romtools.workflows.inverse.full_covariance_vi_drivers#
Full-covariance Gaussian VI wrappers.
This module adds a true full-covariance Gaussian variational family without
changing the established diagonal VI and MF-VI implementations. The Gaussian
lives in optimizer coordinates, q(x)=N(mu, Sigma), with Sigma=L L.T.
The Cholesky factor is the persisted numerical state while updates use the
Gaussian Fisher geometry in mean/covariance coordinates.
Functions
|
|
|
|
|
Run VI, adding a true full-covariance Gaussian variational family. |
- romtools.workflows.inverse.full_covariance_vi_drivers.run_vi(model, prior_parameter_space, observations, observations_covariance, parameter_mins=None, parameter_maxes=None, initial_variational_parameter_space=None, restart_file=None, optimizer_method='gradient', optimizer_config=None, line_search_method='stochastic_nonmonotone', line_search_config=None, absolute_work_dir=None, sample_size=30, random_seed=1, sampling_method='mc', evaluation_concurrency=1, covariance_regularization=1e-08, restart_files_to_keep=10, elbo_scaling_factor='diag_mean', elbo_relative_tolerance=None, baseline_method=None, bounded_parameter_handling='transform', transform_interior_margin=1e-06, transform_map='sigmoid', min_physical_variational_std_fraction=1e-06, dispatcher=None, *, absolute_vi_directory=None, score_function_entropy_strategy='analytic', variational_distribution=None, max_covariance_log_step=1.0, create_run_directories=True, sample_reuse_config=None)[source]#
Run VI, adding a true full-covariance Gaussian variational family.
- Parameters:
observations (ndarray)
observations_covariance (ndarray)
parameter_mins (ndarray)
parameter_maxes (ndarray)
restart_file (str)
optimizer_method (str)
line_search_method (str)
absolute_work_dir (str)
sample_size (int)
random_seed (int)
sampling_method (str)
covariance_regularization (float)
restart_files_to_keep (int)
elbo_relative_tolerance (float)
baseline_method (str)
bounded_parameter_handling (str)
transform_interior_margin (float)
transform_map (str)
min_physical_variational_std_fraction (float)
absolute_vi_directory (str)
score_function_entropy_strategy (str)
variational_distribution (str)
max_covariance_log_step (float)
create_run_directories (bool)