Titre :

Efficient preconditioned stochastic gradient descent for estimation in latent variable models

Auteur(s) :

Baey, Charlotte [Auteur]
Laboratoire Paul Painlevé - UMR 8524 [LPP]
Delattre, Maud [Auteur]
Mathématiques et Informatique Appliquées du Génome à l'Environnement [Jouy-En-Josas] [MaIAGE]
Kuhn, Estelle [Auteur]
Mathématiques et Informatique Appliquées du Génome à l'Environnement [Jouy-En-Josas] [MaIAGE]
Leger, Jean-Benoist [Auteur]
Heuristique et Diagnostic des Systèmes Complexes [Compiègne] [Heudiasyc]
Lemler, Sarah [Auteur]
Mathématiques et Informatique pour la Complexité et les Systèmes [MICS]

Titre de la manifestation scientifique :

40th International Conference on Machine Learning

Ville :

Honolulu

Pays :

Etats-Unis d'Amérique

Date de début de la manifestation scientifique :

2023-07-23

Titre de l’ouvrage :

Proceedings of the 40th International Conference on Machine Learning, Honolulu, Hawaii, USA. PMLR 202, 2023.

Date de publication :

2023-06

Discipline(s) HAL :

Mathématiques [math]/Statistiques [math.ST]

Résumé en anglais : [en]

Latent variable models are powerful tools for modeling complex phenomena involving in particular partially observed data, unobserved variables or underlying complex unknown structures. Inference is often difficult due to ...
Lire la suite >Latent variable models are powerful tools for modeling complex phenomena involving in particular partially observed data, unobserved variables or underlying complex unknown structures. Inference is often difficult due to the latent structure of the model. To deal with parameter estimation in the presence of latent variables, well-known efficient methods exist, such as gradient-based and EM-type algorithms, but with practical and theoretical limitations. In this paper, we propose as an alternative for parameter estimation an efficient preconditioned stochastic gradient algorithm. Our method includes a preconditioning step based on a positive definite Fisher information matrix estimate. We prove convergence results for the proposed algorithm under mild assumptions for very general latent variables models. We illustrate through relevant simulations the performance of the proposed methodology in a nonlinear mixed effects model and in a stochastic block model.Lire moins >

Langue :

Anglais

Comité de lecture :

Oui

Audience :

Internationale

Vulgarisation :

Non

Collections :

• Laboratoire Paul Painlevé - UMR 8524

Source :

Harvested from HAL