Do Deep Minds Think Alike? Selective Adversarial Attacks for Fine-Grained Manipulation of Multiple Deep Neural Networks

by   Zain Khan, et al.

Recent works have demonstrated the existence of adversarial examples targeting a single machine learning system. In this paper we ask a simple but fundamental question of "selective fooling": given multiple machine learning systems assigned to solve the same classification problem and taking the same input signal, is it possible to construct a perturbation to the input signal that manipulates the outputs of these multiple machine learning systems simultaneously in arbitrary pre-defined ways? For example, is it possible to selectively fool a set of "enemy" machine learning systems but does not fool the other "friend" machine learning systems? The answer to this question depends on the extent to which these different machine learning systems "think alike". We formulate the problem of "selective fooling" as a novel optimization problem, and report on a series of experiments on the MNIST dataset. Our preliminary findings from these experiments show that it is in fact very easy to selectively manipulate multiple MNIST classifiers simultaneously, even when the classifiers are identical in their architectures, training algorithms and training datasets except for random initialization during training. This suggests that two nominally equivalent machine learning systems do not in fact "think alike" at all, and opens the possibility for many novel applications and deeper understandings of the working principles of deep neural networks.


page 7

page 8


Adversarial attacks on an optical neural network

Adversarial attacks have been extensively investigated for machine learn...

Learning to Extrapolate: A Transductive Approach

Machine learning systems, especially with overparameterized deep neural ...

Examining Redundancy in the Context of Safe Machine Learning

This paper describes a set of experiments with neural network classifier...

Algorithmic Censoring in Dynamic Learning Systems

Dynamic learning systems subject to selective labeling exhibit censoring...

On the combined effect of class imbalance and concept complexity in deep learning

Structural concept complexity, class overlap, and data scarcity are some...

Modern Machine and Deep Learning Systems as a way to achieve Man-Computer Symbiosis

Man-Computer Symbiosis (MCS) was originally envisioned by the famous com...

Design of Supervision-Scalable Learning Systems: Methodology and Performance Benchmarking

The design of robust learning systems that offer stable performance unde...

Please sign up or login with your details

Forgot password? Click here to reset