BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Clinical data based optimal STI strategies for HIV: a reinforcemen
 t learning approach - Dr Guy-Bart Stan (University of Cambridge)
DTSTART:20071121T140000Z
DTEND:20071121T150000Z
UID:TALK9176@talks.cam.ac.uk
CONTACT:Carl Edward Rasmussen
DESCRIPTION:This  research  addresses the  problem  of  computing optimal 
 structured treatment interruption strategies (STI) for HIV infected  patie
 nts.\n\nSTI represent a class of treatments in which patients are cycled o
 n  and off  drug therapy at specific time instants. The problem that we co
 nsider consists in designing efficient drug-scheduling strategies\, i.e. s
 trategies which bring the immune system  into a state that allows it to in
 dependently (without help from any drug) maintain immune control over the 
 virus.  Also\, this transfer to a drug-independent viral  control   situat
 ion  should  be  done  with as low as possible drug-related systemic effec
 ts for the patients.\n\nIn this presentation\, we show that reinforcement 
  learning may be useful to  extract (close-to) optimal STI strategies dire
 ctly  from  clinical  data\,   without  the  need of identifying a mathema
 tical model of HIV  infection dynamics.  To support our claims\, we report
   simulation results obtained by running  a recently proposed batch-mode  
 reinforcement  learning   algorithm\,  known  as  fitted Q iteration\, on 
 numerically generated data.\n\nThe corresponding paper can be found at htt
 p://www.montefiore.ulg.ac.be/~stan/CDC_2006.pdf\n
LOCATION:Engineering Department\, CBL Room 438
END:VEVENT
END:VCALENDAR