Saved in:
Bibliographic Details
Main Authors: Douglas, Pk, Farahani, Farzad Vasheghani
Format: Preprint
Published: 2020
Subjects:
Online Access:https://arxiv.org/abs/2002.06816
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909317506007040
author Douglas, Pk
Farahani, Farzad Vasheghani
author_facet Douglas, Pk
Farahani, Farzad Vasheghani
contents The increasing use of deep neural networks (DNNs) has motivated a parallel endeavor: the design of adversaries that profit from successful misclassifications. However, not all adversarial examples are crafted for malicious purposes. For example, real world systems often contain physical, temporal, and sampling variability across instrumentation. Adversarial examples in the wild may inadvertently prove deleterious for accurate predictive modeling. Conversely, naturally occurring covariance of image features may serve didactic purposes. Here, we studied the stability of deep learning representations for neuroimaging classification across didactic and adversarial conditions characteristic of MRI acquisition variability. We show that representational similarity and performance vary according to the frequency of adversarial examples in the input space.
format Preprint
id arxiv_https___arxiv_org_abs_2002_06816
institution arXiv
publishDate 2020
record_format arxiv
spellingShingle On the Similarity of Deep Learning Representations Across Didactic and Adversarial Examples
Douglas, Pk
Farahani, Farzad Vasheghani
Computer Vision and Pattern Recognition
Machine Learning
Image and Video Processing
Neurons and Cognition
I.2.4
The increasing use of deep neural networks (DNNs) has motivated a parallel endeavor: the design of adversaries that profit from successful misclassifications. However, not all adversarial examples are crafted for malicious purposes. For example, real world systems often contain physical, temporal, and sampling variability across instrumentation. Adversarial examples in the wild may inadvertently prove deleterious for accurate predictive modeling. Conversely, naturally occurring covariance of image features may serve didactic purposes. Here, we studied the stability of deep learning representations for neuroimaging classification across didactic and adversarial conditions characteristic of MRI acquisition variability. We show that representational similarity and performance vary according to the frequency of adversarial examples in the input space.
title On the Similarity of Deep Learning Representations Across Didactic and Adversarial Examples
topic Computer Vision and Pattern Recognition
Machine Learning
Image and Video Processing
Neurons and Cognition
I.2.4
url https://arxiv.org/abs/2002.06816