GraphGuard: Detecting and Counteracting Training Data Misuse in Graph Neural Networks.
Bang WuHe ZhangXiangwen YangShuo WangMinhui XueShirui PanXingliang YuanPublished in: CoRR (2023)
Keyphrases
- training data
- neural network
- training process
- decision trees
- data sets
- artificial neural networks
- graph theory
- training examples
- learning algorithm
- training set
- neural network model
- recurrent neural networks
- classification models
- directed graph
- graph model
- graph structure
- weighted graph
- training samples
- anomaly detection
- fuzzy logic
- misuse detection
- directed acyclic graph
- machine learning
- structured data
- genetic algorithm
- pattern recognition
- supervised learning
- multi layer
- learned from training data
- graph based algorithm
- graph theoretic
- graph representation
- spanning tree
- noisy data
- test set
- class labels
- bipartite graph
- test data
- intrusion detection system
- graph matching
- connected components