BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Europe/Stockholm
X-LIC-LOCATION:Europe/Stockholm
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260421T090513Z
LOCATION:Bldg. 8 - Room B 101
DTSTART;TZID=Europe/Stockholm:20260629T143000
DTEND;TZID=Europe/Stockholm:20260629T153000
UID:submissions.pasc-conference.org_PASC26_sess110_msa237@linklings.com
SUMMARY:Hybrid Imitation–Reinforcement Learning for Stroke Rehabilitation:
  Toward Adaptive and Human-Compatible VR Therapy
DESCRIPTION:Mahdiyeh Moosavi (LISPEN Arts et Métiers Institute of Technolo
 gy Chalon-Sur-Saône, France)\n\nStroke rehabilitation is not just about re
 peating movements, it is about helping patients gradually regain control, 
 confidence, and natural motor behavior. While virtual reality has opened n
 ew possibilities for immersive therapy, current systems still struggle to 
 adapt to individual patients and often rely on rigid, repetitive training 
 patterns.\nIn this talk, I will present a learning framework that brings u
 s a step closer to more natural and adaptive rehabilitation in VR. The ide
 a is simple: instead of forcing an agent to either copy human movement or 
 learn everything from scratch, we combine both. By leveraging human demons
 trations together with reinforcement learning, we allow the system to lear
 n efficiently while still adapting beyond what it has seen. \nWe implement
 ed this approach in a VR environment where human motion is captured and us
 ed to guide the learning process. What is interesting is not just that the
  agent learns to complete the task, but how it moves. The results show tha
 t the agent develops motion patterns that are smooth, consistent, and surp
 risingly close to human behavior without being explicitly programmed to do
  so.\n\nDomain: Engineering, Life Sciences, Computational Methods and Appl
 ied Mathematics\n\nSession Chairs: John Anderson Garcia Henao (University 
 of Bern, ARTORG Center for Biomedical Engineering Research) and Carlos Bar
 rios Hernandez (Universidad Industrial de Santander, LIG/INRIA - CITI Labo
 ratory)\n\n
END:VEVENT
END:VCALENDAR