Deep Learning Backdoors

2020/07/16 by Shaofeng Li, Shiqing Ma, Li, Shaofeng +5 · 1 citation
Computer Science · #Adversarial Robustness in Machine Learning #Cryptography and Security (cs.CR) #Digital Media Forensic Detection #FOS: Computer and information sciences #Generative Adversarial Networks and Image Synthesis #Machine Learning (cs.LG)

paper · pdf · doi:10.48550/arxiv.2007.08273

openalex publication_date 2020/07/16 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Intuitively, a backdoor attack against Deep Neural Networks (DNNs) is to inject hidden malicious behaviors into DNNs such that the backdoor model behaves legitimately for benign inputs, yet invokes a predefined malicious behavior when its input contains a malicious trigger. The trigger can take a plethora of forms, including a special object present in the image (e.g., a yellow pad), a shape filled with custom textures (e.g., logos with particular colors) or even image-wide stylizations with special filters (e.g., images altered by Nashville or Gotham filters). These filters can be applied to the original image by replacing or perturbing a set of image pixels.

Cited by

• Backdoor Scanning for Deep Neural Networks through K-Arm Optimization

Related

• Defending against Backdoor Attack on Deep Neural Networks
• Backdoor Attack in the Physical World
• An Invisible Backdoor Attack Based On Semantic Feature
• BATT: Backdoor Attack with Transformation-based Triggers
• Rethinking the Trigger of Backdoor Attack