Multiscale graph neural network autoencoders for interpretable scientific machine learning.
Shivam BarweyVarun ShankarVenkatasubramanian ViswanathanRomit MaulikPublished in: J. Comput. Phys. (2023)
Keyphrases
- multiscale
- machine learning
- neural network
- pattern recognition
- denoising
- data mining
- artificial intelligence
- image processing
- information extraction
- coarse to fine
- computer vision
- graph structure
- hidden layer
- image representation
- genetic algorithm
- graph representation
- machine learning methods
- artificial neural networks
- graph model
- back propagation
- wavelet transform
- learning algorithm
- image segmentation
- recurrent neural networks
- random walk
- graph theory
- computational intelligence
- scientific data
- neural nets
- weighted graph
- knowledge acquisition
- multi layer perceptron
- feedforward neural networks
- directed graph
- self organizing maps
- structured data
- decision trees
- edge detection
- natural language processing
- neural network is trained
- feature selection
- reinforcement learning
- multiresolution
- support vector machine
- directed acyclic graph
- supervised learning
- classification rules
- graphical models
- scale space
- image fusion
- graph matching
- neural network model