pymc-devs/pymc

Compute only dependent logp in step samplers

Open

#7591 aperta il 26 nov 2024

Vedi su GitHub
 (0 commenti) (0 reazioni) (0 assegnatari)Python (1902 fork)batch import
enhancementshelp wantedsamplers

Metriche repository

Star
 (7926 star)
Metriche merge PR
 (Merge medio 11g 6h) (12 PR mergiate in 30 g)

Descrizione

Description

As discussed in: https://discourse.pymc.io/t/semantik-description-of-kruschke-diagrams-model-to-graphviz/16142/8 we could be making more use of the conditional partition of the logp graph for partial step samplers.

When the proposal depends on a ratio between probabilities (BinaryMetropolois, BinaryGibbsMetropolis, CategoricalGibbsMetropolis, Slice, NUTS), we don't need to compute terms that don't depend on the updated variables.

Step samples that explicitly request delta_logp (Metropolis, DEMetropolis) achieve this implicitly via PyTensor rewrites.

I suggest adding a model.dependent_logp(vars) that helps us get the logp we need (the variables of interest + conditional dependent variables)

Guida contributor