Differentiable PAC–Bayes Objectives with Partially Aggregated Neural Networks

We make two related contributions motivated by the challenge of training stochastic neural networks, particularly in a PAC–Bayesian setting: (1) we show how averaging over an ensemble of stochastic neural networks enables a new class of partially-aggregated estimators, proving that these lead to unb...

Descripción completa

Detalles Bibliográficos
Autores principales: Biggs, Felix, Guedj, Benjamin
Formato: Online Artículo Texto
Lenguaje:English
Publicado: MDPI 2021
Materias:
Article
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8535105/
https://www.ncbi.nlm.nih.gov/pubmed/34682004
http://dx.doi.org/10.3390/e23101280

Internet

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8535105/
https://www.ncbi.nlm.nih.gov/pubmed/34682004
http://dx.doi.org/10.3390/e23101280

Ejemplares similares