Christian Lorentzen

@luis261 What happened? Why did you close and delete your branch? This PR would be a valuable addition. We are just very short on reviewer time and the little we...

@Micky774 @lmcinnes @jjerphan @glemaitre any chance to get a review?

I‘m not -1 on adding losses to MLP. Just if we do so, we should use the losses that we already have in our private _loss module. They are already...

I‘m -1 on Log-Cosh loss because we have it nowhere else (and because I think it’s not a consistent scoring functions, neither for the expectation nor for the median).

@JosephBARBIERDARNAL Could you please address all the comments from https://github.com/scikit-learn/scikit-learn/pull/28752?

2 things are important for me: - Even if this PR focuses on binary classification, it should be implemented in a way such that (positive) regression can be added easily,...

> We need to make a choice: either we implement support for regressors right away. That is fine for me if we release it together. Maybe merge into a feature...

My main conclusion is that we should make an official distinction of scores/metrics: probabilistic ones and decision metrics.

I don’t intent to sound harsh, still: Could we incorporate Adrin‘s comments, do a bit less magic warnings here for the moment and discuss scorer API redesign elsewhere?

@thomasjpfan ping