Show simple item record

dc.contributor.advisorSeltzer, Margo I.
dc.contributor.advisorAdams, Ryan Prescott
dc.contributor.authorAngelino, Elaine Lee
dc.date.accessioned2014-10-21T19:07:06Z
dc.date.issued2014-10-21
dc.date.submitted2014
dc.identifier.citationAngelino, Elaine Lee. 2014. Accelerating Markov chain Monte Carlo via parallel predictive prefetching. Doctoral dissertation, Harvard University.en_US
dc.identifier.otherhttp://dissertations.umi.com/gsas.harvard.inactive:11849en
dc.identifier.urihttp://nrs.harvard.edu/urn-3:HUL.InstRepos:13070022
dc.description.abstractWe present a general framework for accelerating a large class of widely used Markov chain Monte Carlo (MCMC) algorithms. This dissertation demonstrates that MCMC inference can be accelerated in a model of parallel computation that uses speculation to predict and complete computational work ahead of when it is known to be useful. By exploiting fast, iterative approximations to the target density, we can speculatively evaluate many potential future steps of the chain in parallel. In Bayesian inference problems, this approach can accelerate sampling from the target distribution, without compromising exactness, by exploiting subsets of data. It takes advantage of whatever parallel resources are available, but produces results exactly equivalent to standard serial execution. In the initial burn-in phase of chain evaluation, it achieves speedup over serial evaluation that is close to linear in the number of available cores.en_US
dc.description.sponsorshipEngineering and Applied Sciencesen_US
dc.language.isoen_USen_US
dash.licenseLAA
dc.subjectComputer scienceen_US
dc.subjectBayesian inferenceen_US
dc.subjectMarkov chain Monte Carloen_US
dc.subjectMCMCen_US
dc.subjectparallelen_US
dc.subjectprefetchingen_US
dc.subjectspeculative executionen_US
dc.titleAccelerating Markov chain Monte Carlo via parallel predictive prefetchingen_US
dc.typeThesis or Dissertationen_US
dash.depositing.authorAngelino, Elaine Lee
dc.date.available2014-10-21T19:07:06Z
thesis.degree.date2014en_US
thesis.degree.disciplineEngineering and Applied Sciencesen_US
thesis.degree.grantorHarvard Universityen_US
thesis.degree.leveldoctoralen_US
thesis.degree.namePh.D.en_US
dc.contributor.committeeMemberSeltzer, Margoen_US
dc.contributor.committeeMemberAdams, Ryanen_US
dc.contributor.committeeMemberKohler, Eddieen_US
dash.contributor.affiliatedAngelino, Elaine Lee


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record