Scalable Stochastic Gradient Riemannian Langevin Dynamics in Non-Diagonal Metrics
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2023
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866916184675319808 |
|---|---|
| author | Yu, Hanlin Hartmann, Marcelo Williams, Bernardo Klami, Arto |
| author_facet | Yu, Hanlin Hartmann, Marcelo Williams, Bernardo Klami, Arto |
| contents | Stochastic-gradient sampling methods are often used to perform Bayesian inference on neural networks. It has been observed that the methods in which notions of differential geometry are included tend to have better performances, with the Riemannian metric improving posterior exploration by accounting for the local curvature. However, the existing methods often resort to simple diagonal metrics to remain computationally efficient. This loses some of the gains. We propose two non-diagonal metrics that can be used in stochastic-gradient samplers to improve convergence and exploration but have only a minor computational overhead over diagonal metrics. We show that for fully connected neural networks (NNs) with sparsity-inducing priors and convolutional NNs with correlated priors, using these metrics can provide improvements. For some other choices the posterior is sufficiently easy also for the simpler metrics. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2303_05101 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | Scalable Stochastic Gradient Riemannian Langevin Dynamics in Non-Diagonal Metrics Yu, Hanlin Hartmann, Marcelo Williams, Bernardo Klami, Arto Machine Learning Computation Stochastic-gradient sampling methods are often used to perform Bayesian inference on neural networks. It has been observed that the methods in which notions of differential geometry are included tend to have better performances, with the Riemannian metric improving posterior exploration by accounting for the local curvature. However, the existing methods often resort to simple diagonal metrics to remain computationally efficient. This loses some of the gains. We propose two non-diagonal metrics that can be used in stochastic-gradient samplers to improve convergence and exploration but have only a minor computational overhead over diagonal metrics. We show that for fully connected neural networks (NNs) with sparsity-inducing priors and convolutional NNs with correlated priors, using these metrics can provide improvements. For some other choices the posterior is sufficiently easy also for the simpler metrics. |
| title | Scalable Stochastic Gradient Riemannian Langevin Dynamics in Non-Diagonal Metrics |
| topic | Machine Learning Computation |
| url | https://arxiv.org/abs/2303.05101 |