August 2023
Volume 23, Issue 9
Open Access
Vision Sciences Society Annual Meeting Abstract  |   August 2023
Feature Visualizations do not sufficiently explain hidden units of Artificial Neural Networks
Author Affiliations & Notes
  • Thomas Klein
    Neural Information Processing Group, University of Tübingen
    Max Planck Institute for Intelligent Systems, Tübingen
  • Wieland Brendel
    Max Planck Institute for Intelligent Systems, Tübingen
  • Felix Wichmann
    Neural Information Processing Group, University of Tübingen
  • Footnotes
    Acknowledgements  Funded by EXC number 2064/1 – project number 390727645 and by the German Research Foundation (DFG): SFB 1233 – project number 276693517. The authors would like to thank the International Max Planck Research School for Intelligent Systems (IMPRS-IS) for supporting Thomas Klein.
Journal of Vision August 2023, Vol.23, 5062. doi:https://doi.org/10.1167/jov.23.9.5062
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Thomas Klein, Wieland Brendel, Felix Wichmann; Feature Visualizations do not sufficiently explain hidden units of Artificial Neural Networks. Journal of Vision 2023;23(9):5062. https://doi.org/10.1167/jov.23.9.5062.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Artificial Neural Networks (ANNs) have been proposed as computational models of the primate ventral stream, because their performance on tasks such as image classification rivals or exceeds human baselines. But useful models should not only predict data well, but also offer insights into the systems they represent, which remains a challenge for ANNs. We here investigate a specific method that has been proposed to shed light on the representations learned by ANNs: Feature Visualizations (FVs), that is, synthetic images specifically designed to excite individual units ("neurons") of the target network. Theoretically, these images should visualize the features that a unit is sensitive to, like receptive fields in neurophysiology. We conduct a psychophysical experiment to establish an upper bound on the interpretability afforded by FVs, in which participants need to match five sets of exemplars (natural images that highly activate certain units) to five sets of FVs of the same units---a task that should be trivial if FVs were informative. Extending earlier work that has cast doubts on the utility of this method, we show that (1) even human experts perform hardly better than chance when trying to match a unit's FVs to its exemplars and that (2) matching exemplars to each other is much easier, even if only a single exemplar is shown per set. Presumably, this difficulty is not caused by so-called polysemantic units (neurons that code for multiple unrelated features, possibly mixing them in their visualizations) but by the unnatural visual appearance of FVs themselves. We also investigate the effect of visualizing units from different layers and find that the interpretability of FVs declines in later layers---contrary to what one might expect, since later layers should represent semantic concepts. These findings highlight the need for better interpretability-techniques, if ANNs are ever to become useful models of human vision.

×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×