English
 
User Manual Privacy Policy Disclaimer Contact us
  Advanced SearchBrowse

Item

ITEM ACTIONSEXPORT
  Generalisation in humans and deep neural networks

Geirhos, R., Medina Temme, C., Rauber, J., Schuett, H., Bethge, M., & Wichmann, F. (2019). Generalisation in humans and deep neural networks. In S. Bengio, H. Wallach, H. Larochelle, K. Grauman, N. Cesa-Bianchi, & R. Garnett (Eds.), Advances in Neural Information Processing Systems 31 (pp. 7549-7561). Red Hook, NY, USA: Curran.

Item is

Basic

show hide
Item Permalink: http://hdl.handle.net/21.11116/0000-0002-0582-5 Version Permalink: http://hdl.handle.net/21.11116/0000-0005-4B65-6
Genre: Conference Paper

Files

show Files

Creators

show
hide
 Creators:
Geirhos, R, Author
Medina Temme, CR, Author
Rauber, J, Author
Schuett, HH, Author
Bethge, M1, 2, Author              
Wichmann, FA, Author              
Affiliations:
1Research Group Computational Vision and Neuroscience, Max Planck Institute for Biological Cybernetics, Max Planck Society, ou_1497805              
2Max Planck Institute for Biological Cybernetics, Max Planck Society, Spemannstrasse 38, 72076 Tübingen, DE, ou_1497794              

Content

show
hide
Free keywords: -
 Abstract: We compare the robustness of humans and current convolutional deep neural networks (DNNs) on object recognition under twelve different types of image degradations. First, using three well known DNNs (ResNet-152, VGG-19, GoogLeNet) we find the human visual system to be more robust to nearly all of the tested image manipulations, and we observe progressively diverging classification error-patterns between humans and DNNs when the signal gets weaker. Secondly, we show that DNNs trained directly on distorted images consistently surpass human performance on the exact distortion types they were trained on, yet they display extremely poor generalisation abilities when tested on other distortion types. For example, training on salt-and-pepper noise does not imply robustness on uniform white noise and vice versa. Thus, changes in the noise distribution between training and testing constitutes a crucial challenge to deep learning vision systems that can be systematically addressed in a lifelong machine learning approach. Our new dataset consisting of 83K carefully measured human psychophysical trials provide a useful reference for lifelong robustness against image degradations set by the human visual system.

Details

show
hide
Language(s):
 Dates: 2018-082018-122019-07
 Publication Status: Published in print
 Pages: -
 Publishing info: -
 Table of Contents: -
 Rev. Method: -
 Identifiers: -
 Degree: -

Event

show
hide
Title: Thirty-second Annual Conference on Neural Information Processing Systems 2018 (NeurIPS 2018)
Place of Event: Montréal, Canada
Start-/End Date: 2018-12-03 - 2018-12-08

Legal Case

show

Project information

show

Source 1

show
hide
Title: Advances in Neural Information Processing Systems 31
Source Genre: Proceedings
 Creator(s):
Bengio, S, Editor
Wallach, HM, Editor
Larochelle, H, Editor
Grauman, K, Editor
Cesa-Bianchi, N, Editor
Garnett, R, Editor
Affiliations:
-
Publ. Info: Red Hook, NY, USA : Curran
Pages: - Volume / Issue: - Sequence Number: - Start / End Page: 7549 - 7561 Identifier: ISBN: 978-1-5108-8447-2