Bayesian Lasso and multinomial logistic regression on GPU.
We describe an efficient Bayesian parallel GPU implementation of two classic statistical models-the Lasso and multinomial logistic regression. We focus on parallelizing the key components: matrix multiplication, matrix inversion, and sampling from the full conditionals. Our GPU implementations of Ba...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Public Library of Science (PLoS)
2017-01-01
|
| Series: | PLoS ONE |
| Online Access: | http://europepmc.org/articles/PMC5489220?pdf=render |
| id |
doaj-a577c0a1b58e489bb090d5131dfa228c |
|---|---|
| record_format |
Article |
| spelling |
doaj-a577c0a1b58e489bb090d5131dfa228c2020-11-24T21:50:20ZengPublic Library of Science (PLoS)PLoS ONE1932-62032017-01-01126e018034310.1371/journal.pone.0180343Bayesian Lasso and multinomial logistic regression on GPU.Rok ČešnovarErik ŠtrumbeljWe describe an efficient Bayesian parallel GPU implementation of two classic statistical models-the Lasso and multinomial logistic regression. We focus on parallelizing the key components: matrix multiplication, matrix inversion, and sampling from the full conditionals. Our GPU implementations of Bayesian Lasso and multinomial logistic regression achieve 100-fold speedups on mid-level and high-end GPUs. Substantial speedups of 25 fold can also be achieved on older and lower end GPUs. Samplers are implemented in OpenCL and can be used on any type of GPU and other types of computational units, thereby being convenient and advantageous in practice compared to related work.http://europepmc.org/articles/PMC5489220?pdf=render |
| collection |
DOAJ |
| language |
English |
| format |
Article |
| sources |
DOAJ |
| author |
Rok Češnovar Erik Štrumbelj |
| spellingShingle |
Rok Češnovar Erik Štrumbelj Bayesian Lasso and multinomial logistic regression on GPU. PLoS ONE |
| author_facet |
Rok Češnovar Erik Štrumbelj |
| author_sort |
Rok Češnovar |
| title |
Bayesian Lasso and multinomial logistic regression on GPU. |
| title_short |
Bayesian Lasso and multinomial logistic regression on GPU. |
| title_full |
Bayesian Lasso and multinomial logistic regression on GPU. |
| title_fullStr |
Bayesian Lasso and multinomial logistic regression on GPU. |
| title_full_unstemmed |
Bayesian Lasso and multinomial logistic regression on GPU. |
| title_sort |
bayesian lasso and multinomial logistic regression on gpu. |
| publisher |
Public Library of Science (PLoS) |
| series |
PLoS ONE |
| issn |
1932-6203 |
| publishDate |
2017-01-01 |
| description |
We describe an efficient Bayesian parallel GPU implementation of two classic statistical models-the Lasso and multinomial logistic regression. We focus on parallelizing the key components: matrix multiplication, matrix inversion, and sampling from the full conditionals. Our GPU implementations of Bayesian Lasso and multinomial logistic regression achieve 100-fold speedups on mid-level and high-end GPUs. Substantial speedups of 25 fold can also be achieved on older and lower end GPUs. Samplers are implemented in OpenCL and can be used on any type of GPU and other types of computational units, thereby being convenient and advantageous in practice compared to related work. |
| url |
http://europepmc.org/articles/PMC5489220?pdf=render |
| work_keys_str_mv |
AT rokcesnovar bayesianlassoandmultinomiallogisticregressionongpu AT erikstrumbelj bayesianlassoandmultinomiallogisticregressionongpu |
| _version_ |
1725884778551443456 |