New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Adaptation for static Bayesian models #17

Closed
LeahPrice opened this Issue Aug 4, 2017 · 2 comments

Comments

Projects
None yet
2 participants
@LeahPrice
Collaborator

LeahPrice commented Aug 4, 2017

Following on from the previous issue ticket (#16), the plan in this pull request is to make some of the more common methods for parameter adaptation for static Bayesian models available. Specifically,

  • Estimating the covariance matrix from the population of particles (e.g. for use in an MCMC proposal)
  • Adapting the 'temperature' schedule in likelihood annealing SMC
  • Adapting the number of MCMC repeats
@adamjohansen

This comment has been minimized.

Show comment
Hide comment
@adamjohansen

adamjohansen Aug 4, 2017

Collaborator

I'm certainly in favour of this, but these do seem to be two distinct things: I'd make sure that generic code for facilitating adaptation is clearly distinguished from the specific example. The former is likely to be reused in other places and we want it to be clear to users that this is the intention.

I can see the logic in separating the examples from the base code changes and personally I'd always err on the side of more, smaller commits/pulls, there's not much cost associated with them and there can be significant advantages.

Collaborator

adamjohansen commented Aug 4, 2017

I'm certainly in favour of this, but these do seem to be two distinct things: I'd make sure that generic code for facilitating adaptation is clearly distinguished from the specific example. The former is likely to be reused in other places and we want it to be clear to users that this is the intention.

I can see the logic in separating the examples from the base code changes and personally I'd always err on the side of more, smaller commits/pulls, there's not much cost associated with them and there can be significant advantages.

@LeahPrice

This comment has been minimized.

Show comment
Hide comment
@LeahPrice

LeahPrice Aug 4, 2017

Collaborator

Okay, I'd be happy to do the commits for these two things separately. I'll edit my original issue ticket and add another.

Collaborator

LeahPrice commented Aug 4, 2017

Okay, I'd be happy to do the commits for these two things separately. I'll edit my original issue ticket and add another.

@LeahPrice LeahPrice changed the title from Examples of adaptation (with linear regression example) to Adaptation for static Bayesian models Aug 4, 2017

@LeahPrice LeahPrice closed this Aug 8, 2017

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment