-
Notifications
You must be signed in to change notification settings - Fork 3
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
free lunch killed it on this problem #42
Comments
Very cool! I've not read the paper but very interested in the approach. Just from the figures in the article I wonder about if there is sufficient sampling close to the verticies of the simplex? I remember that was something we had trouble getting around. I would guess that a lot of 'optimal' solutiuons to real world problems (portfolios, discrete event probability etc.) are going to be on or near to verticies... but hard to know for sure. Cool to see freelunch doing well. I wonder why SA is performing so well here? |
The boundary is an issue, yes. I'm working on smoothing that out. Question: does freelunch support constaints? I'm reading your docs now... |
By the way I'm thinking of creating a new chatgpt-written objective test suite of objective functions. It is not bad at suggesting them! |
In answer to your question, freelunch is extensible to arbitrary bounding strategies 👍 we offer hypercuboid bounding out of the box, but custom logic can be implemented to overwrite the Interesting that GPT can suggest obj functions... does it write the code or just describe the function? |
It writes the code !
https://github.com/microprediction/humpday/blob/main/humpday/objectives/chatgptobjectives.py
…On Mon, Mar 6, 2023 at 9:32 AM Max Champneys ***@***.***> wrote:
In answer to your question, freelunch is extensible to arbitrary bounding
strategies 👍 we offer hypercuboid bounding out of the box, but custom
logic can be implemented to overwrite the opt.bounder method.
Interesting that GPT can suggest obj functions... does it write the code
or just describe the function?
—
Reply to this email directly, view it on GitHub
<#42 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ANWLINOQFT4CA6QCVUY7EVTW2XYPRANCNFSM6AAAAAAVOFXBTY>
.
You are receiving this because you authored the thread.Message ID:
***@***.***>
|
Check this out:
https://medium.com/@microprediction/a-new-family-of-diffeomorphisms-from-the-simplex-to-the-cube-6d358714f429
The text was updated successfully, but these errors were encountered: