Hijacking Attacks against Neural Networks by Analyzing Training Data.
Yunjie GeQian WangHuayang HuangQi LiCong WangChao ShenLingchen ZhaoPeipei JiangZheng FangShenyi ZhangPublished in: CoRR (2024)
Keyphrases
- training data
- neural network
- training process
- training set
- learning algorithm
- pattern recognition
- training patterns
- artificial neural networks
- terrorist attacks
- data sets
- countermeasures
- back propagation
- training examples
- multilayer perceptron
- supervised learning
- training dataset
- classification accuracy
- traffic analysis
- decision trees
- neural nets
- test set
- neural network model
- self organizing maps
- test data
- fuzzy logic
- genetic algorithm
- support vector machine
- prior knowledge
- chosen plaintext
- attack detection
- data mining
- malicious attacks
- watermarking technique
- security mechanisms
- active learning
- training algorithm
- generalization error
- labeled data
- recurrent neural networks
- unlabeled data