Improving Molecular Graph Neural Network Explainability with Orthonormalization and Induced Sparsity.
Ryan HendersonDjork-Arné ClevertFloriane MontanariPublished in: ICML (2021)
Keyphrases
- neural network
- graph representation
- graph theory
- fuzzy logic
- graph structure
- artificial neural networks
- graph matching
- back propagation
- directed graph
- genetic algorithm
- weighted graph
- feed forward
- random walk
- pattern recognition
- high dimensional
- structured data
- graph theoretic
- three dimensional
- fuzzy neural network
- network architecture
- bipartite graph
- neural nets
- associative memory
- sparse representation
- neural network model
- fault diagnosis
- bayesian networks
- training algorithm
- multilayer perceptron
- graph databases
- graph partitioning
- fuzzy artmap
- chemical reactions