On the Hyperparameters influencing a PINN's generalization beyond the training domain.
Andrea BonfantiRoberto SantanaMarco ElleroBabak GholamiPublished in: CoRR (2023)
Keyphrases
- hyperparameters
- model selection
- bayesian inference
- cross validation
- bayesian framework
- support vector
- gaussian process
- em algorithm
- random sampling
- noise level
- prior information
- closed form
- maximum a posteriori
- incremental learning
- training set
- grid search
- generative model
- maximum likelihood
- gaussian processes
- image processing
- sample size
- incomplete data
- supervised learning
- regularization parameter