Bounding the Expected Robustness of Graph Neural Networks Subject to Node Feature Attacks.
Yassine AbbahaddouSofiane EnnadirJohannes F. LutzeyerMichalis VazirgiannisHenrik BoströmPublished in: CoRR (2024)
Keyphrases
- neural network
- graph structure
- directed graph
- edge weights
- pattern recognition
- feature vectors
- digital image watermarking
- graph representation
- artificial neural networks
- upper bound
- finding the shortest path
- undirected graph
- nodes of a graph
- graph theory
- graph model
- watermarking scheme
- root node
- random walk
- image watermarking scheme
- tree traversal
- semi fragile watermarking
- weighted graph
- bipartite graph
- self organizing maps
- image features
- genetic algorithm
- directed acyclic graph
- countermeasures
- graph partitioning
- path length
- watermark embedding
- structured data
- back propagation
- feature set
- geometric attacks
- fuzzy logic
- lower bound