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

this looks cool! #5

Closed
kellertuer opened this issue Aug 29, 2024 · 2 comments
Closed

this looks cool! #5

kellertuer opened this issue Aug 29, 2024 · 2 comments

Comments

@kellertuer
Copy link

kellertuer commented Aug 29, 2024

Hi Olivier,
this looks cool!

Should we link to this somewhere from Manifolds maybe (Readme/docs)?

In the long run it would be super great to have something like a ManifoldsDistribution package, see this super old issue at JuliaManifolds/Manifolds.jl#57, just as a remark if you feel that would be something we could start (over on JuliaManifolds).
Similarly if you feel that one of your manifold packages would “live” over there nicely as well, we happily could host them as well; an advantage might be that there is more than one maintainer with rights on the repository – and when you decide, we can surely and would happily of course also add you to the org. I am confident that Mateusz would not have anything against that.

(maybe this is more of a discussion point and not an issue, sorry for that, but the repo does not have discussions activated).

@olivierverdier
Copy link
Owner

Thank you for your comment! I have now activated discussions (I was not aware of that possibility, thanks!).

I was not aware either of the issue you link to. This repo implements various push-forwards of normal distributions. That corresponds to Retracted and Riemannian as special cases, but there are others as well (ActionNoise, for instance).

I have to apologize for "hijacking" the Manifold namespace for some of my projects. I hope you do not mind too much.

Hosting some of these projects in JuliaManifolds is a great idea.

You can definitely link to this project in the Manifolds documentation, that would be a good start, thank you for suggesting that.

@kellertuer
Copy link
Author

Hi,

absolutely no worries, it is not easy to have all the issues and ideas in mind that “float around”; just that the “proper randomness” thing is something we have in there for quite a while; I personally usually just need “some random point on M” to start my optimisation at – therefore rand(M) is sometimes also just some poor-mans-randomness.
Having that properly worked out would be great.
Push forwards of noises is a very nice group of them definelty.

For the highjacking – no worries either! We do not own the name, quite the opposite, it is great to see more packages!
The idea of having that in JuliaManifolds is really to not be dependent on a single person, it is easier (and by now quite common) to organise packages in Julia in organisations for certain topics.
I can check with Mateusz what we could do here – he will also be at NTNU for the autumn semester (coming here mid September).

So – no worries, this was really just a point to check whether you would be interested, and great to hear you are!

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

No branches or pull requests

2 participants