Isotropic SGD: a Practical Approach to Bayesian Posterior Sampling - Archive ouverte HAL Accéder directement au contenu
Pré-Publication, Document De Travail Année : 2020

Isotropic SGD: a Practical Approach to Bayesian Posterior Sampling

Résumé

In this work we define a unified mathematical framework to deepen our understanding of the role of stochastic gradient (SG) noise on the behavior of Markov chain Monte Carlo sampling (SGMCMC) algorithms. Our formulation unlocks the design of a novel, practical approach to posterior sampling, which makes the SG noise isotropic using a fixed learning rate that we determine analytically, and that requires weaker assumptions than existing algorithms. In contrast, the common traits of existing \sgmcmc algorithms is to approximate the isotropy condition either by drowning the gradients in additive noise (annealing the learning rate) or by making restrictive assumptions on the \sg noise covariance and the geometry of the loss landscape. Extensive experimental validations indicate that our proposal is competitive with the state-of-the-art on \sgmcmc, while being much more practical to use.

Dates et versions

hal-03344953 , version 1 (15-09-2021)

Identifiants

Citer

Giulio Franzese, Rosa Candela, Dimitrios Milios, Maurizio Filippone, Pietro Michiardi. Isotropic SGD: a Practical Approach to Bayesian Posterior Sampling. 2020. ⟨hal-03344953⟩
26 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More