PatchBackdoor: Backdoor Attack against Deep Neural Networks without Model Modification

Backdoor attack is a major threat to deep learning systems in safety-critical scenarios, which aims to trigger misbehavior of neural network models under attacker-controlled conditions. However, most backdoor attacks have to modify the neural network models through training with poisoned data and/or...

Ausführliche Beschreibung

Bibliographische Detailangaben

Veröffentlicht in:	arXiv.org 2023-08
Hauptverfasser:	Yuan, Yizhen, Kong, Rui, Xie, Shenghao, Li, Yuanchun, Liu, Yunxin
Format:	Artikel
Sprache:	eng
Schlagworte:	Artificial neural networks Computer Science - Computer Vision and Pattern Recognition Computer Science - Cryptography and Security Computer Science - Learning Deep learning Machine learning Neural networks Safety critical Training
Online-Zugang:	Volltext
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!