Search Results for author: Tomasz Krzyżek

Found 1 papers, 0 papers with code

Adversarial Examples Detection and Analysis with Layer-wise Autoencoders

no code implementations17 Jun 2020 Bartosz Wójcik, Paweł Morawiecki, Marek Śmieja, Tomasz Krzyżek, Przemysław Spurek, Jacek Tabor

We present a mechanism for detecting adversarial examples based on data representations taken from the hidden layers of the target network.

Cannot find the paper you are looking for? You can Submit a new open access paper.