What do adversarial images tell us about human vision?

Marin Dujmović,Jeffrey S Bowers,Gaurav Malhotra

doi:10.7554/elife.55978

Abstract

Deep convolutional neural networks (DCNNs) are frequently described as the best current models of human and primate vision. An obvious challenge to this claim is the existence of adversarial images that fool DCNNs but are uninterpretable to humans. However, recent research has suggested that there may be similarities in how humans and DCNNs interpret these seemingly nonsense images. We reanalysed data from a high-profile paper and conducted five experiments controlling for different ways in which these images can be generated and selected. We show human-DCNN agreement is much weaker and more variable than previously reported, and that the weak agreement is contingent on the choice of adversarial images and the design of the experiment. Indeed, we find there are well-known methods of generating images for which humans show no agreement with DCNNs. We conclude that adversarial images still pose a challenge to theorists using DCNNs as models of human vision.

Highlights

Deep convolutional neural networks (DCNNs) are models of computer vision that have reached, and in some cases exceeded, human performance in many image classification benchmarks such as ImageNet [18]
In trying to understand the surprisingly large agreement between humans and DCNNs observed by Z&F, was to reassess how they measured this agreement on classification of adversarial images
Chance level is 1/48, so if the participant chooses the same label as the DCNN on two or more trials, they were labelled as agreeing with the DCNN

Summary

Introduction

Deep convolutional neural networks (DCNNs) are models of computer vision that have reached, and in some cases exceeded, human performance in many image classification benchmarks such as ImageNet [18]. DCNNs identify objects in a similar way to the infereotemproal cortex (IT) that supports object recognition in humans and primates. If so, these models may provide important new insights into the underlying computations performed in IT. These models may provide important new insights into the underlying computations performed in IT Consistent with this possibility, a number of researchers have highlighted various functional similarities between DCNNs and human vision [29] as well. Preprint submitted to bioRXiv as similarities in patterns of activation of neurons in IT and units in DCNNs [35]. Kubilius et al [24] write: “Deep artificial neural networks with spatially repeated processing (a.k.a., deep convolutional [Artificial Neural Networks]) have been established as the best class of candidate models of visual processing in primate ventral visual processing stream” (p.1)

Methods

Results

Conclusion