Evasion Attacks to Graph Neural Networks via Influence Function.
Binghui WangTianxiang ZhouMinhua LinPan ZhouAng LiMeng PangCai FuHai LiYiran ChenPublished in: CoRR (2020)
Keyphrases
- neural network
- countermeasures
- pattern recognition
- graph theory
- graph representation
- connected components
- recurrent neural networks
- back propagation
- genetic algorithm
- neural nets
- random walk
- artificial neural networks
- watermarking algorithm
- watermarking scheme
- graph theoretic
- function approximators
- denial of service attacks
- multi layer
- weighted graph
- multilayer perceptron
- graph matching
- structured data
- level set
- reinforcement learning
- image segmentation
- social networks