BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//132.216.98.100//NONSGML kigkonsult.se iCalcreator 2.20.4//
BEGIN:VEVENT
UID:20260405T101250EDT-2093LhCVN4@132.216.98.100
DTSTAMP:20260405T141250Z
DESCRIPTION:Title:\n\nRobust control perspectives on algorithm analysis and
  design\n\nAbstract: \n\nMost complicated optimization problems\, in parti
 cular those involving a large number of variables\, are solved in practice
  using iterative algorithms. The problem of selecting a suitable algorithm
  is currently more of an art than a science\; a great deal of expertise is
  required to know which algorithms to try and how to properly tune them. M
 oreover\, there are seldom performance guarantees. In this talk\, I will s
 how how the problem of algorithm selection can be approached using tools f
 rom robust control theory. By solving simple semidefinite programs (that d
 o not scale with problem size)\, we can derive robust bounds on convergenc
 e rates for popular algorithms such as the gradient method\, proximal meth
 ods\, fast/accelerated methods\, and operator-splitting methods such as AD
 MM. The bounds derived in this manner either match or improve upon the bes
 t known bounds from the literature. The bounds also lead to a natural ener
 gy dissipation interpretation and an associated Lyapunov function. Finally
 \, our framework can be used to efficiently search for algorithms that mee
 t desired performance specifications\, thus establishing a principled meth
 odology for designing new algorithms. We give examples of novel algorithm 
 designs to address distributed optimization and stochastic optimization pr
 oblems.\n
DTSTART:20230306T213000Z
DTEND:20230306T223000Z
LOCATION:Room 1104\, Burnside Hall\, CA\, QC\, Montreal\, H3A 0B9\, 805 rue
  Sherbrooke Ouest
SUMMARY:Laurent Lessard (Northeastern University)
URL:https://www.mcgill.ca/mathstat/channels/event/laurent-lessard-northeast
 ern-university-346389
END:VEVENT
END:VCALENDAR