BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Gaussian process optimization with simulation failures
DTSTART:20190227T110000
DTEND:20190227T120000
DTSTAMP:20260406T171707Z
UID:2c9e16262a3ecf36669d5e88eb04ce03c24cb747ae41eac500e42592
CATEGORIES:Conferences - Seminars
DESCRIPTION:Dr. François Bachoc\nWe address the optimization of a compute
 r model\, where each simulation either fails or returns a valid output per
 formance. We suggest a joint Gaussian process model for classification of 
 the inputs (computation failure or success) and for regression of the perf
 ormance function. We discuss the maximum likelihood estimation of the cova
 riance parameters\, with a stochastic approximation of the gradient. We th
 en extend the celebrated expected improvement criterion to our setting of 
 joint classification and regression\, thus obtaining a global optimization
  algorithm. We prove the convergence of this algorithm. We also study its 
 practical performances\, on simulated data\, and on a real computer model 
 in the context of automotive fan design.
LOCATION:Martigny\, room 106 https://www.google.com/maps/d/viewer?mid=1Wi7
 CoK1Apqmnxt-dLMI8FpRgE0s&msa=0&ll=46.108789091068445%2C7.082373000000075&s
 pn=0.006337%2C0.018915&z=16
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR