BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/New_York
X-LIC-LOCATION:America/New_York
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20210402T160550Z
LOCATION:Poster Module
DTSTART;TZID=America/New_York:20201119T083000
DTEND;TZID=America/New_York:20201119T170000
UID:submissions.supercomputing.org_SC20_sess341_spostg108@linklings.com
SUMMARY:Sparsity-Aware Distributed Tensor Decomposition
DESCRIPTION:ACM Student Research Competition: Graduate Poster, ACM Student
  Research Competition: Undergraduate Poster, Posters\n\nSparsity-Aware Dis
 tributed Tensor Decomposition\n\nMiao\n\nTensors are used by a wide range 
 of applications as data structures to model multi-dimensional data.  Tenso
 r decomposition is a class of methods for latent data analytics. This work
  presents a sparsity-aware tensor decomposition on a distributed memory sy
 stem. We optimize the CANDECOMP/PARAFAC decomposition, a popular low-rank 
 tensor decomposition used in a wide variety of applications. We first thro
 ughly investigate and gain some insights of a state-of-the-art implementat
 ion, which guide our optimization direction. To solve these problems, we p
 ropose three optimization strategies: predicting the optimal grid configur
 ation, tensor dimensions-oriented data distribution, and overlapping compu
 tation and communication. Our proposed sparsity-aware distributed CANDECOM
 P/PARAFAC decomposition, outperforms the state-of-the-art distributed SPLA
 TT library by up to 2.3 X on 64 CPU nodes.\n\nTag: Student Program\n\nRegi
 stration Category: Tech Program Reg Pass, Exhibits Reg Pass
END:VEVENT
END:VCALENDAR