Multi-task Sparse Structure Learning With Gaussian Copula Models

dc.creatorGoncalves
dc.creatorAndre R.; Von Zuben
dc.creatorFernando J.; Banerjee
dc.creatorArindam
dc.date2016
dc.date2017-11-13T13:22:24Z
dc.date2017-11-13T13:22:24Z
dc.date.accessioned2018-03-29T05:55:10Z
dc.date.available2018-03-29T05:55:10Z
dc.identifierJournal Of Machine Learning Research. Microtome Publ, v. 17, p. , 2016.
dc.identifier1532-4435
dc.identifierWOS:000391480500001
dc.identifierhttp://jmlr.org/papers/v17/15-215.html
dc.identifierhttp://repositorio.unicamp.br/jspui/handle/REPOSIP/327878
dc.identifier.urihttp://repositorioslatinoamericanos.uchile.cl/handle/2250/1364903
dc.descriptionConselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)
dc.descriptionConselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)
dc.descriptionMulti-task learning (MTL) aims to improve generalization performance by learning multiple related tasks simultaneously. While sometimes the underlying task relationship structure is known, often the structure needs to be estimated from data at hand. In this paper, we present a novel family of models for MTL, applicable to regression and classification problems, capable of learning the structure of tasks relationship. In particular, we consider a joint estimation problem of the tasks relationship structure and the individual task parameters, which is solved using alternating minimization. The task relationship revealed by structure learning is founded on recent advances in Gaussian graphical models endowed with sparse estimators of the precision (inverse covariance) matrix. An extension to include flexible Gaussian copula models that relaxes the Gaussian marginal assumption is also proposed. We illustrate the e ff ectiveness of the proposed model on a variety of synthetic and benchmark data sets for regression and classi fi cation. We also consider the problem of combining Earth System Model (ESM) outputs for better projections of future climate, with focus on projections of temperature by combining ESMs in South and North America, and show that the proposed model outperforms several existing methods for the problem.
dc.description17
dc.descriptionNSF [IIS-1029711, IIS-0916750, IIS-0953274, CNS-1314560, IIS-1422557, CCF-1451986, IIS-1447566]
dc.descriptionNASA [NNX12AQ39A]
dc.descriptionIBM
dc.descriptionYahoo
dc.descriptionCNPq
dc.descriptionCNPq, Brazil
dc.descriptionConselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)
dc.descriptionConselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)
dc.languageEnglish
dc.publisherMicrotome Publ
dc.publisherBrookline
dc.relationJournal of Machine Learning Research
dc.rightsaberto
dc.sourceWOS
dc.subjectMulti-task Learning
dc.subjectStructure Learning
dc.subjectGaussian Copula
dc.subjectProbabilistic Graphical Model
dc.subjectSparse Modeling
dc.titleMulti-task Sparse Structure Learning With Gaussian Copula Models
dc.typeArtículos de revistas


Este ítem pertenece a la siguiente institución