Turing.jl
Turing.jl copied to clipboard
Particle Gibbs does not return log probability
Would it be possible to return the log probability when using PG?
If you have usage questions, please post a minimal working example. These questions are also more suited to the Julia slack (https://julialang.slack.com) channel #turing.
@colehurwitz31 You should be able to extract the log probability from the chain for each sample after you've run it using chain[:lp]
, with chain
being the variable containing the value returned from sample
. Does that work for you?
@mohamed82008 @cpfiffer Sorry for the confusing - I was helping @colehurwitz31 with Turing and found this issue so I asked him to create an issue on this.
Yes PG indeed doesn't report log-joint in chain correct - all :lp
in the chain are 0s.
using Turing
@model gdemo(x, y) = begin
s ~ InverseGamma(10,3)
m ~ Normal(0,sqrt(s))
x ~ Normal(m, sqrt(s))
y ~ Normal(m, sqrt(s))
return s, m
end
x, y = 1.5, 2.0
chn = sample(gdemo(x, y), PG(20, 500))
chn[:lp] # => all 0s
Because PG doesn't track the log-joint, there is no way to extract it directly - is my understanding correct? @yebai
I guess a simple solution is to evaluate the log-joint after each MCMC step, though it adds some more computation (ofc we should have an interface to make it optional). Otherwise we should provide some utility functions to re-evaluate the log-joint for each sample in the MCMC chain.
Because PG doesn't track the log-joint, there is no way to extract it directly - is my understanding correct? @yebai
That's right, the weight associated with each particle is reset to 1
after each resampling step. I think this feature has been brought up in the past, see related issues: https://github.com/TuringLang/Turing.jl/issues/493 https://github.com/TuringLang/Turing.jl/issues/426. It's actually really simple to support this feature using API from https://github.com/TuringLang/Turing.jl/issues/634
logp(model, vi)
However, if we accumulate log weights and store them somewhere, we can avoid an extra call to runmodel
(or logp(model, vi)
using the new API from https://github.com/TuringLang/Turing.jl/issues/634).
However, if we accumulate log weights and store them somewhere, ...
So it seems to me we'd add another field to the ParticleContainer
right?
Likely out of date due to the effort of separating particle MCMC and SMC samplers into https://github.com/TuringLang/AdvancedPS.jl