SG-PALM: a Fast Physically Interpretable Tensor Graphical Model
We propose a new graphical model inference procedure, called SG-PALM, for learning conditional dependency structure of high-dimensional tensor-variate data. Unlike most other tensor graphical models the proposed model is interpretable and computationally scalable to high dimension. Physical interpre...
Gorde:
| Egile Nagusiak: | , |
|---|---|
| Formatua: | Artikulua |
| Argitaratua: |
2021
|
| Gaiak: | |
| Sarrera elektronikoa: | https://arxiv.org/abs/2105.12271 http://arxiv.org/abs/2105.12271 |
| Etiketak: |
Etiketarik gabe, Izan zaitez lehena erregistro honi etiketa jartzen!
|
| _version_ | 1879172997881790464 |
|---|---|
| author | Wang, Yu Hero, Alfred |
| author_browse | Hero, Alfred Wang, Yu |
| author_facet | Wang, Yu Hero, Alfred |
| author_sort | Wang, Yu |
| building | arXiv |
| collection | Acceso abierto |
| description | We propose a new graphical model inference procedure, called SG-PALM, for learning conditional dependency structure of high-dimensional tensor-variate data. Unlike most other tensor graphical models the proposed model is interpretable and computationally scalable to high dimension. Physical interpretability follows from the Sylvester generative (SG) model on which SG-PALM is based: the model is exact for any observation process that is a solution of a partial differential equation of Poisson type. Scalability follows from the fast proximal alternating linearized minimization (PALM) procedure that SG-PALM uses during training. We establish that SG-PALM converges linearly (i.e., geometric convergence rate) to a global optimum of its objective function. We demonstrate the scalability and accuracy of SG-PALM for an important but challenging climate prediction problem: spatio-temporal forecasting of solar flares from multimodal imaging data. |
| first_indexed | 2026-10-15T23:50:19Z |
| format | Article |
| fullrecord | <oai_dc:dc xmlns:oai_dc="http://www.openarchives.org/OAI/2.0/oai_dc/" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/oai_dc/ https://www.openarchives.org/OAI/2.0/oai_dc.xsd"><identifier>arxiv-2105.12271</identifier><datestamp>2021-05-27</datestamp>
<dc:title>SG-PALM: a Fast Physically Interpretable Tensor Graphical Model</dc:title>
<dc:creator>Wang, Yu</dc:creator>
<dc:creator>Hero, Alfred</dc:creator>
<dc:subject>Machine Learning</dc:subject>
<dc:subject>Applications</dc:subject>
<dc:subject>Computation</dc:subject>
<dc:description>We propose a new graphical model inference procedure, called SG-PALM, for learning conditional dependency structure of high-dimensional tensor-variate data. Unlike most other tensor graphical models the proposed model is interpretable and computationally scalable to high dimension. Physical interpretability follows from the Sylvester generative (SG) model on which SG-PALM is based: the model is exact for any observation process that is a solution of a partial differential equation of Poisson type. Scalability follows from the fast proximal alternating linearized minimization (PALM) procedure that SG-PALM uses during training. We establish that SG-PALM converges linearly (i.e., geometric convergence rate) to a global optimum of its objective function. We demonstrate the scalability and accuracy of SG-PALM for an important but challenging climate prediction problem: spatio-temporal forecasting of solar flares from multimodal imaging data. </dc:description>
<dc:description>Accepted in ICML 2021</dc:description>
<dc:date>2021-05-26</dc:date>
<dc:type>text</dc:type>
<dc:identifier>https://arxiv.org/abs/2105.12271</dc:identifier>
<dc:identifier>http://arxiv.org/abs/2105.12271</dc:identifier>
</oai_dc:dc> |
| id | arxiv-2105.12271 |
| institution | arXiv |
| last_indexed | 2026-10-15T23:50:19Z |
| publishDate | 2021 |
| publishDateRange | 2021 |
| publishDateSort | 2021 |
| record_format | oai_dc |
| spelling | arxiv-2105.122712021-05-27 SG-PALM: a Fast Physically Interpretable Tensor Graphical Model Wang, Yu Hero, Alfred Machine Learning Applications Computation We propose a new graphical model inference procedure, called SG-PALM, for learning conditional dependency structure of high-dimensional tensor-variate data. Unlike most other tensor graphical models the proposed model is interpretable and computationally scalable to high dimension. Physical interpretability follows from the Sylvester generative (SG) model on which SG-PALM is based: the model is exact for any observation process that is a solution of a partial differential equation of Poisson type. Scalability follows from the fast proximal alternating linearized minimization (PALM) procedure that SG-PALM uses during training. We establish that SG-PALM converges linearly (i.e., geometric convergence rate) to a global optimum of its objective function. We demonstrate the scalability and accuracy of SG-PALM for an important but challenging climate prediction problem: spatio-temporal forecasting of solar flares from multimodal imaging data. Accepted in ICML 2021 2021-05-26 text https://arxiv.org/abs/2105.12271 http://arxiv.org/abs/2105.12271 |
| spellingShingle | Machine Learning Applications Computation Wang, Yu Hero, Alfred SG-PALM: a Fast Physically Interpretable Tensor Graphical Model |
| title | SG-PALM: a Fast Physically Interpretable Tensor Graphical Model |
| title_full | SG-PALM: a Fast Physically Interpretable Tensor Graphical Model |
| title_fullStr | SG-PALM: a Fast Physically Interpretable Tensor Graphical Model |
| title_full_unstemmed | SG-PALM: a Fast Physically Interpretable Tensor Graphical Model |
| title_short | SG-PALM: a Fast Physically Interpretable Tensor Graphical Model |
| title_sort | sg palm a fast physically interpretable tensor graphical model |
| topic | Machine Learning Applications Computation |
| url | https://arxiv.org/abs/2105.12271 http://arxiv.org/abs/2105.12271 |
| work_keys_str_mv | AT wangyu sgpalmafastphysicallyinterpretabletensorgraphicalmodel AT heroalfred sgpalmafastphysicallyinterpretabletensorgraphicalmodel |