BayesianActiveLearningSampler

Fast Bayesian inference with the parallel active learning (partly inspired from El Gammal et al. 2023).

Description

The BayesianActiveLearningSampler derives off of PMCMCBase and is intended for parallel active learning for Bayesian inverse UQ problems. It's function is similar to that of GenericActiveLearningSampler. Differently, BayesianActiveLearningSampler considers the experimental configurations and combines them with the proposed samples in each iteration of active learning.

Input Parameters

file_nameName of the CSV file with configuration values.
C++ Type:FileName
Controllable:No
Description:Name of the CSV file with configuration values.
initial_valuesThe starting values of the inputs to be calibrated.
C++ Type:std::vector<double>
Unit:(no unit assumed)
Controllable:No
Description:The starting values of the inputs to be calibrated.
num_parallel_proposalsNumber of proposals to make and corresponding subApps executed in parallel.
C++ Type:unsigned int
Controllable:No
Description:Number of proposals to make and corresponding subApps executed in parallel.
num_triesNumber of samples to propose in each iteration (not all are sent for subApp evals).
C++ Type:unsigned int
Range:num_tries>0
Controllable:No
Description:Number of samples to propose in each iteration (not all are sent for subApp evals).
prior_distributionsThe prior distributions of the parameters to be calibrated.
C++ Type:std::vector<DistributionName>
Controllable:No
Description:The prior distributions of the parameters to be calibrated.
sorted_indicesThe sorted sample indices in order of importance to evaluate the subApp.
C++ Type:ReporterName
Controllable:No
Description:The sorted sample indices in order of importance to evaluate the subApp.

Required Parameters

execute_onINITIALThe list of flag(s) indicating when this object should be executed. For a description of each flag, see https://mooseframework.inl.gov/source/interfaces/SetupInterface.html.
Default:INITIAL
C++ Type:ExecFlagEnum
Options:XFEM_MARK, FORWARD, ADJOINT, HOMOGENEOUS_FORWARD, ADJOINT_TIMESTEP_BEGIN, ADJOINT_TIMESTEP_END, NONE, INITIAL, LINEAR, LINEAR_CONVERGENCE, NONLINEAR, NONLINEAR_CONVERGENCE, POSTCHECK, TIMESTEP_END, TIMESTEP_BEGIN, MULTIAPP_FIXED_POINT_END, MULTIAPP_FIXED_POINT_BEGIN, MULTIAPP_FIXED_POINT_CONVERGENCE, FINAL, CUSTOM, PRE_MULTIAPP_SETUP
Controllable:No
Description:The list of flag(s) indicating when this object should be executed. For a description of each flag, see https://mooseframework.inl.gov/source/interfaces/SetupInterface.html.
file_column_nameName of column in CSV file to use, by default first column is used.
C++ Type:std::string
Controllable:No
Description:Name of column in CSV file to use, by default first column is used.
limit_get_global_samples429496729The maximum allowed number of items in the DenseMatrix returned by getGlobalSamples method.
Default:429496729
C++ Type:unsigned long
Controllable:No
Description:The maximum allowed number of items in the DenseMatrix returned by getGlobalSamples method.
limit_get_local_samples429496729The maximum allowed number of items in the DenseMatrix returned by getLocalSamples method.
Default:429496729
C++ Type:unsigned long
Controllable:No
Description:The maximum allowed number of items in the DenseMatrix returned by getLocalSamples method.
limit_get_next_local_row429496729The maximum allowed number of items in the std::vector returned by getNextLocalRow method.
Default:429496729
C++ Type:unsigned long
Controllable:No
Description:The maximum allowed number of items in the std::vector returned by getNextLocalRow method.
lower_boundLower bounds for making the next proposal.
C++ Type:std::vector<double>
Unit:(no unit assumed)
Controllable:No
Description:Lower bounds for making the next proposal.
max_procs_per_row4294967295This will ensure that the sampler is partitioned properly when 'MultiApp/*/max_procs_per_app' is specified. It is not recommended to use otherwise.
Default:4294967295
C++ Type:unsigned int
Controllable:No
Description:This will ensure that the sampler is partitioned properly when 'MultiApp/*/max_procs_per_app' is specified. It is not recommended to use otherwise.
min_procs_per_row1This will ensure that the sampler is partitioned properly when 'MultiApp/*/min_procs_per_app' is specified. It is not recommended to use otherwise.
Default:1
C++ Type:unsigned int
Controllable:No
Description:This will ensure that the sampler is partitioned properly when 'MultiApp/*/min_procs_per_app' is specified. It is not recommended to use otherwise.
num_columnsNumber of columns to be used in the CSV file with the configuration values.
C++ Type:unsigned int
Controllable:No
Description:Number of columns to be used in the CSV file with the configuration values.
num_random_seeds100000Initialize a certain number of random seeds. Change from the default only if you have to.
Default:100000
C++ Type:unsigned int
Controllable:No
Description:Initialize a certain number of random seeds. Change from the default only if you have to.
prior_varianceThe prior distribution of the variance parameter to be calibrated.
C++ Type:DistributionName
Controllable:No
Description:The prior distribution of the variance parameter to be calibrated.
seed0Random number generator initial seed
Default:0
C++ Type:unsigned int
Controllable:No
Description:Random number generator initial seed
upper_boundUpper bounds for making the next proposal.
C++ Type:std::vector<double>
Unit:(no unit assumed)
Controllable:No
Description:Upper bounds for making the next proposal.
variance_bound1.79769e+308Upper bound for variance for making the next proposal.
Default:1.79769e+308
C++ Type:double
Unit:(no unit assumed)
Controllable:No
Description:Upper bound for variance for making the next proposal.

Optional 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/samplers/ParallelActiveLearning/BayesianALTrain.i)

(modules/stochastic_tools/test/tests/samplers/ParallelActiveLearning/BayesianALTrain.i)

[StochasticTools]
[]

[Distributions]
  [left]
    type = Normal
    mean = 0.0
    standard_deviation = 1.0
  []
  [right]
    type = Normal
    mean = 0.0
    standard_deviation = 1.0
  []
[]

[Likelihood]
  [gaussian]
    type = Gaussian
    noise = 'noise_specified/noise_specified'
    file_name = 'exp_0_05.csv'
    # log_likelihood = true
  []
[]

[ParallelAcquisition]
  [BayesianPosterior]
    type = BayesianPosteriorTargeted
  []
[]

[Samplers]
  [sample]
    type = BayesianActiveLearningSampler
    prior_distributions = 'left right'
    sorted_indices = 'conditional/sorted_indices'
    num_parallel_proposals = 5
    num_tries = 100
    seed = 200
    num_random_seeds = 100
    file_name = 'confg.csv'
    initial_values = '0.1 0.1'
    execute_on = 'PRE_MULTIAPP_SETUP TIMESTEP_END'
  []
[]

[MultiApps]
  [sub]
    type = SamplerFullSolveMultiApp
    input_files = sub.i
    sampler = sample
  []
[]

[Transfers]
  [reporter_transfer]
    type = SamplerReporterTransfer
    from_reporter = 'average/value'
    stochastic_reporter = 'constant'
    from_multi_app = sub
    sampler = sample
  []
[]

[Controls]
  [cmdline]
    type = MultiAppSamplerControl
    multi_app = sub
    sampler = sample
    param_names = 'left_bc right_bc mesh1'
  []
[]

[Reporters]
  [constant]
    type = StochasticReporter
  []
  [noise_specified]
    type = ConstantReporter
    real_names = 'noise_specified'
    real_values = '0.05'
  []
  [conditional]
    type = BayesianActiveLearner
    output_value = constant/reporter_transfer:average:value
    sampler = sample
    al_gp = GP_al_trainer
    gp_evaluator = GP_eval
    acquisition = 'BayesianPosterior'
    likelihoods = 'gaussian'
    execute_on = timestep_begin
  []
[]

[Trainers]
  [GP_al_trainer]
    type = ActiveLearningGaussianProcess
    covariance_function = 'covar'
    standardize_params = 'true'
    standardize_data = 'true'
    tune_parameters = 'covar:signal_variance covar:length_factor'
    num_iters = 500
    learning_rate = 0.01
    # show_every_nth_iteration = 100
    batch_size = 350
  []
[]

[Surrogates]
  [GP_eval]
    type = GaussianProcessSurrogate
    trainer = GP_al_trainer
  []
[]

[Covariance]
  [covar]
    type = SquaredExponentialCovariance
    signal_variance = 4.0
    noise_variance = 1e-6
    length_factor = '4.0 4.0'
  []
[]

[Executioner]
  type = Transient
  num_steps = 5
[]

[Outputs]
  file_base = 'b_al'
  [out1_parallelAL]
    type = JSON
    execute_system_information_on = NONE
  []
  [out2_BayesianAL]
    type = SurrogateTrainerOutput
    trainers = 'GP_al_trainer'
    execute_on = FINAL
  []
[]

Description
Input Parameters
Input Files