BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//132.216.98.100//NONSGML kigkonsult.se iCalcreator 2.20.4//
BEGIN:VEVENT
UID:20260705T092808EDT-7935gZd9V4@132.216.98.100
DTSTAMP:20260705T132808Z
DESCRIPTION:Dr. Joelle Pineau\, School of Computer Science\, 91ºÚÁÏÍø.    Ele
 ctrical stimulation has recently emerged as a promising therapy for patien
 ts with medically intractable epilepsy\, but little is known about the bes
 t stimulation patterns to use in order to get maximal seizure reduction wh
 ile minimizing long-term damage to the brain. This project's overall goal 
 is to automatically optimize a closed-loop strategy for the control of dee
 p brain stimulation using reinforcement learning methods. I will present s
 ome of the key methodological challenges that must be addressed in order t
 o automatically learn adaptive treatment strategies in this context. In pa
 rticular\, I'll discuss (1) the use of ensemble methods to automatically d
 etect seizures\, (2) the design of a computational model of epilepsy to pr
 ovide synthetic training data for the reinforcement learning agent\, (3) i
 nitial results of applying reinforcement learning using in vitro recording
 s.    This is joint work with Massimo Avoli\, Aaron Courville\, Arthur Gue
 z\, Philip de Guzman and Robert Vincent.\n
DTSTART:20070919T183000Z
DTEND:20070919T193000Z
LOCATION:Duff Medical Building\, CA\, QC\, Montreal\, H3A 2B4\, 3775 rue Un
 iversity
SUMMARY:Adaptive stimulation design for the treatment of epilepsy
URL:/channels/event/adaptive-stimulation-design-treatm
 ent-epilepsy-26755
END:VEVENT
END:VCALENDAR