Joint control variate for faster black-box variational inference
Fuente:
arXiv
Guardado en:
| Autores principales: | , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2022
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866910357841248256 |
|---|---|
| author | Wang, Xi Geffner, Tomas Domke, Justin |
| author_facet | Wang, Xi Geffner, Tomas Domke, Justin |
| contents | Black-box variational inference performance is sometimes hindered by the use of gradient estimators with high variance. This variance comes from two sources of randomness: Data subsampling and Monte Carlo sampling. While existing control variates only address Monte Carlo noise, and incremental gradient methods typically only address data subsampling, we propose a new "joint" control variate that jointly reduces variance from both sources of noise. This significantly reduces gradient variance, leading to faster optimization in several applications. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2210_07290 |
| institution | arXiv |
| publishDate | 2022 |
| record_format | arxiv |
| spellingShingle | Joint control variate for faster black-box variational inference Wang, Xi Geffner, Tomas Domke, Justin Machine Learning Black-box variational inference performance is sometimes hindered by the use of gradient estimators with high variance. This variance comes from two sources of randomness: Data subsampling and Monte Carlo sampling. While existing control variates only address Monte Carlo noise, and incremental gradient methods typically only address data subsampling, we propose a new "joint" control variate that jointly reduces variance from both sources of noise. This significantly reduces gradient variance, leading to faster optimization in several applications. |
| title | Joint control variate for faster black-box variational inference |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2210.07290 |