Skip to main content

Main menu

  • Home
  • Articles
    • Current
    • Special Feature Articles - Most Recent
    • Special Features
    • Colloquia
    • Collected Articles
    • PNAS Classics
    • List of Issues
  • Front Matter
    • Front Matter Portal
    • Journal Club
  • News
    • For the Press
    • This Week In PNAS
    • PNAS in the News
  • Podcasts
  • Authors
    • Information for Authors
    • Editorial and Journal Policies
    • Submission Procedures
    • Fees and Licenses
  • Submit
  • Submit
  • About
    • Editorial Board
    • PNAS Staff
    • FAQ
    • Accessibility Statement
    • Rights and Permissions
    • Site Map
  • Contact
  • Journal Club
  • Subscribe
    • Subscription Rates
    • Subscriptions FAQ
    • Open Access
    • Recommend PNAS to Your Librarian

User menu

  • Log in
  • My Cart

Search

  • Advanced search
Home
Home
  • Log in
  • My Cart

Advanced Search

  • Home
  • Articles
    • Current
    • Special Feature Articles - Most Recent
    • Special Features
    • Colloquia
    • Collected Articles
    • PNAS Classics
    • List of Issues
  • Front Matter
    • Front Matter Portal
    • Journal Club
  • News
    • For the Press
    • This Week In PNAS
    • PNAS in the News
  • Podcasts
  • Authors
    • Information for Authors
    • Editorial and Journal Policies
    • Submission Procedures
    • Fees and Licenses
  • Submit
Colloquium Paper

Controversial stimuli: Pitting neural networks against each other as models of human cognition

View ORCID ProfileTal Golan, View ORCID ProfilePrashant C. Raju, and Nikolaus Kriegeskorte
  1. aZuckerman Mind Brain Behavior Institute, Columbia University, New York, NY 10027;
  2. bDepartment of Computer Science, Columbia University, New York, NY 10027;
  3. cDepartment of Psychology, Columbia University, New York, NY 10027;
  4. dDepartment of Neuroscience, Columbia University, New York, NY 10027;
  5. eDepartment of Electrical Engineering, Columbia University, New York, NY 10027

See allHide authors and affiliations

PNAS November 24, 2020 117 (47) 29330-29337; first published November 23, 2020; https://doi.org/10.1073/pnas.1912334117
Tal Golan
aZuckerman Mind Brain Behavior Institute, Columbia University, New York, NY 10027;
  • Find this author on Google Scholar
  • Find this author on PubMed
  • Search for this author on this site
  • ORCID record for Tal Golan
  • For correspondence: n.kriegeskorte@columbia.edu tal.golan@columbia.edu
Prashant C. Raju
bDepartment of Computer Science, Columbia University, New York, NY 10027;
  • Find this author on Google Scholar
  • Find this author on PubMed
  • Search for this author on this site
  • ORCID record for Prashant C. Raju
Nikolaus Kriegeskorte
aZuckerman Mind Brain Behavior Institute, Columbia University, New York, NY 10027;
cDepartment of Psychology, Columbia University, New York, NY 10027;
dDepartment of Neuroscience, Columbia University, New York, NY 10027;
eDepartment of Electrical Engineering, Columbia University, New York, NY 10027
  • Find this author on Google Scholar
  • Find this author on PubMed
  • Search for this author on this site
  • For correspondence: n.kriegeskorte@columbia.edu tal.golan@columbia.edu
  1. Edited by Joshua B. Tenenbaum, Massachusetts Institute of Technology, Cambridge, MA, and accepted by Editorial Board Member Dale Purves September 15, 2020 (received for review November 13, 2019)

  • Article
  • Figures & SI
  • Info & Metrics
  • PDF
Loading

Abstract

Distinct scientific theories can make similar predictions. To adjudicate between theories, we must design experiments for which the theories make distinct predictions. Here we consider the problem of comparing deep neural networks as models of human visual recognition. To efficiently compare models’ ability to predict human responses, we synthesize controversial stimuli: images for which different models produce distinct responses. We applied this approach to two visual recognition tasks, handwritten digits (MNIST) and objects in small natural images (CIFAR-10). For each task, we synthesized controversial stimuli to maximize the disagreement among models which employed different architectures and recognition algorithms. Human subjects viewed hundreds of these stimuli, as well as natural examples, and judged the probability of presence of each digit/object category in each image. We quantified how accurately each model predicted the human judgments. The best-performing models were a generative analysis-by-synthesis model (based on variational autoencoders) for MNIST and a hybrid discriminative–generative joint energy model for CIFAR-10. These deep neural networks (DNNs), which model the distribution of images, performed better than purely discriminative DNNs, which learn only to map images to labels. None of the candidate models fully explained the human responses. Controversial stimuli generalize the concept of adversarial examples, obviating the need to assume a ground-truth model. Unlike natural images, controversial stimuli are not constrained to the stimulus distribution models are trained on, thus providing severe out-of-distribution tests that reveal the models’ inductive biases. Controversial stimuli therefore provide powerful probes of discrepancies between models and human perception.

  • visual object recognition
  • deep neural networks
  • optimal experimental design
  • adversarial examples
  • generative modeling

Footnotes

  • ↵1To whom correspondence may be addressed. Email: n.kriegeskorte{at}columbia.edu or tal.golan{at}columbia.edu.
  • Author contributions: T.G., P.C.R., and N.K. designed research; T.G. and P.C.R. performed research; and T.G. and N.K. wrote the paper.

  • The authors declare no competing interest.

  • This paper results from the Arthur M. Sackler Colloquium of the National Academy of Sciences, “Brain Produces Mind by Modeling,” held May 1–3, 2019, at the Arnold and Mabel Beckman Center of the National Academies of Sciences and Engineering in Irvine, CA. NAS colloquia began in 1991 and have been published in PNAS since 1995. From February 2001 through May 2019, colloquia were supported by a generous gift from The Dame Jillian and Dr. Arthur M. Sackler Foundation for the Arts, Sciences, & Humanities, in memory of Dame Sackler’s husband, Arthur M. Sackler. The complete program and video recordings of most presentations are available on the NAS website at http://www.nasonline.org/brain-produces-mind-by.

  • This article is a PNAS Direct Submission. J.B.T. is a guest editor invited by the Editorial Board.

  • This article contains supporting information online at https://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1912334117/-/DCSupplemental.

Data Availability.

Optimization source code, synthesized images, and anonymized, detailed behavioral testing results are available at GitHub, https://github.com/kriegeskorte-lab/PNAS_2020_Controversial_Stimuli.

Published under the PNAS license.

View Full Text
PreviousNext
Back to top
Article Alerts
Email Article

Thank you for your interest in spreading the word on PNAS.

NOTE: We only request your email address so that the person you are recommending the page to knows that you wanted them to see it, and that it is not junk mail. We do not capture any email address.

Enter multiple addresses on separate lines or separate them with commas.
Controversial stimuli: Pitting neural networks against each other as models of human cognition
(Your Name) has sent you a message from PNAS
(Your Name) thought you would like to see the PNAS web site.
CAPTCHA
This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.
Citation Tools
Controversial stimuli: Pitting neural networks against each other as models of human cognition
Tal Golan, Prashant C. Raju, Nikolaus Kriegeskorte
Proceedings of the National Academy of Sciences Nov 2020, 117 (47) 29330-29337; DOI: 10.1073/pnas.1912334117

Citation Manager Formats

  • BibTeX
  • Bookends
  • EasyBib
  • EndNote (tagged)
  • EndNote 8 (xml)
  • Medlars
  • Mendeley
  • Papers
  • RefWorks Tagged
  • Ref Manager
  • RIS
  • Zotero
Request Permissions
Share
Controversial stimuli: Pitting neural networks against each other as models of human cognition
Tal Golan, Prashant C. Raju, Nikolaus Kriegeskorte
Proceedings of the National Academy of Sciences Nov 2020, 117 (47) 29330-29337; DOI: 10.1073/pnas.1912334117
del.icio.us logo Digg logo Reddit logo Twitter logo CiteULike logo Facebook logo Google logo Mendeley logo
  • Tweet Widget
  • Facebook Like
  • Mendeley logo Mendeley

Article Classifications

  • Biological Sciences
  • Psychological and Cognitive Sciences
Proceedings of the National Academy of Sciences: 117 (47)
Table of Contents

Submit

Sign up for Article Alerts

Jump to section

  • Article
    • Abstract
    • Results
    • Experiment 1: Adjudicating among MNIST Models
    • Experiment 2: Adjudicating among CIFAR-10 Models
    • Discussion
    • Materials and Methods
    • Data Availability.
    • Acknowledgments
    • Footnotes
    • References
  • Figures & SI
  • Info & Metrics
  • PDF

You May Also be Interested in

Setting sun over a sun-baked dirt landscape
Core Concept: Popular integrated assessment climate policy models have key caveats
Better explicating the strengths and shortcomings of these models will help refine projections and improve transparency in the years ahead.
Image credit: Witsawat.S.
Model of the Amazon forest
News Feature: A sea in the Amazon
Did the Caribbean sweep into the western Amazon millions of years ago, shaping the region’s rich biodiversity?
Image credit: Tacio Cordeiro Bicudo (University of São Paulo, São Paulo, Brazil), Victor Sacek (University of São Paulo, São Paulo, Brazil), and Lucy Reading-Ikkanda (artist).
Syrian archaeological site
Journal Club: In Mesopotamia, early cities may have faltered before climate-driven collapse
Settlements 4,200 years ago may have suffered from overpopulation before drought and lower temperatures ultimately made them unsustainable.
Image credit: Andrea Ricci.
Steamboat Geyser eruption.
Eruption of Steamboat Geyser
Mara Reed and Michael Manga explore why Yellowstone's Steamboat Geyser resumed erupting in 2018.
Listen
Past PodcastsSubscribe
Birds nestling on tree branches
Parent–offspring conflict in songbird fledging
Some songbird parents might improve their own fitness by manipulating their offspring into leaving the nest early, at the cost of fledgling survival, a study finds.
Image credit: Gil Eckrich (photographer).

Similar Articles

Site Logo
Powered by HighWire
  • Submit Manuscript
  • Twitter
  • Facebook
  • RSS Feeds
  • Email Alerts

Articles

  • Current Issue
  • Special Feature Articles – Most Recent
  • List of Issues

PNAS Portals

  • Anthropology
  • Chemistry
  • Classics
  • Front Matter
  • Physics
  • Sustainability Science
  • Teaching Resources

Information

  • Authors
  • Editorial Board
  • Reviewers
  • Subscribers
  • Librarians
  • Press
  • Site Map
  • PNAS Updates
  • FAQs
  • Accessibility Statement
  • Rights & Permissions
  • About
  • Contact

Feedback    Privacy/Legal

Copyright © 2021 National Academy of Sciences. Online ISSN 1091-6490