Zombori, Dániel / Bánhelyi, Balázs: Attacking Neural …

Előadás címe: Attacking Neural Network Verifier with Adversarial Examples

Helyszín: Online, https://meet.google.com/ete-gaku-aex

Időpont: 2020.06.29., 15:15 – 15:35

Kivonat: One of the hottest topics in present artificial intelligence research is to understand the phenomenon of adversarial examples for machine learning techniques applying artificial neural networks  The typical problem is that in many practical cases, e.g. in image classification, after the proper training of the network, surprisingly similar to the actual  pictur result in a wrong denial decision.  That is, an attacker may easily mislead a well-performing image classification system by altering some pixels. However, proving that a network will have correct output when changing some regions of the images, is quite challenging. Because of this, only a few works targeted this problem. Although there are an increasing number of studies on this field, reliable robustness evaluation is still an open issue. In this work, we will attempt to contribute in this direction. We will present new interval arithmetic based algorithms to provide adversarial example free image patches for trained artificial neural networks. surprisingly similar to the actual  pictur result in a wrong denial decision