Server favicon of sorts

[ValwareIRC]

Perhaps a nice idea for a server to provide a favicon for the client to display somewhere appropriate like the server window icon or cough like discord down the left hand side =]]] cough

[mathurinm]

Welcome back @tomaszkacprzak ! $\epsilon$ is a vector, so I'm not sure how $1[\epsilon > 0]$ is to be interpreted, when multiplying $\Vert \epsilon \Vert^2$?

Is it not something like:

$L (\epsilon, \alpha) = 2 · \sum_i ( \alpha + (1 − 2\alpha) · 1[\epsilon_i > 0] ) \epsilon_i^2$

Can you give the implementation a try through a PR, like you did for positive group lasso ? Thanks a lot

[Badr-MOUFAD]

Also I believe it should be $1 / 2$ instead of $2$ to get ride of the coefficient that pops up after differentiation

$L (\epsilon, \alpha) = \frac{1}{2} \sum_i ( \alpha + (1 − 2\alpha) · 1[\epsilon_i > 0] ) \epsilon_i^2$

[tomaszkacprzak]

Hi @mathurinm thank you for a fast reply. Indeed there was a mistake in my equation, the one you gave should be right. I can try a PR. I could start with a Quadratic and modify all the functions. Do you expect any difficulties for optimisation?

[mathurinm]

Yes that sounds like a plan, and no I don't foresee any difficulty!

[mathurinm]

@tomaszkacprzak could you also please let us know in which context you use skglm, and what it brings you compared to alternatives?

[mathurinm]

Hi @tomaszkacprzak, any news on this?

[Aishwarya0811]

Hi @mathurinm i recently gone through your discussion approach is understandable. I can try a PR

[Aishwarya0811]

Hi @mathurinm @tomaszkacprzak!

I've successfully implemented the DoubleQuadratic datafit as discussed.

Implementation highlights:

• Asymmetric quadratic loss: L(ε,α) = (1/2n) * Σᵢ (2α + 2(1-2α) * 1[εᵢ>0]) * εᵢ²
• When α=0.5, gives identical results to standard Quadratic (verified )
• Support for both dense and sparse matrices
• Full compatibility with existing penalties (L1, MCP, etc.)
• Comprehensive test coverage

Verification:

• α=0.5 matches Quadratic exactly (loss and gradient)
• Integration with AndersonCD solver works
• Asymmetric behavior confirmed for α≠0.5
• All tests pass

I'll submit the PR shortly. Thanks for the clear mathematical specification!

[Aishwarya0811]

Hi @mathurinm @Badr-MOUFAD @tomaszkacprzak 👋

Just wanted to follow up on this PR adding the DoubleQuadratic datafit discussed in issue #272.
The implementation is complete, all tests are green , and the behavior matches Quadratic when α=0.5 while supporting asymmetric loss otherwise.

Would you be able to take a look when you get a chance?
Happy to adjust anything if needed. Thanks a lot for your time and review 🙏