Graph convolutional neural networks with node transition probability-based message passing and DropNode regularization.
Tien Huu DoDuc Minh NguyenGiannis BekoulisAdrian MunteanuNikos DeligiannisPublished in: Expert Syst. Appl. (2021)
Keyphrases
- message passing
- transition probabilities
- convolutional neural networks
- random walk
- singly connected
- directed graph
- graph structure
- belief propagation
- markov random walk
- markov chain
- distributed systems
- markov models
- graphical models
- undirected graph
- shared memory
- markov random field
- junction tree
- sum product algorithm
- max product
- sum product
- weighted graph
- belief networks
- markov model
- hidden variables
- image segmentation
- post processing
- probabilistic model