Year: 2019
Type: article
Abstract: Training data is crucial in ensuring robust neural inference, and deep neural networks (DNNs) are heavily dependent on this assumption. However, DNNs can be exploited by adversaries that facilitate various attacks. Adversarial defenses include several techniques, some of which happen during the preprocessing stages (i.e., noise filtering, etc.). This article analyzes the impact of some preprocessing filters, and proposes a selective preprocessing method which increases robustness and reduces the computational complexity.
Source: IEEE Design and Test
Authors: Hassan Ali, Faiq Khalid, Hammad Ali Tariq, Muhammad Abdullah Hanif, Rehan Ahmed, Semeen Rehman
Institutions: National University of Sciences and Technology, TU Wien
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2997572583