Implicit differentiation of Lasso-type models for hyperparameter optimization - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2020

Implicit differentiation of Lasso-type models for hyperparameter optimization

Résumé

Setting regularization parameters for Lasso-type estimators is notoriously difficult, though crucial in practice. The most popular hyperparam-eter optimization approach is grid-search using held-out validation data. Grid-search however requires to choose a predefined grid for each parameter , which scales exponentially in the number of parameters. Another approach is to cast hyperparameter optimization as a bi-level optimization problem, one can solve by gradient descent. The key challenge for these methods is the estimation of the gradient w.r.t. the hyperpa-rameters. Computing this gradient via forward or backward automatic differentiation is possible yet usually suffers from high memory consumption. Alternatively implicit differentiation typically involves solving a linear system which can be prohibitive and numerically unstable in high dimension. In addition, implicit differentiation usually assumes smooth loss functions, which is not the case for Lasso-type problems. This work introduces an efficient implicit differentiation algorithm, without matrix inversion, tailored for Lasso-type problems. Our approach scales to high-dimensional data by leveraging the sparsity of the solutions. Experiments demonstrate that the proposed method outperforms a large number of standard methods to optimize the error on held-out data, or the Stein Unbiased Risk Esti-mator (SURE).
Fichier principal
Vignette du fichier
icml.pdf (650.74 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-02532683 , version 1 (05-04-2020)
hal-02532683 , version 2 (07-09-2020)

Identifiants

  • HAL Id : hal-02532683 , version 2

Citer

Quentin Bertrand, Quentin Klopfenstein, Mathieu Blondel, Samuel Vaiter, Alexandre Gramfort, et al.. Implicit differentiation of Lasso-type models for hyperparameter optimization. ICML 2020 - 37th International Conference on Machine Learning, Jul 2020, Vienna / Virtuel, Austria. ⟨hal-02532683v2⟩
243 Consultations
197 Téléchargements

Partager

Gmail Facebook X LinkedIn More