Can bin-wise scaling improve consistency and adaptivity of prediction uncertainty for machine learning regression ? - Université Paris-Saclay Accéder directement au contenu
Pré-Publication, Document De Travail Année : 2023

Can bin-wise scaling improve consistency and adaptivity of prediction uncertainty for machine learning regression ?

Pascal Pernot

Résumé

Binwise Variance Scaling (BVS) has recently been proposed as a post hoc recalibration method for prediction uncertainties of machine learning regression problems that is able of more efficient corrections than uniform variance (or temperature) scaling. The original version of BVS uses uncertainty-based binning, which is aimed to improve calibration conditionally on uncertainty, i.e. consistency. I explore here several adaptations of BVS, in particular with alternative loss functions and a binning scheme based on an input-feature (X) in order to improve adaptivity, i.e. calibration conditional on X. The performances of BVS and its proposed variants are tested on a benchmark dataset for the prediction of atomization energies and compared to the results of isotonic regression.

Dates et versions

hal-04249284 , version 1 (19-10-2023)

Identifiants

Citer

Pascal Pernot. Can bin-wise scaling improve consistency and adaptivity of prediction uncertainty for machine learning regression ?. 2023. ⟨hal-04249284⟩
10 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More