Exploració per altres contribucions "Kalita, Jugal"
Ara es mostren els items 1-1 de 1
-
Defense-GAN for text: protecting text classifiers against Adversarial Attacks
(Universitat Politècnica de Catalunya, 2019-06-18)
Treball Final de Grau
Accés restringit per decisió de l'autorIn computer vision, Defense-GAN is a framework that leverages the capability of generative models to defend neural networks (NNs) against adversarial attacks. This paper investigates how Defense-GAN will behave when it is ...