Skip to content
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

Ensemble Kalman Sampler in docs doesn't match the implementation #50

Closed
odunbar opened this issue Sep 6, 2021 · 2 comments · Fixed by #69
Closed

Ensemble Kalman Sampler in docs doesn't match the implementation #50

odunbar opened this issue Sep 6, 2021 · 2 comments · Fixed by #69
Assignees

Comments

@odunbar
Copy link
Collaborator

odunbar commented Sep 6, 2021

We talk about the vanilla EKS implementation, however in implementation we use a linear-corrected version proposed through the paper chain of here, and subsequently here . This is shown to have better numerical and analytical properties than the original proposed EKS.

The authors of the final paper I believe give the correction of the method a new name too. We should take care with how it is presented. Perhaps I can work with @agarbuno with the documentation for this

@odunbar
Copy link
Collaborator Author

odunbar commented Sep 8, 2021

We should also mention the adaptive timestepping used.

@agarbuno
Copy link
Contributor

Yep, so we basically need to incorporate:

  • Wording and references.
  • Time-stepping scheme and references.

bors bot added a commit that referenced this issue Dec 9, 2021
69: EKS docs update. r=agarbuno a=agarbuno

Updates EKS documentation to match the finite-sample correction. It should resolve #50. 

Co-authored-by: agarbuno <[email protected]>
Co-authored-by: Alfredo Garbuno Iñigo <[email protected]>
Co-authored-by: odunbar <[email protected]>
@bors bors bot closed this as completed in a48fe6d Dec 9, 2021
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging a pull request may close this issue.

2 participants