magni.cs.reconstruction.amp.threshold_operator module¶
Module providing threshold functions for the Approximate Message Passing (AMP) algorithm.
Routine listings¶
- ValidatedThresholdOperator(magni.utils.validation.types.ThresholdOperator)
- A base class for validated
magni.cs.reconstruction.amp
threshold operator - SoftThreshold(ValidatedThresholdOperator)
- A soft threshold operator.
-
class
magni.cs.reconstruction.amp.threshold_operator.
ValidatedThresholdOperator
(var)[source]¶ Bases:
magni.utils.validation.types.ThresholdOperator
A base class for validated
magni.cs.reconstruction.amp
threshold operatorParameters: var (dict) – The threshold operator state variables. -
compute_deriv_threshold
(var)[source]¶ Compute the entrywise derivative threshold.
Parameters: var (dict) – The variables used in computing the derivative threshold. Returns: eta_deriv (ndarray) – The computed entrywise derivative threshold. Notes
This method honors magni.utils.validation.enable_allow_validate_once.
-
-
class
magni.cs.reconstruction.amp.threshold_operator.
SoftThreshold
(var)[source]¶ Bases:
magni.cs.reconstruction.amp.threshold_operator.ValidatedThresholdOperator
A soft threshold operator.
This soft threshold operator is based on the description of it and its use in AMP as given in [1] with corrections from [2].
Parameters: - threshold_level_update_method ({‘residual’, ‘median’}) – The method to use for updating the threshold level.
- theta (float) – The tunable regularisation parameter in the threshold level.
- tau_hat_sq (float) – The mean squared error of the (approximated) un-thresholded estimate used to determine the threshold level.
Notes
The above Parameters are the threshold parameters that must be passed in a var dict to the threshold constructor.
References
[1] A. Montanari, “Graphical models concepts in compressed sensing” in Compressed Sensing: Theory and Applications, Y. C. Eldar and G. Kutyniok (Ed.), Cambridge University Press, ch. 9, pp. 394-438, 2012. [2] J. T. Parker, “Approximate Message Passing Algorithms for Generalized Bilinear Inference”, PhD Thesis, Graduate School of The Ohio State University, 2014 -
compute_deriv_threshold
(var)[source]¶ Compute the entrywise derivative soft threshold.
Parameters: var (dict) – The variables used in computing the derivative threshold. Returns: eta_deriv (ndarray) – The computed entrywise derivative threshold.