Explainability-based Backdoor Attacks Against Graph Neural Networks.
Jing XuMinhui XueStjepan PicekPublished in: WiseML@WiSec (2021)
Keyphrases
- neural network
- pattern recognition
- graph structure
- random walk
- countermeasures
- graph theory
- feed forward
- artificial neural networks
- neural network model
- graph mining
- multi layer
- random graphs
- traffic analysis
- structured data
- graph theoretic
- malicious attacks
- terrorist attacks
- spanning tree
- bipartite graph
- directed acyclic graph
- neural nets
- multilayer perceptron
- connected components
- self organizing maps
- learning algorithm
- data mining