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:20260522T150119Z
LOCATION:C141/143/149
DTSTART;TZID=America/Chicago:20181113T160000
DTEND;TZID=America/Chicago:20181113T163000
UID:submissions.supercomputing.org_SC18_sess212_pap273@linklings.com
SUMMARY:Distributed Memory Sparse Inverse Covariance Matrix Estimation on 
 High-Performance Computing Architectures
DESCRIPTION:Aryan Eftekhari (University of Lugano), Matthias Bollhöfer (Br
 aunschweig University of Technology), and Olaf Schenk (University of Lugan
 o)\n\nWe consider the problem of estimating sparse inverse covariance matr
 ices for high-dimensional datasets using the l1-regularized Gaussian maxim
 um likelihood method. This task is particularly challenging as the require
 d computational resources increase superlinearly with the dimensionality o
 f the dataset. We introduce a performant and scalable algorithm which buil
 ds on the current advancements of second-order, maximum likelihood methods
 . The routine leverages the intrinsic parallelism in the linear algebra op
 erations and exploits the underlying sparsity of the problem. The computat
 ional bottlenecks are identified and the respective subroutines are parall
 elized using an MPI-OpenMP approach. Experiments conducted on a Cray XC50 
 system at the Swiss National Supercomputing Center show that, in compariso
 n to the state-of-the-art algorithms, the proposed routine provides signif
 icant strong scaling speedup with ideal scalability up to 128 nodes. The d
 eveloped framework is used to estimate the sparse inverse covariance matri
 x of both synthetic and real-world datasets with up to 10 million dimensio
 ns.\n\nTag: Algorithms, Graph Algorithms, Linear Algebra, Machine Learning
 , Sparse Computation\n\nRegistration Category: Tech Program Reg Pass\n\nFi
 nalist: BSP Finalist\n\nSession Chair: Howie Huang (George Washington Univ
 ersity)\n\n
END:VEVENT
END:VCALENDAR