search

probcomp / GenExperimental.jl

Featherweight embedded probabilistic programming language and compositional inference programming library
MIT License
17 stars 2 forks source link

Write fork(trace) = deepcopy(trace) or something equivalent #19

Open marcoct opened 7 years ago

vkmvkmvkmvkm commented 7 years ago

We may want Taylor to implement a persistent data structure here, so we have support for copy-on-write trace forking. Does that make sense?

On Tue, Jun 20, 2017 at 10:10 PM, Marco Cusumano-Towner < notifications@github.com> wrote:

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/probcomp/Gen.jl/issues/19, or mute the thread https://github.com/notifications/unsubscribe-auth/ABXzoOaSnn-kdT1479LwqVT9eeTE_pF5ks5sGHuSgaJpZM4OAXyP .

marcoct commented 7 years ago

Yes, that makes sense. I will need to think about this a bit more. In addition to forking traces for use in SIR / SMC, we have proposal programs proposing trace fragments into model programs (https://github.com/probcomp/Gen.jl/issues/1). Persistent trace data structure may also be relevant for that.