Skip to content

update cholesky in UKI#125

Merged
bors[bot] merged 1 commit intomainfrom
dh/uki_cholesky
Mar 7, 2022
Merged

update cholesky in UKI#125
bors[bot] merged 1 commit intomainfrom
dh/uki_cholesky

Conversation

@Zhengyu-Huang
Copy link
Contributor

@Zhengyu-Huang Zhengyu-Huang commented Mar 4, 2022

Update the Cholesky factor computation in UKI

When the matrix is not positive definite, compute SVD of the matrix, replace 0 singular values as the smallest nonzero singular value (with a threshold 1e-8), and then compute the Cholesky factor.

@Zhengyu-Huang Zhengyu-Huang requested a review from ilopezgp March 4, 2022 23:49
@ilopezgp ilopezgp linked an issue Mar 5, 2022 that may be closed by this pull request
@ilopezgp
Copy link
Contributor

ilopezgp commented Mar 5, 2022

Can you add a unit test of construct_sigma_ensemble passing a positive semidefinite covariance matrix?

Copy link
Contributor

@ilopezgp ilopezgp left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM, thank you @Zhengyu-Huang !

@ilopezgp
Copy link
Contributor

ilopezgp commented Mar 7, 2022

bors r+

@ilopezgp
Copy link
Contributor

ilopezgp commented Mar 7, 2022

Closes #126

@bors
Copy link
Contributor

bors bot commented Mar 7, 2022

Build succeeded:

@bors bors bot merged commit 1a0d195 into main Mar 7, 2022
@bors bors bot deleted the dh/uki_cholesky branch March 7, 2022 22:00
bors bot added a commit that referenced this pull request Mar 7, 2022
127: Patch release v0.5.1, internal changes to Cholesky factorization of UKI. r=ilopezgp a=ilopezgp

This patch release closes issue #126 with PR #125 . 

Co-authored-by: ilopezgp <ilopezgp@gmail.com>
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 this pull request may close these issues.

UKI does not allow for positive semidefinite covariances

3 participants