Skip to content

Conversation

@ppebay
Copy link
Contributor

@ppebay ppebay commented Jul 21, 2025

Resolves #611

@ppebay ppebay self-assigned this Jul 21, 2025
@ppebay ppebay marked this pull request as ready for review July 26, 2025 19:31
@ppebay ppebay requested a review from lifflander July 26, 2025 19:33
@ppebay
Copy link
Contributor Author

ppebay commented Jul 26, 2025

@lifflander this is now fully working, and the optional choice to use the update formulae (or not) is determined by the use of the TemperedWithUpdates criterion, instead of the full-information Tempered one. Eventually, we may simply deprecate the latter and replace it with the former.

Currently, the debug verification is performed, by comparing the update vs. full recompute results; until we convince ourselves fully that the update formulae are correctly implemented (and no theoretical errors were made in the paper either...).

Follow-on issues as created to keep track of these 2 points.

@ppebay ppebay requested a review from pierrepebay July 27, 2025 00:13
@ppebay
Copy link
Contributor Author

ppebay commented Aug 13, 2025

@lifflander I made some additional testing following our recent conversation, no problem found. Still awaiting your review when you get a chance.

Copy link
Contributor

@lifflander lifflander left a comment

Choose a reason for hiding this comment

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

Overall, this looks great. I've walked through all the update formula and I think they are correct.

@ppebay ppebay marked this pull request as draft August 25, 2025 13:35
…y detected with a configuration non even feasible to begin with and therefore not part of the test harness)
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.

Implement all update formulae and make them optionally usable

3 participants