Gaussian KL to N(0, I) (Geron Ch 17, morie.fn grkldg)

KL = -0.5 sum(1 + logvar - mu^2 - exp(logvar)); zero exactly at the prior.

Usage

morie_geron_kl_divergence_gaussian(mu, logvar)

Arguments

mu, logvar

Numeric objects of the same shape.

Value

List with kl, per_dimension, per_sample, variance, n_active_dims.

Examples

V <- c(1, 2, 3, 4, 5, 6, 7, 8)
morie_geron_kl_divergence_gaussian(V, V)
#> $kl
#> [1] 2437.112
#> 
#> $per_dimension
#> [1]    0.8591409    4.1945280   12.5427685   32.7990750   83.7065796
#> [6]  216.2143967  568.8165792 1517.9789935
#> 
#> $per_sample
#> [1] 2437.112
#> 
#> $variance
#> [1]    2.718282    7.389056   20.085537   54.598150  148.413159  403.428793
#> [7] 1096.633158 2980.957987
#> 
#> $n_active_dims
#> [1] 8
#> 
#> $estimate
#> [1] 2437.112
#> 
#> $n
#> [1] 1
#>