Massif: Interactive Interpretation of Adversarial Attacks on Deep Learning

01/21/2020
by   Nilaksh Das, et al.
33

Deep neural networks (DNNs) are increasingly powering high-stakes applications such as autonomous cars and healthcare; however, DNNs are often treated as "black boxes" in such applications. Recent research has also revealed that DNNs are highly vulnerable to adversarial attacks, raising serious concerns over deploying DNNs in the real world. To overcome these deficiencies, we are developing Massif, an interactive tool for deciphering adversarial attacks. Massif identifies and interactively visualizes neurons and their connections inside a DNN that are strongly activated or suppressed by an adversarial attack. Massif provides both a high-level, interpretable overview of the effect of an attack on a DNN, and a low-level, detailed description of the affected neurons. These tightly coupled views in Massif help people better understand which input features are most vulnerable or important for correct predictions.

READ FULL TEXT

page 1

page 2

page 3

page 4

page 5

research
09/05/2020

Bluff: Interactively Deciphering Adversarial Attacks on Deep Neural Networks

Deep neural networks (DNNs) are now commonly used in many domains. Howev...
research
12/03/2020

Detecting Trojaned DNNs Using Counterfactual Attributions

We target the problem of detecting Trojans or backdoors in DNNs. Such mo...
research
06/19/2019

Global Adversarial Attacks for Assessing Deep Learning Robustness

It has been shown that deep neural networks (DNNs) may be vulnerable to ...
research
01/13/2023

On the feasibility of attacking Thai LPR systems with adversarial examples

Recent advances in deep neural networks (DNNs) have significantly enhanc...
research
06/05/2023

Adversarial alignment: Breaking the trade-off between the strength of an attack and its relevance to human perception

Deep neural networks (DNNs) are known to have a fundamental sensitivity ...
research
09/23/2020

Detection of Iterative Adversarial Attacks via Counter Attack

Deep neural networks (DNNs) have proven to be powerful tools for process...
research
09/29/2019

Libraries of hidden layer activity patterns can lead to better understanding of operating principles of deep neural networks

Deep neural networks (DNNs) can outperform human brains in specific task...

Please sign up or login with your details

Forgot password? Click here to reset