Building powerful and equivariant graph neural networks with structural message-passing.
Clément VignacAndreas LoukasPascal FrossardPublished in: NeurIPS (2020)
Keyphrases
- message passing
- neural network
- belief propagation
- distributed systems
- junction tree
- markov random field
- distributed shared memory
- singly connected
- probabilistic inference
- factor graphs
- shared memory
- inference in graphical models
- sum product
- graphical models
- tree reweighted
- approximate inference
- back propagation
- max product
- ldpc codes
- energy minimization
- post processing
- bayesian networks
- image processing