{"315831":{"#nid":"315831","#data":{"type":"event","title":"Ph.D. Thesis Defense by Bo Xiao","body":[{"value":"\u003Cp\u003EPh.D. Thesis Defense Announcement\u003C\/p\u003E\u003Cp\u003ETitle: \u003Cstrong\u003EParallel Algorithms for Generalized N-Body Problems in High Dimensions and Their Applications for Bayesian Inference and Image Analysis\u003C\/strong\u003E\u003C\/p\u003E\u003Cp\u003E\u003Cstrong\u003EBo Xiao\u003C\/strong\u003E\u003Cbr \/\u003EPh.D. Candidate\u003Cbr \/\u003ESchool of Computational Science and Engineering\u003Cbr \/\u003ECollege of Computing\u003Cbr \/\u003EGeorgia Institute of Technology\u003C\/p\u003E\u003Cp\u003EDate: Aug 18th, Monday\u003Cbr \/\u003ETime: 1:30pm-3:30pm\u003Cbr \/\u003ELocation: Klaus 1212\u003C\/p\u003E\u003Cp\u003E\u003Cstrong\u003ECommittee:\u003C\/strong\u003E\u003Cbr \/\u003EDr. Richard Vuduc, School of Computational Science and Engineering, Georgia Tech\u003Cbr \/\u003EDr. George Biros, Institute of Computational Engineering and Sciences, UT Austin (advisor)\u003Cbr \/\u003EDr. Hongyuan Zha, School of Computational Science and Engineering, Georgia Tech\u003Cbr \/\u003EDr. David Bader, School of Computational Science and Engineering, Georgia Tech\u003Cbr \/\u003EDr. Edmond Chow, School of Computational Science and Engineering, Georgia Tech\u003C\/p\u003E\u003Cp\u003E\u003Cstrong\u003EAbstract:\u003C\/strong\u003E\u003C\/p\u003E\u003Cp\u003EIn this dissertation, I explore parallel algorithms for general N-Body problems \u003Cbr \/\u003Ein high dimensions, and their applications in machine learning and image analysis \u003Cbr \/\u003Eon distributed infrastructures.\u003C\/p\u003E\u003Cp\u003EIn the first part of this work, we proposed and developed a set of basic tools built \u003Cbr \/\u003Eon top of Message Passing Interface and OpenMP for massively parallel nearest neighbors \u003Cbr \/\u003Esearch. In particular, we present a distributed tree structure to index data in arbitrary \u003Cbr \/\u003Enumber of dimensions, and a novel algorithm that eliminate the need for collective \u003Cbr \/\u003Ecoordinate exchanges during tree construction. To the best of our knowledge, our nearest \u003Cbr \/\u003Eneighbors package is the first attempt that scales to millions of cores in up to a \u003Cbr \/\u003Ethousand dimensions.\u003C\/p\u003E\u003Cp\u003EBased on our nearest neighbors search algorithms, we present \u0022ASKIT\u0022, a parallel fast \u003Cbr \/\u003Ekernel summation tree code with a new near-far field decomposition and a new compact \u003Cbr \/\u003Erepresentation for the far field. Specially our algorithm is kernel independent. \u003Cbr \/\u003EThe efficiency of new near far decomposition depends only on the intrinsic dimensionality \u003Cbr \/\u003Eof data, and the new far field representation only relies on the rand of sub-blocks of \u003Cbr \/\u003Ethe kernel matrix.\u003C\/p\u003E\u003Cp\u003EIn the second part, we developed a Bayesian inference framework and a variational \u003Cbr \/\u003Eformulation for a MAP estimation of the label field for medical image segmentation. \u003Cbr \/\u003EIn particular, we propose new representations for both likelihood probability and \u003Cbr \/\u003Eprior probability functions, as well as their fast calculation. Then a parallel \u003Cbr \/\u003Ematrix free optimization algorithm is given to solve the MAP estimation. Our new \u003Cbr \/\u003Eprior function is suitable for lots of spatial inverse problems. Experimental results \u003Cbr \/\u003Eshow our framework is robust to noise, variations of shapes and artifacts.\u003C\/p\u003E","summary":null,"format":"limited_html"}],"field_subtitle":"","field_summary":"","field_summary_sentence":[{"value":"Parallel Algorithms for Generalized N-Body Problems in High Dimensions and Their Applications for Bayesian Inference and Image Analysis"}],"uid":"28077","created_gmt":"2014-08-15 09:16:19","changed_gmt":"2016-10-08 02:08:43","author":"Danielle Ramirez","boilerplate_text":"","field_publication":"","field_article_url":"","field_event_time":{"event_time_start":"2014-08-18T14:30:00-04:00","event_time_end":"2014-08-18T16:30:00-04:00","event_time_end_last":"2014-08-18T16:30:00-04:00","gmt_time_start":"2014-08-18 18:30:00","gmt_time_end":"2014-08-18 20:30:00","gmt_time_end_last":"2014-08-18 20:30:00","rrule":null,"timezone":"America\/New_York"},"extras":[],"groups":[{"id":"221981","name":"Graduate Studies"}],"categories":[],"keywords":[{"id":"4549","name":"coc"},{"id":"4305","name":"cse"},{"id":"1808","name":"graduate students"},{"id":"5603","name":"thesis"}],"core_research_areas":[],"news_room_topics":[],"event_categories":[{"id":"1788","name":"Other\/Miscellaneous"}],"invited_audience":[{"id":"78771","name":"Public"}],"affiliations":[],"classification":[],"areas_of_expertise":[],"news_and_recent_appearances":[],"phone":[],"contact":[],"email":[],"slides":[],"orientation":[],"userdata":""}}}