File Download
There are no files associated with this item.
Supplementary
-
Citations:
- Scopus: 0
- Appears in Collections:
Conference Paper: Quartz: Randomized dual coordinate ascent with arbitrary sampling
Title | Quartz: Randomized dual coordinate ascent with arbitrary sampling |
---|---|
Other Titles | Randomized dual coordinate ascent with arbitrary sampling |
Authors | |
Keywords | Arbitrary sampling Data-driven speedup Dual coordinate ascent Empirical risk minimization |
Issue Date | 2015 |
Publisher | Morgan Kaufmann Publishers, Inc. The Proceedings' web site is located at https://papers.nips.cc/ |
Citation | The 29th Annual Conference on Neural Information Processing Systems (NIPS 2015), Montreal, Canada, 7-12 December 2015. In Conference Proceedings, 2015, v. 28, p. 865-873 How to Cite? |
Abstract | We study the problem of minimizing the average of a large number of smooth convex functions penalized with a strongly convex regularizer. We propose and analyze a novel primal-dual method (Quartz) which at every iteration samples and updates a random subset of the dual variables, chosen according to an arbitrary distribution. In contrast to typical analysis, we directly bound the decrease of the primal-dual error (in expectation), without the need to first analyze the dual error. Depending on the choice of the sampling, we obtain efficient serial and mini-batch variants of the method. In the serial case, our bounds match the best known bounds for SDCA (both with uniform and importance sampling). With standard mini-batching, our bounds predict initial data-independent speedup as well as additional data-driven speedup which depends on spectral and sparsity properties of the data. |
Description | Free Access of NIPS Proceedings' website located at: http://papers.nips.cc/ |
Persistent Identifier | http://hdl.handle.net/10722/235016 |
ISSN | 2020 SCImago Journal Rankings: 1.399 |
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Qu, Z | - |
dc.contributor.author | Richtarik, P | - |
dc.contributor.author | Zhang, T | - |
dc.date.accessioned | 2016-10-14T13:50:44Z | - |
dc.date.available | 2016-10-14T13:50:44Z | - |
dc.date.issued | 2015 | - |
dc.identifier.citation | The 29th Annual Conference on Neural Information Processing Systems (NIPS 2015), Montreal, Canada, 7-12 December 2015. In Conference Proceedings, 2015, v. 28, p. 865-873 | - |
dc.identifier.issn | 1049-5258 | - |
dc.identifier.uri | http://hdl.handle.net/10722/235016 | - |
dc.description | Free Access of NIPS Proceedings' website located at: http://papers.nips.cc/ | - |
dc.description.abstract | We study the problem of minimizing the average of a large number of smooth convex functions penalized with a strongly convex regularizer. We propose and analyze a novel primal-dual method (Quartz) which at every iteration samples and updates a random subset of the dual variables, chosen according to an arbitrary distribution. In contrast to typical analysis, we directly bound the decrease of the primal-dual error (in expectation), without the need to first analyze the dual error. Depending on the choice of the sampling, we obtain efficient serial and mini-batch variants of the method. In the serial case, our bounds match the best known bounds for SDCA (both with uniform and importance sampling). With standard mini-batching, our bounds predict initial data-independent speedup as well as additional data-driven speedup which depends on spectral and sparsity properties of the data. | - |
dc.language | eng | - |
dc.publisher | Morgan Kaufmann Publishers, Inc. The Proceedings' web site is located at https://papers.nips.cc/ | - |
dc.relation.ispartof | Advances in Neural Information Processing Systems | - |
dc.subject | Arbitrary sampling | - |
dc.subject | Data-driven speedup | - |
dc.subject | Dual coordinate ascent | - |
dc.subject | Empirical risk minimization | - |
dc.title | Quartz: Randomized dual coordinate ascent with arbitrary sampling | - |
dc.title.alternative | Randomized dual coordinate ascent with arbitrary sampling | - |
dc.type | Conference_Paper | - |
dc.identifier.email | Qu, Z: zhengqu@hku.hk | - |
dc.identifier.authority | Qu, Z=rp02096 | - |
dc.identifier.scopus | eid_2-s2.0-84965123044 | - |
dc.identifier.hkuros | 269838 | - |
dc.identifier.volume | 28 | - |
dc.identifier.spage | 865 | - |
dc.identifier.epage | 873 | - |
dc.publisher.place | United States | - |
dc.customcontrol.immutable | sml 161130 | - |
dc.identifier.issnl | 1049-5258 | - |