Ling-Qi Zhang & Cheng Qiu & Alan A. Stocker & Konrad P. Kording, 2022. "Efficient neural codes naturally emerge through gradient descent learning," Nature Communications, Nature, vol. 13(1), pages 1-12, December.
  • Handle: RePEc:nat:natcom:v:13:y:2022:i:1:d:10.1038_s41467-022-35659-7
    DOI: 10.1038/s41467-022-35659-7
    as

    Download full text from publisher

    File URL: https://www.nature.com/articles/s41467-022-35659-7
    File Function: Abstract
    Download Restriction: no

    File URL: https://libkey.io/10.1038/s41467-022-35659-7?utm_source=ideas
    LibKey link: if access is restricted and if your library uses this service, LibKey will redirect you to where you can use your library subscription to access this item
    ---><---

    References listed on IDEAS

    as
    1. J. Gold & P. J. Bennett & A. B. Sekuler, 1999. "Signal but not noise changes with perceptual learning," Nature, Nature, vol. 402(6758), pages 176-178, November.
    2. Aniek Schoups & Rufin Vogels & Ning Qian & Guy Orban, 2001. "Practising orientation identification improves orientation coding in V1 neurons," Nature, Nature, vol. 412(6846), pages 549-553, August.
    Full references (including those not matched with items on IDEAS)

    Most related items

    These are the items that most often cite the same works as this one and are cited by the same works as this one.
    1. Alan Consorti & Gabriele Sansevero & Irene Marco & Silvia Floridia & Elena Novelli & Nicoletta Berardi & Alessandro Sale, 2024. "An essential role for the latero-medial secondary visual cortex in the acquisition and retention of visual perceptual learning in mice," Nature Communications, Nature, vol. 15(1), pages 1-16, December.
    2. Richard F Murray & Khushbu Patel & Alan Yee, 2015. "Posterior Probability Matching and Human Perceptual Decision Making," PLOS Computational Biology, Public Library of Science, vol. 11(6), pages 1-16, June.
    3. Jiashu Liu & Yingtian He & Andreanne Lavoie & Guy Bouvier & Bao-hua Liu, 2023. "A direction-selective cortico-brainstem pathway adaptively modulates innate behaviors," Nature Communications, Nature, vol. 14(1), pages 1-23, December.
    4. Gesa Lange & Eric Lowet & Mark J Roberts & Peter De Weerd, 2018. "Within-quadrant position and orientation specificity after extensive orientation discrimination learning is related to performance gains during late learning," PLOS ONE, Public Library of Science, vol. 13(9), pages 1-37, September.
    5. Stefanie Duyck & Hans Op de Beeck, 2019. "An investigation of far and near transfer in a gamified visual learning paradigm," PLOS ONE, Public Library of Science, vol. 14(12), pages 1-23, December.
    6. Hojin Jang & Frank Tong, 2024. "Improved modeling of human vision by incorporating robustness to blur in convolutional neural networks," Nature Communications, Nature, vol. 15(1), pages 1-14, December.
    7. Jeyadarshan Jeyabalaratnam & Vishal Bharmauria & Lyes Bachatene & Sarah Cattan & Annie Angers & Stéphane Molotchnikoff, 2013. "Adaptation Shifts Preferred Orientation of Tuning Curve in the Mouse Visual Cortex," PLOS ONE, Public Library of Science, vol. 8(5), pages 1-8, May.

    More about this item

    Statistics

    Access and download statistics

    Corrections

    All material on this site has been provided by the respective publishers and authors. You can help correct errors and omissions. When requesting a correction, please mention this item's handle: RePEc:nat:natcom:v:13:y:2022:i:1:d:10.1038_s41467-022-35659-7. See general information about how to correct material in RePEc.

    If you have authored this item and are not yet registered with RePEc, we encourage you to do it here. This allows to link your profile to this item. It also allows you to accept potential citations to this item that we are uncertain about.

    If CitEc recognized a bibliographic reference but did not link an item in RePEc to it, you can help with this form .

    If you know of missing items citing this one, you can help us creating those links by adding the relevant references in the same way as above, for each refering item. If you are a registered author of this item, you may also want to check the "citations" tab in your RePEc Author Service profile, as there may be some citations waiting for confirmation.

    For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: Sonal Shukla or Springer Nature Abstracting and Indexing (email available below). General contact details of provider: http://www.nature.com .

    Please note that corrections may take a couple of weeks to filter through the various RePEc services.

    IDEAS is a RePEc service. RePEc uses bibliographic data supplied by the respective publishers.