- filenameFilename containing the trained data.
C++ Type:FileName
Controllable:No
Description:Filename containing the trained data.
- trainerThe SurrogateTrainer object. If this is specified the trainer data is automatically gathered and available in this SurrogateModel object.
C++ Type:UserObjectName
Controllable:No
Description:The SurrogateTrainer object. If this is specified the trainer data is automatically gathered and available in this SurrogateModel object.
GaussianProcess
Computes and evaluates Gaussian Process surrogate model.
The theory and use this object is provided within a discussion of the GaussianProcessTrainer training object.
A desirable aspect of Gaussian process modeling is that in addition to returning a predicted value at the evaluation point, it can also provide a measure of uncertainty in the form of a standard deviation. To facilitate this an overloaded evaluate() function which sets the standard deviation by reference is provided.
GaussianProcess::evaluate(const std::vector<Real> & x, Real & std_dev) const
Input Parameters
- control_tagsAdds user-defined labels for accessing object parameters via control logic.
C++ Type:std::vector<std::string>
Controllable:No
Description:Adds user-defined labels for accessing object parameters via control logic.
- enableTrueSet the enabled status of the MooseObject.
Default:True
C++ Type:bool
Controllable:No
Description:Set the enabled status of the MooseObject.
Advanced Parameters
Input Files
- (modules/stochastic_tools/test/tests/surrogates/gaussian_process/GP_squared_exponential.i)
- (modules/stochastic_tools/examples/surrogates/gaussian_process/gaussian_process_uniform_1D_tuned.i)
- (modules/stochastic_tools/test/tests/surrogates/gaussian_process/GP_exponential_tuned.i)
- (modules/stochastic_tools/test/tests/surrogates/gaussian_process/GP_exponential.i)
- (modules/stochastic_tools/examples/surrogates/gaussian_process/gaussian_process_uniform_2D.i)
- (modules/stochastic_tools/test/tests/surrogates/gaussian_process/GP_squared_exponential_tuned.i)
- (modules/stochastic_tools/test/tests/surrogates/gaussian_process/GP_squared_exponential_testing.i)
- (modules/stochastic_tools/examples/surrogates/gaussian_process/GP_normal.i)
- (modules/stochastic_tools/examples/surrogates/gaussian_process/gaussian_process_uniform_1D.i)
- (modules/stochastic_tools/test/tests/surrogates/gaussian_process/GP_Matern_half_int_tuned.i)
- (modules/stochastic_tools/test/tests/surrogates/gaussian_process/GP_Matern_half_int.i)
- (modules/stochastic_tools/examples/surrogates/gaussian_process/gaussian_process_uniform_2D_tuned.i)
(modules/stochastic_tools/src/surrogates/GaussianProcess.C)
// This file is part of the MOOSE framework
// https://www.mooseframework.org
//
// All rights reserved, see COPYRIGHT for full restrictions
// https://github.com/idaholab/moose/blob/master/COPYRIGHT
//
// Licensed under LGPL 2.1, please see LICENSE for details
// https://www.gnu.org/licenses/lgpl-2.1.html
#include "GaussianProcess.h"
#include "Sampler.h"
#include "CovarianceFunctionBase.h"
registerMooseObject("StochasticToolsApp", GaussianProcess);
InputParameters
GaussianProcess::validParams()
{
InputParameters params = SurrogateModel::validParams();
params.addClassDescription("Computes and evaluates Gaussian Process surrogate model.");
return params;
}
GaussianProcess::GaussianProcess(const InputParameters & parameters)
: SurrogateModel(parameters),
CovarianceInterface(parameters),
_gp_handler(
isParamValid("trainer")
? dynamic_cast<GaussianProcessTrainer &>(getSurrogateTrainer("trainer")).gpHandler()
: setModelData<StochasticTools::GaussianProcessHandler>("_gp_handler")),
_training_params(getModelData<RealEigenMatrix>("_training_params"))
{
}
void
GaussianProcess::setupCovariance(UserObjectName covar_name)
{
if (_gp_handler.getCovarFunctionPtr() != nullptr)
::mooseError("Attempting to redefine covariance function using setupCovariance.");
_gp_handler.linkCovarianceFunction(getCovarianceFunctionByName(covar_name));
}
Real
GaussianProcess::evaluate(const std::vector<Real> & x) const
{
// Overlaod for evaluate to maintain general compatibility. Only returns mean
Real dummy = 0;
return this->evaluate(x, dummy);
}
Real
GaussianProcess::evaluate(const std::vector<Real> & x, Real & std_dev) const
{
unsigned int _n_params = _training_params.cols();
unsigned int _num_tests = 1;
mooseAssert(x.size() == _n_params,
"Number of parameters provided for evaluation does not match number of parameters "
"used for training.");
RealEigenMatrix test_points(_num_tests, _n_params);
for (unsigned int ii = 0; ii < _n_params; ++ii)
test_points(0, ii) = x[ii];
_gp_handler.getParamStandardizer().getStandardized(test_points);
RealEigenMatrix K_train_test(_training_params.rows(), test_points.rows());
_gp_handler.getCovarFunction().computeCovarianceMatrix(
K_train_test, _training_params, test_points, false);
RealEigenMatrix K_test(test_points.rows(), test_points.rows());
_gp_handler.getCovarFunction().computeCovarianceMatrix(K_test, test_points, test_points, true);
// Compute the predicted mean value (centered)
RealEigenMatrix pred_value = (K_train_test.transpose() * _gp_handler.getKResultsSolve());
// De-center/scale the value and store for return
_gp_handler.getDataStandardizer().getDestandardized(pred_value);
RealEigenMatrix pred_var =
K_test - (K_train_test.transpose() * _gp_handler.getKCholeskyDecomp().solve(K_train_test));
// Vairance computed, take sqrt for standard deviation, scale up by training data std and store
RealEigenMatrix std_dev_mat = pred_var.array().sqrt();
_gp_handler.getDataStandardizer().getDescaled(std_dev_mat);
std_dev = std_dev_mat(0, 0);
return pred_value(0, 0);
}
(modules/stochastic_tools/test/tests/surrogates/gaussian_process/GP_squared_exponential.i)
[StochasticTools]
[]
[Distributions]
[k_dist]
type = Uniform
lower_bound = 1
upper_bound = 10
[]
[q_dist]
type = Uniform
lower_bound = 9000
upper_bound = 11000
[]
[]
[Samplers]
[train_sample]
type = MonteCarlo
num_rows = 10
distributions = 'k_dist q_dist'
execute_on = PRE_MULTIAPP_SETUP
[]
[test_sample]
type = MonteCarlo
num_rows = 100
distributions = 'k_dist q_dist'
execute_on = PRE_MULTIAPP_SETUP
[]
[]
[MultiApps]
[sub]
type = SamplerFullSolveMultiApp
input_files = sub.i
sampler = train_sample
[]
[]
[Controls]
[cmdline]
type = MultiAppCommandLineControl
multi_app = sub
sampler = train_sample
param_names = 'Materials/conductivity/prop_values Kernels/source/value'
[]
[]
[Transfers]
[data]
type = SamplerReporterTransfer
from_multi_app = sub
sampler = train_sample
stochastic_reporter = results
from_reporter = 'avg/value'
[]
[]
[Reporters]
[results]
type = StochasticReporter
parallel_type = ROOT
[]
[samp_avg]
type = EvaluateSurrogate
model = GP_avg
sampler = test_sample
evaluate_std = 'true'
parallel_type = ROOT
execute_on = final
[]
[train_avg]
type = EvaluateSurrogate
model = GP_avg
sampler = train_sample
evaluate_std = 'true'
parallel_type = ROOT
execute_on = final
[]
[]
[VectorPostprocessors]
[hyperparams]
type = GaussianProcessData
gp_name = 'GP_avg'
execute_on = final
[]
[]
[Trainers]
[GP_avg_trainer]
type = GaussianProcessTrainer
execute_on = timestep_end
covariance_function = 'covar' #Choose a squared exponential for the kernel
standardize_params = 'true' #Center and scale the training params
standardize_data = 'true' #Center and scale the training data
sampler = train_sample
response = results/data:avg:value
[]
[]
[Surrogates]
[GP_avg]
type = GaussianProcess
trainer = GP_avg_trainer
[]
[]
[Covariance]
[covar]
type=SquaredExponentialCovariance
signal_variance = 1 #Use a signal variance of 1 in the kernel
noise_variance = 1e-6 #A small amount of noise can help with numerical stability
length_factor = '0.38971 0.38971' #Select a length factor for each parameter (k and q)
[]
[]
[Outputs]
[out]
type = CSV
execute_on = FINAL
[]
[]
(modules/stochastic_tools/examples/surrogates/gaussian_process/gaussian_process_uniform_1D_tuned.i)
[StochasticTools]
[]
[Distributions]
[k_dist]
type = Uniform
lower_bound = 1
upper_bound = 10
[]
[q_dist]
type = Uniform
lower_bound = 9000
upper_bound = 11000
[]
[L_dist]
type = Uniform
lower_bound = 0.01
upper_bound = 0.05
[]
[Tinf_dist]
type = Uniform
lower_bound = 290
upper_bound = 310
[]
[]
[Samplers]
[train_sample]
type = MonteCarlo
num_rows = 6
distributions = 'q_dist'
execute_on = PRE_MULTIAPP_SETUP
[]
[cart_sample]
type = CartesianProduct
linear_space_items = '9000 20 100'
execute_on = initial
[]
[]
[MultiApps]
[sub]
type = SamplerFullSolveMultiApp
input_files = sub.i
sampler = train_sample
[]
[]
[Controls]
[cmdline]
type = MultiAppCommandLineControl
multi_app = sub
sampler = train_sample
param_names = 'Kernels/source/value'
[]
[]
[Transfers]
[data]
type = SamplerReporterTransfer
from_multi_app = sub
sampler = train_sample
stochastic_reporter = results
from_reporter = 'avg/value'
[]
[]
[Reporters]
[results]
type = StochasticReporter
[]
[]
[Trainers]
[GP_avg_trainer]
type = GaussianProcessTrainer
execute_on = timestep_end
covariance_function = 'rbf'
standardize_params = 'true' #Center and scale the training params
standardize_data = 'true' #Center and scale the training data
sampler = train_sample
response = results/data:avg:value
tao_options = '-tao_bncg_type kd'
tune_parameters = ' signal_variance length_factor'
tuning_min = ' 1e-9 1e-9'
tuning_max = ' 1e16 1e16'
tuning_algorithm = 'tao'
[]
[]
[Covariance]
[rbf]
type=SquaredExponentialCovariance
signal_variance = 1 #Use a signal variance of 1 in the kernel
noise_variance = 1e-3 #A small amount of noise can help with numerical stability
length_factor = '0.38971' #Select a length factor for each parameter (k and q)
[]
[]
[Surrogates]
[gauss_process_avg]
type = GaussianProcess
trainer = 'GP_avg_trainer'
[]
[]
# # Computing statistics
[VectorPostprocessors]
[hyperparams]
type = GaussianProcessData
gp_name = 'gauss_process_avg'
execute_on = final
[]
[]
[Reporters]
[cart_avg]
type = EvaluateSurrogate
model = gauss_process_avg
sampler = cart_sample
evaluate_std = 'true'
parallel_type = ROOT
execute_on = final
[]
[train_avg]
type = EvaluateSurrogate
model = gauss_process_avg
sampler = train_sample
evaluate_std = 'true'
parallel_type = ROOT
execute_on = final
[]
[]
[Outputs]
csv = true
execute_on = FINAL
[]
(modules/stochastic_tools/test/tests/surrogates/gaussian_process/GP_exponential_tuned.i)
[StochasticTools]
[]
[Distributions]
[k_dist]
type = Uniform
lower_bound = 1
upper_bound = 10
[]
[q_dist]
type = Uniform
lower_bound = 9000
upper_bound = 11000
[]
[]
[Samplers]
[train_sample]
type = MonteCarlo
num_rows = 10
distributions = 'k_dist q_dist'
execute_on = PRE_MULTIAPP_SETUP
[]
[test_sample]
type = MonteCarlo
num_rows = 100
distributions = 'k_dist q_dist'
execute_on = PRE_MULTIAPP_SETUP
[]
[]
[MultiApps]
[sub]
type = SamplerFullSolveMultiApp
input_files = sub.i
sampler = train_sample
[]
[]
[Controls]
[cmdline]
type = MultiAppCommandLineControl
multi_app = sub
sampler = train_sample
param_names = 'Materials/conductivity/prop_values Kernels/source/value'
[]
[]
[Transfers]
[data]
type = SamplerReporterTransfer
from_multi_app = sub
sampler = train_sample
stochastic_reporter = results
from_reporter = 'avg/value'
[]
[]
[Reporters]
[results]
type = StochasticReporter
parallel_type = ROOT
[]
[samp_avg]
type = EvaluateSurrogate
model = GP_avg
sampler = test_sample
evaluate_std = 'true'
parallel_type = ROOT
execute_on = final
[]
[train_avg]
type = EvaluateSurrogate
model = GP_avg
sampler = train_sample
evaluate_std = 'true'
parallel_type = ROOT
execute_on = final
[]
[]
[VectorPostprocessors]
[hyperparams]
type = GaussianProcessData
gp_name = 'GP_avg'
execute_on = final
[]
[]
[Trainers]
[GP_avg_trainer]
type = GaussianProcessTrainer
execute_on = timestep_end
covariance_function = 'covar' #Choose an exponential for the kernel
standardize_params = 'true' #Center and scale the training params
standardize_data = 'true' #Center and scale the training data
sampler = train_sample
response = results/data:avg:value
tao_options = '-tao_bncg_type ssml_bfgs'
tune_parameters = 'signal_variance length_factor'
tuning_min = ' 1e-9 1e-9'
tuning_max = ' 1e16 1e16'
tuning_algorithm = 'tao'
[]
[]
[Surrogates]
[GP_avg]
type = GaussianProcess
trainer = GP_avg_trainer
[]
[]
[Covariance]
[covar]
type=ExponentialCovariance
gamma = 2 #Define the exponential factor
signal_variance = 1 #Use a signal variance of 1 in the kernel
noise_variance = 1e-3 #A small amount of noise can help with numerical stability
length_factor = '0.551133 0.551133' #Select a length factor for each parameter (k and q)
[]
[]
[Outputs]
[out]
type = CSV
execute_on = FINAL
[]
[]
(modules/stochastic_tools/test/tests/surrogates/gaussian_process/GP_exponential.i)
[StochasticTools]
[]
[Distributions]
[k_dist]
type = Uniform
lower_bound = 1
upper_bound = 10
[]
[q_dist]
type = Uniform
lower_bound = 9000
upper_bound = 11000
[]
[]
[Samplers]
[train_sample]
type = MonteCarlo
num_rows = 10
distributions = 'k_dist q_dist'
execute_on = PRE_MULTIAPP_SETUP
[]
[test_sample]
type = MonteCarlo
num_rows = 100
distributions = 'k_dist q_dist'
execute_on = PRE_MULTIAPP_SETUP
[]
[]
[MultiApps]
[sub]
type = SamplerFullSolveMultiApp
input_files = sub.i
sampler = train_sample
[]
[]
[Controls]
[cmdline]
type = MultiAppCommandLineControl
multi_app = sub
sampler = train_sample
param_names = 'Materials/conductivity/prop_values Kernels/source/value'
[]
[]
[Transfers]
[data]
type = SamplerReporterTransfer
from_multi_app = sub
sampler = train_sample
stochastic_reporter = results
from_reporter = 'avg/value'
[]
[]
[Reporters]
[results]
type = StochasticReporter
parallel_type = ROOT
[]
[samp_avg]
type = EvaluateSurrogate
model = GP_avg
sampler = test_sample
evaluate_std = 'true'
parallel_type = ROOT
execute_on = final
[]
[train_avg]
type = EvaluateSurrogate
model = GP_avg
sampler = train_sample
evaluate_std = 'true'
parallel_type = ROOT
execute_on = final
[]
[]
[VectorPostprocessors]
[hyperparams]
type = GaussianProcessData
gp_name = 'GP_avg'
execute_on = final
[]
[]
[Trainers]
[GP_avg_trainer]
type = GaussianProcessTrainer
execute_on = timestep_end
covariance_function = 'covar' #Choose an exponential for the kernel
standardize_params = 'true' #Center and scale the training params
standardize_data = 'true' #Center and scale the training data
sampler = train_sample
response = results/data:avg:value
[]
[]
[Surrogates]
[GP_avg]
type = GaussianProcess
trainer = GP_avg_trainer
[]
[]
[Covariance]
[covar]
type=ExponentialCovariance
gamma = 1 #Define the exponential factor
signal_variance = 1 #Use a signal variance of 1 in the kernel
noise_variance = 1e-6 #A small amount of noise can help with numerical stability
length_factor = '0.551133 0.551133' #Select a length factor for each parameter (k and q)
[]
[]
[Outputs]
[out]
type = CSV
execute_on = FINAL
[]
[]
(modules/stochastic_tools/examples/surrogates/gaussian_process/gaussian_process_uniform_2D.i)
[StochasticTools]
[]
[Distributions]
[k_dist]
type = Uniform
lower_bound = 1
upper_bound = 10
[]
[q_dist]
type = Uniform
lower_bound = 9000
upper_bound = 11000
[]
[]
[Samplers]
[train_sample]
type = MonteCarlo
num_rows = 50
distributions = 'k_dist q_dist'
execute_on = PRE_MULTIAPP_SETUP
[]
[cart_sample]
type = CartesianProduct
linear_space_items = '1 0.09 100
9000 20 100 '
execute_on = initial
[]
[]
[MultiApps]
[sub]
type = SamplerFullSolveMultiApp
input_files = sub.i
sampler = train_sample
[]
[]
[Controls]
[cmdline]
type = MultiAppCommandLineControl
multi_app = sub
sampler = train_sample
param_names = 'Materials/conductivity/prop_values Kernels/source/value'
[]
[]
[Transfers]
[data]
type = SamplerReporterTransfer
from_multi_app = sub
sampler = train_sample
stochastic_reporter = results
from_reporter = 'avg/value'
[]
[]
[Reporters]
[results]
type = StochasticReporter
[]
[train_avg]
type = EvaluateSurrogate
model = GP_avg
sampler = train_sample
evaluate_std = 'true'
parallel_type = ROOT
execute_on = final
[]
[cart_avg]
type = EvaluateSurrogate
model = GP_avg
sampler = cart_sample
evaluate_std = 'true'
parallel_type = ROOT
execute_on = final
[]
[]
[Trainers]
[GP_avg_trainer]
type = GaussianProcessTrainer
execute_on = timestep_end
covariance_function = 'rbf'
standardize_params = 'true' #Center and scale the training params
standardize_data = 'true' #Center and scale the training data
sampler = train_sample
response = results/data:avg:value
[]
[]
[Covariance]
[rbf]
type=SquaredExponentialCovariance
signal_variance = 1 #Use a signal variance of 1 in the kernel
noise_variance = 1e-6 #A small amount of noise can help with numerical stability
length_factor = '0.38971 0.38971' #Select a length factor for each parameter (k and q)
[]
[]
[Surrogates]
[GP_avg]
type = GaussianProcess
trainer = 'GP_avg_trainer'
[]
[]
[VectorPostprocessors]
[hyperparams]
type = GaussianProcessData
gp_name = 'GP_avg'
execute_on = final
[]
[]
[Outputs]
[out]
type = CSV
execute_on = FINAL
[]
[]
(modules/stochastic_tools/test/tests/surrogates/gaussian_process/GP_squared_exponential_tuned.i)
[StochasticTools]
[]
[Distributions]
[k_dist]
type = Uniform
lower_bound = 1
upper_bound = 10
[]
[q_dist]
type = Uniform
lower_bound = 9000
upper_bound = 11000
[]
[]
[Samplers]
[train_sample]
type = MonteCarlo
num_rows = 10
distributions = 'k_dist q_dist'
execute_on = PRE_MULTIAPP_SETUP
[]
[test_sample]
type = MonteCarlo
num_rows = 100
distributions = 'k_dist q_dist'
execute_on = PRE_MULTIAPP_SETUP
[]
[]
[MultiApps]
[sub]
type = SamplerFullSolveMultiApp
input_files = sub.i
sampler = train_sample
[]
[]
[Controls]
[cmdline]
type = MultiAppCommandLineControl
multi_app = sub
sampler = train_sample
param_names = 'Materials/conductivity/prop_values Kernels/source/value'
[]
[]
[Transfers]
[data]
type = SamplerReporterTransfer
from_multi_app = sub
sampler = train_sample
stochastic_reporter = results
from_reporter = 'avg/value'
[]
[]
[Reporters]
[results]
type = StochasticReporter
parallel_type = ROOT
[]
[samp_avg]
type = EvaluateSurrogate
model = GP_avg
sampler = test_sample
evaluate_std = 'true'
parallel_type = ROOT
execute_on = final
[]
[train_avg]
type = EvaluateSurrogate
model = GP_avg
sampler = train_sample
evaluate_std = 'true'
parallel_type = ROOT
execute_on = final
[]
[]
[VectorPostprocessors]
[hyperparams]
type = GaussianProcessData
gp_name = 'GP_avg'
execute_on = final
[]
[]
[Trainers]
[GP_avg_trainer]
type = GaussianProcessTrainer
execute_on = timestep_end
covariance_function = 'covar' #Choose a squared exponential for the kernel
standardize_params = 'true' #Center and scale the training params
standardize_data = 'true' #Center and scale the training data
sampler = train_sample
response = results/data:avg:value
tao_options = '-tao_bncg_type ssml_bfgs'
tune_parameters = ' signal_variance length_factor'
tuning_min = ' 1e-9 1e-9'
tuning_max = ' 1e16 1e16'
tuning_algorithm = 'tao'
show_tao=true
[]
[]
[Surrogates]
[GP_avg]
type = GaussianProcess
trainer = GP_avg_trainer
[]
[]
[Covariance]
[covar]
type=SquaredExponentialCovariance
signal_variance = 1 #Use a signal variance of 1 in the kernel
noise_variance = 1e-3 #A small amount of noise can help with numerical stability
length_factor = '0.38971 0.38971' #Select a length factor for each parameter (k and q)
[]
[]
[Outputs]
[out]
type = CSV
execute_on = FINAL
[]
[]
(modules/stochastic_tools/test/tests/surrogates/gaussian_process/GP_squared_exponential_testing.i)
[StochasticTools]
[]
[Distributions]
[k_dist]
type = Uniform
lower_bound = 1
upper_bound = 10
[]
[q_dist]
type = Uniform
lower_bound = 9000
upper_bound = 11000
[]
[]
[Samplers]
[test_sample]
type = MonteCarlo
num_rows = 100
distributions = 'k_dist q_dist'
execute_on = PRE_MULTIAPP_SETUP
[]
[]
[Reporters]
[samp_avg]
type = EvaluateSurrogate
model = GP_avg
sampler = test_sample
evaluate_std = 'true'
parallel_type = ROOT
execute_on = final
[]
[]
[Surrogates]
[GP_avg]
type = GaussianProcess
filename = 'gauss_process_training_GP_avg_trainer.rd'
[]
[]
[Outputs]
[out]
type = CSV
execute_on = FINAL
[]
[]
(modules/stochastic_tools/examples/surrogates/gaussian_process/GP_normal.i)
[StochasticTools]
[]
[Distributions]
[k_dist]
type = TruncatedNormal
mean = 5
standard_deviation = 2
lower_bound = 0
[]
[q_dist]
type = TruncatedNormal
mean = 10000
standard_deviation = 500
lower_bound = 0
[]
[L_dist]
type = TruncatedNormal
mean = 0.03
standard_deviation = 0.01
lower_bound = 0
[]
[Tinf_dist]
type = TruncatedNormal
mean = 300
standard_deviation = 10
lower_bound = 0
[]
[]
[Samplers]
[sample]
type = MonteCarlo
num_rows = 100000
distributions = 'k_dist q_dist L_dist Tinf_dist'
execute_on = initial
[]
[]
[Surrogates]
[GP_avg]
type = GaussianProcess
filename = 'GP_training_normal_GP_avg.rd'
[]
[]
# Computing statistics
[VectorPostprocessors]
[GP_avg_hyperparams]
type = GaussianProcessData
gp_name = 'GP_avg'
[]
[]
[Outputs]
csv = true
[]
(modules/stochastic_tools/examples/surrogates/gaussian_process/gaussian_process_uniform_1D.i)
[StochasticTools]
[]
[Distributions]
[k_dist]
type = Uniform
lower_bound = 1
upper_bound = 10
[]
[q_dist]
type = Uniform
lower_bound = 9000
upper_bound = 11000
[]
[L_dist]
type = Uniform
lower_bound = 0.01
upper_bound = 0.05
[]
[Tinf_dist]
type = Uniform
lower_bound = 290
upper_bound = 310
[]
[]
[Samplers]
[train_sample]
type = MonteCarlo
num_rows = 6
distributions = 'q_dist'
execute_on = PRE_MULTIAPP_SETUP
[]
[cart_sample]
type = CartesianProduct
linear_space_items = '9000 20 100'
execute_on = initial
[]
[]
[MultiApps]
[sub]
type = SamplerFullSolveMultiApp
input_files = sub.i
sampler = train_sample
[]
[]
[Controls]
[cmdline]
type = MultiAppCommandLineControl
multi_app = sub
sampler = train_sample
param_names = 'Kernels/source/value'
[]
[]
[Transfers]
[data]
type = SamplerReporterTransfer
from_multi_app = sub
sampler = train_sample
stochastic_reporter = results
from_reporter = 'avg/value'
[]
[]
[Reporters]
[results]
type = StochasticReporter
[]
[cart_avg]
type = EvaluateSurrogate
model = gauss_process_avg
sampler = cart_sample
evaluate_std = 'true'
parallel_type = ROOT
execute_on = final
[]
[train_avg]
type = EvaluateSurrogate
model = gauss_process_avg
sampler = train_sample
evaluate_std = 'true'
parallel_type = ROOT
execute_on = final
[]
[]
[Trainers]
[GP_avg_trainer]
type = GaussianProcessTrainer
execute_on = timestep_end
sampler = train_sample
response = results/data:avg:value
covariance_function = 'rbf'
standardize_params = 'true' #Center and scale the training params
standardize_data = 'true' #Center and scale the training data
[]
[]
[Covariance]
[rbf]
type=SquaredExponentialCovariance
signal_variance = 1 #Use a signal variance of 1 in the kernel
noise_variance = 1e-3 #A small amount of noise can help with numerical stability
length_factor = '0.38971' #Select a length factor for each parameter (k and q)
[]
[]
[Surrogates]
[gauss_process_avg]
type = GaussianProcess
trainer = 'GP_avg_trainer'
[]
[]
[Outputs]
csv = true
execute_on = FINAL
[]
(modules/stochastic_tools/test/tests/surrogates/gaussian_process/GP_Matern_half_int_tuned.i)
[StochasticTools]
[]
[Distributions]
[k_dist]
type = Uniform
lower_bound = 1
upper_bound = 10
[]
[q_dist]
type = Uniform
lower_bound = 9000
upper_bound = 11000
[]
[]
[Samplers]
[train_sample]
type = MonteCarlo
num_rows = 10
distributions = 'k_dist q_dist'
execute_on = PRE_MULTIAPP_SETUP
[]
[test_sample]
type = MonteCarlo
num_rows = 100
distributions = 'k_dist q_dist'
execute_on = PRE_MULTIAPP_SETUP
[]
[]
[MultiApps]
[sub]
type = SamplerFullSolveMultiApp
input_files = sub.i
sampler = train_sample
[]
[]
[Controls]
[cmdline]
type = MultiAppCommandLineControl
multi_app = sub
sampler = train_sample
param_names = 'Materials/conductivity/prop_values Kernels/source/value'
[]
[]
[Transfers]
[data]
type = SamplerReporterTransfer
from_multi_app = sub
sampler = train_sample
stochastic_reporter = results
from_reporter = 'avg/value'
[]
[]
[Reporters]
[results]
type = StochasticReporter
parallel_type = ROOT
[]
[samp_avg]
type = EvaluateSurrogate
model = GP_avg
sampler = test_sample
evaluate_std = 'true'
parallel_type = ROOT
execute_on = final
[]
[train_avg]
type = EvaluateSurrogate
model = GP_avg
sampler = train_sample
evaluate_std = 'true'
parallel_type = ROOT
execute_on = final
[]
[]
[VectorPostprocessors]
[hyperparams]
type = GaussianProcessData
gp_name = 'GP_avg'
execute_on = final
[]
[]
[Trainers]
[GP_avg_trainer]
type = GaussianProcessTrainer
execute_on = timestep_end
covariance_function = 'covar' #Choose a Matern with half-integer argument for the kernel
standardize_params = 'true' #Center and scale the training params
standardize_data = 'true' #Center and scale the training data
sampler = train_sample
response = results/data:avg:value
tao_options = '-tao_bncg_type ssml_bfgs'
tune_parameters = ' signal_variance length_factor'
tuning_min = ' 1e-9 1e-9'
tuning_max = ' 1e16 1e16'
tuning_algorithm = 'tao'
[]
[]
[Surrogates]
[GP_avg]
type = GaussianProcess
trainer = GP_avg_trainer
[]
[]
[Covariance]
[covar]
type=MaternHalfIntCovariance
p = 2 #Define the exponential factor
signal_variance = 1 #Use a signal variance of 1 in the kernel
noise_variance = 1e-3 #A small amount of noise can help with numerical stability
length_factor = '0.551133 0.551133' #Select a length factor for each parameter (k and q)
[]
[]
[Outputs]
[out]
type = CSV
execute_on = FINAL
[]
[]
(modules/stochastic_tools/test/tests/surrogates/gaussian_process/GP_Matern_half_int.i)
[StochasticTools]
[]
[Distributions]
[k_dist]
type = Uniform
lower_bound = 1
upper_bound = 10
[]
[q_dist]
type = Uniform
lower_bound = 9000
upper_bound = 11000
[]
[]
[Samplers]
[train_sample]
type = MonteCarlo
num_rows = 10
distributions = 'k_dist q_dist'
execute_on = PRE_MULTIAPP_SETUP
[]
[test_sample]
type = MonteCarlo
num_rows = 100
distributions = 'k_dist q_dist'
execute_on = PRE_MULTIAPP_SETUP
[]
[]
[MultiApps]
[sub]
type = SamplerFullSolveMultiApp
input_files = sub.i
sampler = train_sample
[]
[]
[Controls]
[cmdline]
type = MultiAppCommandLineControl
multi_app = sub
sampler = train_sample
param_names = 'Materials/conductivity/prop_values Kernels/source/value'
[]
[]
[Transfers]
[data]
type = SamplerReporterTransfer
from_multi_app = sub
sampler = train_sample
stochastic_reporter = results
from_reporter = 'avg/value'
[]
[]
[Reporters]
[results]
type = StochasticReporter
parallel_type = ROOT
[]
[samp_avg]
type = EvaluateSurrogate
model = GP_avg
sampler = test_sample
evaluate_std = 'true'
parallel_type = ROOT
execute_on = final
[]
[train_avg]
type = EvaluateSurrogate
model = GP_avg
sampler = train_sample
evaluate_std = 'true'
parallel_type = ROOT
execute_on = final
[]
[]
[VectorPostprocessors]
[hyperparams]
type = GaussianProcessData
gp_name = 'GP_avg'
execute_on = final
[]
[]
[Trainers]
[GP_avg_trainer]
type = GaussianProcessTrainer
execute_on = timestep_end
covariance_function = 'covar' #Choose a Matern with half-integer argument for the kernel
standardize_params = 'true' #Center and scale the training params
standardize_data = 'true' #Center and scale the training data
sampler = train_sample
response = results/data:avg:value
[]
[]
[Surrogates]
[GP_avg]
type = GaussianProcess
trainer = GP_avg_trainer
[]
[]
[Covariance]
[covar]
type=MaternHalfIntCovariance
p = 2 #Define the exponential factor
signal_variance = 1 #Use a signal variance of 1 in the kernel
noise_variance = 1e-6 #A small amount of noise can help with numerical stability
length_factor = '0.551133 0.551133' #Select a length factor for each parameter (k and q)
[]
[]
[Outputs]
[out]
type = CSV
execute_on = FINAL
[]
[]
(modules/stochastic_tools/examples/surrogates/gaussian_process/gaussian_process_uniform_2D_tuned.i)
[StochasticTools]
[]
[Distributions]
[k_dist]
type = Uniform
lower_bound = 1
upper_bound = 10
[]
[q_dist]
type = Uniform
lower_bound = 9000
upper_bound = 11000
[]
[]
[Samplers]
[train_sample]
type = MonteCarlo
num_rows = 50
distributions = 'k_dist q_dist'
execute_on = PRE_MULTIAPP_SETUP
[]
[cart_sample]
type = CartesianProduct
linear_space_items = '1 0.09 100
9000 20 100 '
execute_on = initial
[]
[]
[MultiApps]
[sub]
type = SamplerFullSolveMultiApp
input_files = sub.i
sampler = train_sample
[]
[]
[Controls]
[cmdline]
type = MultiAppCommandLineControl
multi_app = sub
sampler = train_sample
param_names = 'Materials/conductivity/prop_values Kernels/source/value'
[]
[]
[Transfers]
[data]
type = SamplerReporterTransfer
from_multi_app = sub
sampler = train_sample
stochastic_reporter = results
from_reporter = 'avg/value'
[]
[]
[Reporters]
[results]
type = StochasticReporter
[]
[]
[Trainers]
[GP_avg_trainer]
type = GaussianProcessTrainer
execute_on = timestep_end
covariance_function = 'rbf'
standardize_params = 'true' #Center and scale the training params
standardize_data = 'true' #Center and scale the training data
tao_options = '-tao_bncg_type kd'
sampler = train_sample
response = results/data:avg:value
tune_parameters = ' signal_variance length_factor'
tuning_min = ' 1e-9 1e-9'
tuning_max = ' 1e16 1e16'
tuning_algorithm = 'tao'
[]
[]
[Covariance]
[rbf]
type=SquaredExponentialCovariance
signal_variance = 1 #Use a signal variance of 1 in the kernel
noise_variance = 1e-3 #A small amount of noise can help with numerical stability
length_factor = '0.38971 0.38971' #Select a length factor for each parameter (k and q)
[]
[]
[Surrogates]
[GP_avg]
type = GaussianProcess
trainer = 'GP_avg_trainer'
[]
[]
[Reporters]
[train_avg]
type = EvaluateSurrogate
model = GP_avg
sampler = train_sample
evaluate_std = 'true'
parallel_type = ROOT
execute_on = final
[]
[cart_avg]
type = EvaluateSurrogate
model = GP_avg
sampler = cart_sample
evaluate_std = 'true'
parallel_type = ROOT
execute_on = final
[]
[]
[VectorPostprocessors]
[hyperparams]
type = GaussianProcessData
gp_name = 'GP_avg'
execute_on = final
[]
[]
[Outputs]
[out]
type = CSV
execute_on = FINAL
[]
[]