Liking cljdoc? Tell your friends :D

anglican.bbvb


*epsilon-one*clj/s

source

*epsilon-two*clj/s

source

add-gradient-to-proposalclj/s

(add-gradient-to-proposal q-dist gradient-map weights stepsize use-adagrad)

update proposals via gradient step

update proposals via gradient step
sourceraw docstring

assoc-gradientclj/s

(assoc-gradient state address proposal value)

store in state the gradient value at an address

store in state the gradient value at an address
sourceraw docstring

get-addressclj/s

(get-address smp)

returns a unique identifier for sample checkpoint and the updated state

returns a unique identifier for sample checkpoint
and the updated state
sourceraw docstring

get-or-create-q!clj/s

(get-or-create-q! state address prior)

get the learned proposal at an address; initialize proposal with prior at new addresses.

get the learned proposal at an address;
initialize proposal with prior at new addresses.
sourceraw docstring

get-variationalclj/s

(get-variational state)

return the learned approximating distributions at each address

return the learned approximating distributions at each address
sourceraw docstring

ignore?clj/s

(ignore? state dist)

determine whether to learn an approximation for a given distribution object

determine whether to learn an approximation for a given distribution object
sourceraw docstring

make-initial-stateclj/s

(make-initial-state only exclude)
source

merge-q!clj/s

(merge-q! state proposals)

force specific approximating distributions: used for initialization, or to use as learned proposals within an importance sampler.

force specific approximating distributions: used for initialization, or to
use as learned proposals within an importance sampler.
sourceraw docstring

optimal-scalingclj/s

(optimal-scaling f g)

given two vectors f=wg and g, estimate Cov(f,g)/Var(g). this gives the optimal scaling for the variance reduction term.

given two vectors f=wg and g, estimate Cov(f,g)/Var(g).
this gives the optimal scaling for the variance reduction term.
sourceraw docstring

proposal-gradientclj/s

(proposal-gradient gradient-samples log-weights)

compute the gradient (and appropriate weightings) at each address

compute the gradient (and appropriate weightings) at each address
sourceraw docstring

update-proposals!clj/s

(update-proposals! particles stepsize use-adagrad)

outer loop of gradient update procedure to update proposal dist atom

outer loop of gradient update procedure to update proposal dist atom
sourceraw docstring

cljdoc is a website building & hosting documentation for Clojure/Script libraries

× close