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:20210402T160548Z
LOCATION:Track 3
DTSTART;TZID=America/New_York:20201112T115500
DTEND;TZID=America/New_York:20201112T122500
UID:submissions.supercomputing.org_SC20_sess209_ws_mlhpce103@linklings.com
SUMMARY:A Benders Decomposition Approach to Correlation Clustering
DESCRIPTION:Workshop\n\nA Benders Decomposition Approach to Correlation Cl
 ustering\n\nLukasik, Keuper, Singh, Yarkony\n\nWe tackle the problem of gr
 aph partitioning for image segmentation using correlation clustering (CC),
  which we treat as an integer linear program (ILP). We reformulate optimiz
 ation in the ILP so as to admit efficient optimization via Benders decompo
 sition, a classic technique from operations research. Our Benders decompos
 ition formulation has many subproblems, each associated with a node in the
  CC instance's graph, which can be solved in parallel. Each Benders subpro
 blem enforces the cycle inequalities corresponding to edges with negative 
 (repulsive) weights attached to its corresponding node in the CC instance.
  We generate Magnanti-Wong Benders rows in addition to standard Benders ro
 ws to accelerate optimization. Our Benders decomposition approach provides
  a promising new avenue to accelerate optimization for CC, and, in contras
 t to previous cutting plane approaches, theoretically allows for massive p
 arallelization.\n\nRegistration Category: Workshop Reg Pass
END:VEVENT
END:VCALENDAR