Text backdoor detection using an interpretable RNN abstract model

• Main Authors: FAN, Ming, SI, Ziliang, XIE, Xiaofei, LIU, Yang, LIU, Ting
• Format: text
• Language: English
• Published: Institutional Knowledge at Singapore Management University 2021
• Subjects: Training; Recurrent neural networks; Task analysis; Motion pictures; Data models; Analytical models; Sentiment analysis; Text backdoor detection; RNN; model abstraction; interpretation; OS and Networks; Software Engineering
• Online Access: https://ink.library.smu.edu.sg/sis_research/7118
• Institution: Singapore Management University

Deep neural networks (DNNs) are known to be inherently vulnerable to malicious attacks such as the adversarial attack and the backdoor attack. The former is crafted by adding small perturbations to benign inputs so as to fool a DNN. The latter generally embeds a hidden pattern in a DNN by poisoning...