{"615074":{"#nid":"615074","#data":{"type":"news","title":"HiCOO Takes Home Best Student Paper Title of Supercomputing 2018 by Creating a New Storage Format","body":[{"value":"\u003Cp\u003E\u003Cstrong\u003E\u003Ca href=\u0022http:\/\/fruitfly1026.github.io\/\u0022\u003EJiajia Li\u003C\/a\u003E\u0026nbsp;\u003C\/strong\u003Eis a recently graduated Ph.D. student from the\u0026nbsp;\u003Ca href=\u0022https:\/\/www.cse.gatech.edu\/\u0022\u003ESchool of Computational Science and Engineering\u003C\/a\u003E\u0026nbsp;(CSE) with a knack for optimizing tensor algorithms and creating memorable pun-filled paper titles. Li\u0026rsquo;s latest work,\u0026nbsp;\u003Ca href=\u0022http:\/\/fruitfly1026.github.io\/static\/files\/sc18-li.pdf\u0022\u003E\u003Cem\u003EHierarchical Storage of Sparse Tensors\u003C\/em\u003E\u003C\/a\u003E, nicknamed HiCOO, focuses on both of these attributes.\u003C\/p\u003E\r\n\r\n\u003Cp\u003EHiCOO was awarded the prestigious title of best student paper at this year\u0026rsquo;s International Conference for\u0026nbsp;\u003Ca href=\u0022https:\/\/sc18.supercomputing.org\/\u0022\u003EHigh Performance Computing, Networking, Storage, and Analysis\u003C\/a\u003E, commonly referred to as Supercomputing (SC).\u003C\/p\u003E\r\n\r\n\u003Cp\u003EWinning the best student paper award at this well-established and increasingly competitive conference program, which boasted an impressive\u0026nbsp;\u003Ca href=\u0022https:\/\/www.hpcwire.com\/off-the-wire\/sc18-papers-program-is-live-and-awards-finalists-are-posted\/\u0022\u003E288 technical paper submissions with 68 ultimately being accepted\u003C\/a\u003E\u0026nbsp;this year, is no small feat.\u003C\/p\u003E\r\n\r\n\u003Cp\u003EHowever, for Li, who holds a second Ph.D. degree in computer architecture, the journey to effectively alter a process is half the battle and half the fun in research.\u003C\/p\u003E\r\n\r\n\u003Cp\u003E\u0026ldquo;For me, it isn\u0026rsquo;t about the outcome, but rather what I can explore in the process along the way,\u0026rdquo; she said. \u0026ldquo;Tensors algorithms are a way to break down data by organizing or viewing it in a certain manner to find what connects the different factors. Sparse tensors are just a specific kind of tensor.\u0026rdquo;\u003C\/p\u003E\r\n\r\n\u003Cp\u003EFor sparse tensor computations, there is tension among storage, speed, and flexibility. With traditional methods, you can usually get just two of these three traits. For example, a computation can be fast and flexible at the price of more storage. Or, it can be compact and flexible, but also slow.\u003C\/p\u003E\r\n\r\n\u003Cp\u003E\u0026ldquo;We wanted to create a storage format for tensors that was all three: compact, fast, and flexible,\u0026rdquo; said Li.\u003C\/p\u003E\r\n\r\n\u003Cp\u003EAuthors of HiCOO, which includes Li and CSE Associate Professors\u0026nbsp;\u003Cstrong\u003E\u003Ca href=\u0022http:\/\/vuduc.org\/v2\/\u0022\u003ERich Vuduc\u003C\/a\u003E\u0026nbsp;\u003C\/strong\u003Eand\u0026nbsp;\u003Ca href=\u0022http:\/\/sunlab.org\/\u0022\u003E\u003Cstrong\u003EJimeng Sun\u003C\/strong\u003E\u003C\/a\u003E, were inspired by the work of Lawrence Berkeley National Lab Staff Scientist\u0026nbsp;\u003Cstrong\u003E\u003Ca href=\u0022https:\/\/people.eecs.berkeley.edu\/~aydin\/\u0022\u003EAydin Bulu\u0026ccedil;\u003C\/a\u003E\u0026nbsp;\u003C\/strong\u003Ewhich focuses on sparse matrices, the lower dimensional analogue of a tensor.\u0026nbsp;\u003C\/p\u003E\r\n\r\n\u003Cp\u003E\u0026ldquo;We saw an opportunity to extend those ideas to sparse tensors,\u0026rdquo; explained Li. \u0026ldquo;We believe HiCOO is smaller, faster, and simpler to update [when the data is changing] than state-of-the-art alternatives. So, we think it will be easier to use in tensor libraries, tools, and data mining applications, which today include e-commerce, healthcare, security, and deep learning, to name a few.\u0026rdquo;\u003C\/p\u003E\r\n\r\n\u003Cp\u003EVuduc said, \u0026ldquo;An example of this use comes from Jimeng Sun\u0026rsquo;s work where his team is\u0026nbsp;\u003Ca href=\u0022https:\/\/www.cc.gatech.edu\/news\/603063\/deep-learning-can-now-help-prevent-heart-failure\u0022\u003Etrying to find structure in electronic health records\u003C\/a\u003E\u0026nbsp;(EHRs). There is always a way to organize the data along certain dimensions, for example, a patient could be considered one dimension, the diagnosis they receive would be a second, and their treatment a third.\u0026rdquo;\u003C\/p\u003E\r\n\r\n\u003Cp\u003EAs one could imagine, the way in which these dimensions could be combined could yield endless outcomes. Which is why the ability to store data of this size in a way that respects order, quickly computes, and uses as minimal amount of storage as possible is a critically needed function.\u0026nbsp;\u0026nbsp;\u003C\/p\u003E\r\n\r\n\u003Cp\u003E\u0026ldquo;Basically, the formats that existed already could give you two of those areas: All the competition could either have a small data structure which took less storage or if you wanted to access the data in a different order it could but would be slower,\u0026rdquo; said Li. \u0026ldquo;HiCOO breaks the mold of its predecessors by proposing a new storage format for sparse tensors [called Hierarchical COOrdinate], that accomplishes all three.\u0026rdquo;\u003C\/p\u003E\r\n\r\n\u003Cp\u003ELi currently works as a computer scientist at the\u0026nbsp;\u003Ca href=\u0022http:\/\/hpc.pnl.gov\/\u0022\u003EHigh Performance Computing Group\u003C\/a\u003E\u0026nbsp;of\u0026nbsp;\u003Ca href=\u0022https:\/\/www.pnnl.gov\/\u0022\u003EPacific Northwest National Laboratory\u003C\/a\u003E\u0026nbsp;(PNNL) in Richland, Washington.\u0026nbsp;\u003C\/p\u003E\r\n\r\n\u003Cp\u003ERead the PNNL press release of HiCOO\u0026nbsp;\u003Ca href=\u0022https:\/\/www.pnnl.gov\/science\/highlights\/highlight.asp?id=5038\u0022\u003Ehere\u003C\/a\u003E.\u003C\/p\u003E\r\n","summary":null,"format":"limited_html"}],"field_subtitle":"","field_summary":"","field_summary_sentence":[{"value":"CSE recent graduate Jiajia Li\u0027s paper, HiCOO, won best student paper award at Supercomputing 2018."}],"uid":"34540","created_gmt":"2018-12-06 16:56:12","changed_gmt":"2019-08-26 13:51:33","author":"Kristen Perez","boilerplate_text":"","field_publication":"","field_article_url":"","dateline":{"date":"2018-12-06T00:00:00-05:00","iso_date":"2018-12-06T00:00:00-05:00","tz":"America\/New_York"},"extras":[],"hg_media":{"615075":{"id":"615075","type":"image","title":"HiCOO Best Student Paper Announcement","body":null,"created":"1544115789","gmt_created":"2018-12-06 17:03:09","changed":"1544115789","gmt_changed":"2018-12-06 17:03:09","alt":"Supercomputing 2018 Best Student Paper Award Winner:\u00a0HiCOO [Image courtesy of The International Conference for High Performance\u00a0Computing, Networking, Storage, and Analysis]","file":{"fid":"234210","name":"HiCOO Awards - courtesy of SC\u002718.jpg","image_path":"\/sites\/default\/files\/images\/HiCOO%20Awards%20-%20courtesy%20of%20SC%2718.jpg","image_full_path":"http:\/\/tlwarc.hg.gatech.edu\/\/sites\/default\/files\/images\/HiCOO%20Awards%20-%20courtesy%20of%20SC%2718.jpg","mime":"image\/jpeg","size":837210,"path_740":"http:\/\/tlwarc.hg.gatech.edu\/sites\/default\/files\/styles\/740xx_scale\/public\/images\/HiCOO%20Awards%20-%20courtesy%20of%20SC%2718.jpg?itok=GnsdEcek"}}},"media_ids":["615075"],"groups":[{"id":"50877","name":"School of Computational Science and Engineering"},{"id":"47223","name":"College of Computing"},{"id":"624060","name":"Center for High Performance Computing (CHiPC)"}],"categories":[],"keywords":[{"id":"181217","name":"cse-hpc"},{"id":"702","name":"hpc"},{"id":"3427","name":"High performance computing"}],"core_research_areas":[],"news_room_topics":[],"event_categories":[],"invited_audience":[],"affiliations":[],"classification":[],"areas_of_expertise":[],"news_and_recent_appearances":[],"phone":[],"contact":[{"value":"\u003Cp\u003EKristen Perez\u003C\/p\u003E\r\n\r\n\u003Cp\u003ECommunications Officer\u003C\/p\u003E\r\n","format":"limited_html"}],"email":["kristen.perez@cc.gatech.edu"],"slides":[],"orientation":[],"userdata":""}}}