Showing 1–1 of 1 results for author: Trost, T A
-
On the security relevance of weights in deep learning
Authors:
Kathrin Grosse,
Thomas A. Trost,
Marius Mosbach,
Michael Backes,
Dietrich Klakow
Abstract:
Recently, a weight-based attack on stochastic gradient descent inducing overfitting has been proposed. We show that the threat is broader: A task-independent permutation on the initial weights suffices to limit the achieved accuracy to for example 50% on the Fashion MNIST dataset from initially more than $90$%. These findings are confirmed on MNIST and CIFAR. We formally confirm that the attack su…
▽ More
Recently, a weight-based attack on stochastic gradient descent inducing overfitting has been proposed. We show that the threat is broader: A task-independent permutation on the initial weights suffices to limit the achieved accuracy to for example 50% on the Fashion MNIST dataset from initially more than $90$%. These findings are confirmed on MNIST and CIFAR. We formally confirm that the attack succeeds with high likelihood and does not depend on the data. Empirically, weight statistics and loss appear unsuspicious, making it hard to detect the attack if the user is not aware. Our paper is thus a call for action to acknowledge the importance of the initial weights in deep learning.
△ Less
Submitted 29 November, 2020; v1 submitted 8 February, 2019;
originally announced February 2019.
