Exploring Inherent Backdoors in Deep Learning Models

Guanhong Tao,Siyuan Cheng,Zhenting Wang,Shiqing Ma,Shengwei An,Yingqi Liu,Guangyu Shen,Zhuo Zhang, Yunshu Mao,Xiangyu Zhang

Annual Computer Security Applications Conference（2024）

引用 0|浏览1

关键词

Deep Learning Models,Training Data,Input Samples,Clear Model,Feature Space,Training Procedure,Model Architecture,Class I,Low-level Features,Input Space,Clean Samples,Adversarial Training,Pixel Spacing,Target Label,Small Threshold,Adversarial Attacks,Threat Model,Normal Training,Vision Transformer,Natural Language Processing Models,Attack Success Rate,False Acceptance Rate,Universality Class,Adversarial Perturbations,Sixth Column,Small Batch Size,Training Dataset,Fourth Column,Frequency Space,Frequency Spectrum

AI 理解论文

溯源树

样例

生成溯源树，研究论文发展脉络

Chat Paper

正在生成论文摘要