BEGIN:VCALENDAR VERSION:2.0 PRODID:Linklings LLC BEGIN:VTIMEZONE TZID:America/Chicago X-LIC-LOCATION:America/Chicago BEGIN:DAYLIGHT TZOFFSETFROM:-0600 TZOFFSETTO:-0500 TZNAME:CDT DTSTART:19700308T020000 RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU END:DAYLIGHT BEGIN:STANDARD TZOFFSETFROM:-0500 TZOFFSETTO:-0600 TZNAME:CST DTSTART:19701101T020000 RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20211207T055409Z LOCATION:Online DTSTART;TZID=America/Chicago:20211114T090000 DTEND;TZID=America/Chicago:20211114T173000 UID:submissions.supercomputing.org_SC21_sess424@linklings.com SUMMARY:DRBSD-7: The 7th International Workshop on Data Analysis and Reduc tion for Big Scientific Data DESCRIPTION:Workshop\n\nDRBSD-7: Morning Break (10-10:30)\n\n\n\n-------- -------------\nDRBSD-7: Afternoon Break (3-3:30)\n\n\n\n----------------- ----\nLightning Talk: In Situ Anomaly Detection and Reduced Order Surrogat e Models for DNS of Turbulent Combustion\n\nChen\n\nExascale computing wi ll provide a unique opportunity to approach device-scale first principles direct numerical simulation (DNS) and enable access to physics regimes pre viously unattainable. With the advantages of access to “bigger, more compl ex” problems come challenges of data management and requi...\n\n---------- -----------\nLightning Talk: Big Scientific Data Visual Analysis\n\nJohnso n\n\nTime-varying computational field simulations are often prohibitively large and pose challenges for accurate interactive analysis and exploratio n. In this talk, I will present visual analysis research for large-scale time-varying simulations, including a new deep neural network-based partic le tracin...\n\n---------------------\nDRBSD-7: Lunch Break (12:30-2)\n\n \n\n---------------------\nExploring Lossy Compressibility through Statist ical Correlations of Scientific Datasets\n\nKrasowska, Bessac, Underwood, Calhoun, Cappello...\n\nLossy compression plays a growing role in scientif ic simulations where the cost of storing their output data can span teraby tes. Using error bounded lossy compression reduces the amount of storage f or each simulation; however, there is no known bound for the upper limit o n lossy compressibility. Cor...\n\n---------------------\nDRBSD-7 Lightnin g Talk: Introduction\n\nFoster\n\n---------------------\nDRBSD-7: The 7th International Workshop on Data Analysis and Reduction for Big Scientific D ata\n\nKlasky, Liu, Foster, Ainsworth\n\nA growing disparity between simul ation speeds and I/O rates makes it increasingly infeasible for applicatio ns to save all results for analysis. In this new world, applications must increasingly perform online data analysis and reduction, tasks that introd uce algorithmic, implementation and programmi...\n\n---------------------\ nUnderstanding Effectiveness of Multi-Error-Bounded Lossy Compression for Preserving Ranges of Interest in Scientific Analysis\n\nLiu, Di, Zhao, Cha rd, Ding...\n\nMultiple lossy compression frameworks have been proposed to address the vast volumes of data being produced by scientific simulations . Setting different precisions to different ranges of data based on resear chers' interests appears to be a promising approach to further improve the compression ratios...\n\n---------------------\nPyParSVD: A Streaming, Di stributed and Randomized Singular-Value-Decomposition Library\n\nMaulik, M engaldo\n\nWe introduce PyParSVD, a Python library that implements a strea ming, distributed and randomized algorithm for the singular value decompos ition. To demonstrate its effectiveness, we extract coherent structures fr om scientific data. Furthermore, we show weak scaling assessments on up t o 256 nodes of ...\n\n---------------------\nLightning Talk: Prospects for Data Reduction in Climate Modeling\n\nJacob\n\nTwo forces have driving cl imate modeling to its current strategies for saving output. As one of the oldest scientific HPC applications, and a coupled application, climate mo dels early on developed a data reduction strategy centered on time averagi ng. At the same time, the extensive post-run analys...\n\n--------------- ------\nLightning Talk: Data Science and AI by Big Memory Supercomputer\n\ nBoku\n\nOne of the biggest issues for high performance data science and A I is the performance and capacity of storage, both for main memory and I/O storage. For the main memory, traditional DRAM solution limits the capaci ty even though the bandwidth and latency are almost sufficient. About I/O storage, SSD ...\n\n---------------------\nTributaryPCA: Distributed, Stre aming PCA for In Situ Dimension Reduction with Application to Space Weathe r Simulations\n\nWang, Klein, Morley, Jordanova, Henderson...\n\nComputer simulations continue to grow in size and complexity and are moving towards exascale. Simulations at this scale can generate outputs that exceed both storage capacity and the bandwidth available for transferring to storage, making traditional offline statistical inference challenging. Theref...\n \n---------------------\nMitigating Catastrophic Forgetting in Deep Learni ng in a Streaming Setting Using Historical Summary\n\nDash, Yin, Shankar, Wang, Feng\n\nTraining deep learning models incrementally on high-velocity data in a streaming setting can help us discover knowledge in a timely fa shion. However, due to catastrophic forgetting, incrementally trained mode ls increasingly perform poorly on the past data. We propose constructing a nd using a histori...\n\n---------------------\nLightning Talk: Data Scien ce Efforts in the Exascale Computing Project\n\nKothe\n\n----------------- ----\nUnbalanced Parallel I/O: An Often-Neglected Side Effect of Lossy Sci entific Data Compression\n\nWang, Wan, Chen, Gong, Whitney...\n\nLossy com pression techniques have demonstrated promising results in significantly r educing the scientific data size while guaranteeing the compression error bounds. However, one important yet often neglected side effect of lossy sc ientific data compression is its impact on the performance of paralle...\n \n---------------------\nProductive and Performant Generic Lossy Data Comp ression with LibPressio\n\nUnderwood, Malvoso, Calhoun, Di, Cappello\n\nIn recent years, lossless and lossy compressors have been developed to cope with the ever increasing volume of scientific floating point data. Howeve r not all compression techniques are appropriate for all datasets, and det ermining which one to use can be time consuming requiring code modificatio ns...\n\n\nTag: Online Only, Applications, Big Data, Data Analytics, Data Management\n\nRegistration Category: Workshop Reg Pass END:VEVENT END:VCALENDAR