SGM-PINN: Sampling Graphical Models for Faster Training of Physics-Informed Neural Networks.
John AnticevAli AghdaeiWuxinlin ChengZhuo FengPublished in: CoRR (2024)
Keyphrases
- graphical models
- neural network
- belief propagation
- probabilistic model
- approximate inference
- markov networks
- probabilistic graphical models
- random variables
- probabilistic inference
- bayesian networks
- belief networks
- structure learning
- conditional independence
- conditional random fields
- message passing
- fuzzy logic
- exact inference
- statistical relational learning
- factor graphs
- map inference
- statistical inference
- structured prediction
- nonparametric belief propagation
- markov chain monte carlo
- generative model
- pairwise
- importance sampling
- artificial neural networks
- undirected graphical models
- search algorithm
- graphical structure
- deep architectures
- possibilistic networks
- image segmentation