search

pytorch / botorch

Bayesian optimization in PyTorch
https://botorch.org/
MIT License
3.11k stars 406 forks source link

Robust Gaussian Processes via Relevance Pursuit #2608

Open SebastianAment opened 3 weeks ago

SebastianAment commented 3 weeks ago

Summary: This commit adds the implementation of the Robust Gaussian Processes via Relevance Pursuit models and algorithms of the NeurIPS 2024 article.

Differential Revision: D65343571

facebook-github-bot commented 3 weeks ago

This pull request was exported from Phabricator. Differential Revision: D65343571

codecov[bot] commented 3 weeks ago

Codecov Report

Attention: Patch coverage is 86.11670% with 69 lines in your changes missing coverage. Please review.

Project coverage is 99.60%. Comparing base (3ca48d0) to head (e6386c6). Report is 14 commits behind head on main.

Files with missing lines Patch % Lines
botorch/models/relevance_pursuit.py 78.38% 67 Missing :warning:
botorch/models/likelihoods/sparse_outlier_noise.py 98.43% 2 Missing :warning:
Additional details and impacted files ```diff @@ Coverage Diff @@ ## main #2608 +/- ## ========================================== - Coverage 99.98% 99.60% -0.39% ========================================== Files 196 198 +2 Lines 17372 17866 +494 ========================================== + Hits 17370 17795 +425 - Misses 2 71 +69 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

🚨 Try these New Features:

facebook-github-bot commented 3 weeks ago

This pull request was exported from Phabricator. Differential Revision: D65343571

facebook-github-bot commented 2 weeks ago

This pull request was exported from Phabricator. Differential Revision: D65343571

facebook-github-bot commented 2 weeks ago

This pull request was exported from Phabricator. Differential Revision: D65343571

facebook-github-bot commented 2 weeks ago

This pull request was exported from Phabricator. Differential Revision: D65343571

jduerholt commented 2 weeks ago

I just read the paper, and really like it! We will definitely integrate it into our workflows, as this problem is very common for our experimental data. Currently we are sometimes using iterative trimming.

Any plan when this will land in main?

SebastianAment commented 2 weeks ago

Great to hear @jduerholt! I'll just need to get test coverage to 100%, will try to get the time for this within the next two weeks. Would be curious to learn about your experience if / when you start using the model.

jduerholt commented 2 weeks ago

I will update you, as soon as it will land here in main, I will integrate it into our workflows ;) And then update you on our experience.