Using pm.DensityDist and customized likelihood with a black-box function

junpenglao · August 21, 2018, 3:43pm

No, gradient information is not available using @as_op
Yes, using a Normal likelihood with constant sigma (ie, not a latent variable) is the same as L2 regularization,

Topic		Replies	Views
Back-box function without specifying gradient Questions	0	483	February 23, 2021
Suggested workaround for blackbox likelihood example? Questions	6	765	February 8, 2021
How to calibrate the sigma of the likelihood within the black-box method? Questions	13	1443	June 30, 2019
Defining a numeric (custom) likelihood function in PyMC3 Questions theano	2	3070	August 25, 2018
Unobserved variables in blackbox likelihood Questions	11	1121	October 14, 2019