Improving Molecular Graph Neural Network Explainability with Orthonormalization and Induced Sparsity.
Ryan HendersonDjork-Arné ClevertFloriane MontanariPublished in: CoRR (2021)
Keyphrases
- neural network
- artificial neural networks
- bipartite graph
- graph representation
- random walk
- back propagation
- bp neural network
- directed graph
- weighted graph
- graph structure
- graph theory
- structured data
- genetic algorithm
- neural network model
- three dimensional
- self organizing maps
- multi layer
- pattern recognition
- high dimensional
- multilayer perceptron
- connected components
- feed forward neural networks
- sparse representation
- fuzzy logic
- fault diagnosis
- graph matching
- graph model
- graph databases
- graph partitioning
- edge weights
- graph based algorithm