Make basic INLA interface and simple marginalisation routine #533

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.

Sign up for GitHub

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

Jump to bottom

Open

Michal-Novomestsky wants to merge 64 commits into pymc-devs:main from Michal-Novomestsky:implement-pmx.fit-option-for-INLA-+-marginalisation-routine

Contributor

Michal-Novomestsky commented Jul 2, 2025 •

edited

Loading

Addresses #532 and #344.

Relies on pymc-devs/pytensor#1582.

review-notebook-app bot commented Jul 19, 2025

Check out this pull request on

See visual diffs & provide feedback on Jupyter Notebooks.

Powered by ReviewNB

theorashid mentioned this pull request

ENH: Adding INLA to PyMC step 1: get a Laplace Approximation pymc-devs/pymc#6992

Open

Michal-Novomestsky commented

View reviewed changes

tests/test_inla.py Outdated Show resolved Hide resolved

Michal-Novomestsky commented

View reviewed changes

pymc_extras/__init__.py Outdated Show resolved Hide resolved

Michal-Novomestsky commented

View reviewed changes

pymc_extras/inference/__init__.py Outdated Show resolved Hide resolved

Michal-Novomestsky commented

View reviewed changes

pymc_extras/model/marginal/marginal_model.py Outdated Show resolved Hide resolved

Michal-Novomestsky commented

View reviewed changes

pymc_extras/model/marginal/distributions.py Outdated Show resolved Hide resolved

Michal-Novomestsky added 3 commits

August 12, 2025 20:55


          changed laplace approx to return MvNormal

51afb9f


          added seperate line for evaluating Q-hess

c326525


          WIP: minor refactor

61d4d89

ricardoV94 reviewed

View reviewed changes

pymc_extras/model/marginal/distributions.py Outdated Show resolved Hide resolved

pymc_extras/model/marginal/marginal_model.py Outdated Show resolved Hide resolved

Michal-Novomestsky added 15 commits

August 12, 2025 21:01


          started writing fit_INLA routine

1960cb9


          changed minimizer tol to 1e-8

6a1d523


          WIP: MarginalLaplaceRV

674d813


          WIP: Minimize inside logp

3b5d49c


          tidied up MarginalLaplaceRV

22d2ef1


          refactor: variable name change

c49de10


          jesse minimize testing

54e394d


          end-to-end implementation

f02e652


          refactor: changed boolean logic

9fb860e


          refactor: changed distributions in test case

68b87ee


          removed jesse's debug notebook

de2d1fc


          added WIP warning to pmx.fit

787a39e


          refactor: added TODO

71f8642


          refactor: re-ran notebook

18747d5


          refactor: temporarily changed gitignore

c6010f3

Michal-Novomestsky force-pushed the implement-pmx.fit-option-for-INLA-+-marginalisation-routine branch from 398979a to c6010f3 Compare

August 12, 2025 11:05


          refactor: rolled gitignore back to default

a473e87

Michal-Novomestsky force-pushed the implement-pmx.fit-option-for-INLA-+-marginalisation-routine branch from dad163c to a473e87 Compare

August 12, 2025 11:12

Contributor Author

Michal-Novomestsky commented Aug 12, 2025

@maresb, @zaxtax suggested that I reach out to you about the Docs not being built. Do you have any ideas why it's failing? Thanks!

Michal-Novomestsky and others added 8 commits

August 26, 2025 13:54


          added docstring for _precision_mv_normal_logp

0779b6e


          added more documentation

d7b198a


          added example 1 to example notebook


          refactor: default return_latent_posteriors to false

0c4fcd5


          Merge branch 'pymc-devs:main' into implement-pmx.fit-option-for-INLA-…

d031008

…+-marginalisation-routine


          refactor: moved sample step inside if-block

e7ccfe2


          added docstring

ece57b1


          added latex to docstring

a675b37

Michal-Novomestsky commented

View reviewed changes

pymc_extras/inference/INLA/inla.py Outdated Show resolved Hide resolved

Michal-Novomestsky commented

View reviewed changes

pymc_extras/inference/INLA/inla.py Outdated Show resolved Hide resolved

Michal-Novomestsky added 6 commits

August 27, 2025 13:22


          refactored unittest

3636e98


          refactor: moved laplace approx into seperate function + more docstrings

47b8dae


          refactor: TensorLike typehint

fb39764


          refactor: labelling of p(x|y,params)

065c6b2


          refactor: text in example notebook

59b623d


          removed old INLA notebook

6cea8ba

ColtAllen reviewed

View reviewed changes

ColtAllen left a comment

Just docstring suggestions for now.

pymc_extras/inference/INLA/inla.py Outdated Show resolved Hide resolved

pymc_extras/model/marginal/distributions.py Show resolved Hide resolved

pymc_extras/model/marginal/distributions.py Show resolved Hide resolved

pymc_extras/inference/INLA/inla.py Outdated Show resolved Hide resolved

pymc_extras/inference/INLA/inla.py Show resolved Hide resolved

pymc_extras/inference/INLA/inla.py Show resolved Hide resolved


          refactor: local import

609156e

theorashid mentioned this pull request

Issue tracker for PyMC implementation of INLA #340

Open

17 tasks

Michal-Novomestsky added 9 commits

August 28, 2025 10:26


          latex-friendly formatting

bc3f1c3


          getting Q as RV

e032c25


          updated inla docstring

e367958


          added warning (INLA experimental)

8ed64fd


          added AR1 testcase

7ca496b


          added normals to notebook

154cc2c


          refactor: changed test case atol to 0.2

fda71d6


          refactor: add warning to d calculation

04db7c3


          refactor: warning message

934d740

This was referenced Aug 30, 2025

Apple Silicon: error: failed to legalize operation 'mhlo.cholesky' jax-ml/jax#16321

Open

Sparsity support in pytensor pymc-devs/pytensor#1127

Open

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

Reviewers

ricardoV94 ricardoV94 left review comments

jessegrabowski jessegrabowski left review comments

+2 more reviewers

ColtAllen ColtAllen left review comments

theorashid theorashid left review comments

At least 1 approving review is required to merge this pull request.

Labels

None yet