Data Generation in Low Sample Size Setting Using Manifold Sampling and a Geometry-Aware VAE - Archive ouverte HAL Accéder directement au contenu
Pré-Publication, Document De Travail Année : 2021

Data Generation in Low Sample Size Setting Using Manifold Sampling and a Geometry-Aware VAE

Résumé

While much efforts have been focused on improving Variational Autoencoders through richer posterior and prior distributions, little interest was shown in amending the way we generate the data. In this paper, we develop two non prior-dependent generation procedures based on the geometry of the latent space seen as a Riemannian manifold. The first one consists in sampling along geodesic paths which is a natural way to explore the latent space while the second one consists in sampling from the inverse of the metric volume element which is easier to use in practice. Both methods are then compared to prior-based methods on various data sets and appear well suited for a limited data regime. Finally, the latter method is used to perform data augmentation in a small sample size setting and is validated across various standard and real-life data sets. In particular, this scheme allows to greatly improve classification results on the OASIS database where balanced accuracy jumps from 80.7% for a classifier trained with the raw data to 89.1% when trained only with the synthetic data generated by our method. Such results were also observed on 4 standard data sets.
Fichier principal
Vignette du fichier
Low_sample_size_setting_data_generation.pdf (5.24 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03180775 , version 1 (25-03-2021)
hal-03180775 , version 2 (21-10-2021)

Identifiants

  • HAL Id : hal-03180775 , version 1

Citer

Clément Chadebec, Stéphanie Allassonnière. Data Generation in Low Sample Size Setting Using Manifold Sampling and a Geometry-Aware VAE. 2021. ⟨hal-03180775v1⟩
345 Consultations
505 Téléchargements

Partager

Gmail Facebook X LinkedIn More