Understanding ADVI approx.sample(n)

junpenglao · April 23, 2018, 6:33am

You can find the transformation here:

pymc-devs/pymc3/blob/fcd7ac57048978f144ba8bdd5b90ff2f1b45c51a/pymc3/distributions/dist_math.py#L97


return .5 + .5 * tt.erf(x / tt.sqrt(2.))




def sd2rho(sd):
"""
`sd -> rho` theano converter
:math:`mu + sd*e = mu + log(1+exp(rho))*e`"""
return tt.log(tt.exp(tt.abs_(sd)) - 1.)




def rho2sd(rho):
"""
`rho -> sd` theano converter
:math:`mu + sd*e = mu + log(1+exp(rho))*e`"""
return tt.nnet.softplus(rho)




def log_normal(x, mean, **kwargs):
"""
Calculate logarithm of normal distribution at point `x`
with given `mean` and `std`

Topic		Replies	Views
Sampling a gaussian using ADVI Questions	3	692	April 19, 2018
ADVI : Collecting Values for Specific Variables Questions	5	892	May 1, 2018
Extracting variable means from ADVI fitted model Questions	11	2766	March 27, 2021
Using ADVI's question Questions	5	559	December 3, 2018
ADVI start with initialization Questions	7	2090	September 23, 2017

Understanding ADVI approx.sample(n)

Related topics