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:20190719T085744Z
LOCATION:HG F 3
DTSTART;TZID=Europe/Stockholm:20190613T121500
DTEND;TZID=Europe/Stockholm:20190613T124500
UID:submissions.pasc-conference.org_PASC19_sess120_msa241@linklings.com
SUMMARY:Machine Learning for Pathological Gait Modelling
DESCRIPTION:Minisymposium\nComputer Science and Applied Mathematics, Emerg
 ing Application Domains, Chemistry and Materials, Climate and Weather, Phy
 sics, Solid Earth Dynamics, Life Sciences, Engineering\n\nMachine Learning
  for Pathological Gait Modelling\n\nKalousis\n\nIn this talk we will prese
 nt how we model pathological gait in order to support treatment selection 
 using machine learning by learning gait models that link clinical data wit
 h the gait's dynamics. Such models would allow medical doctors to explore 
 what-if scenarios for surgery procedure selection. Pathological gait model
 ling is a challenging task. Training data are limited in size. Gait exhibi
 ts high levels of variance, both within a given patient but also between p
 atients. We will show how we learn (conditional) generative models for gai
 t generation and the results we have obtained so far. In addition we will 
 show how we seek to alleviate the small sample size problem by exploiting 
 a neuro-muscular motion simulator. Such a simulator could provide us with 
 infinite amounts of training data; to do so it will have to be tuned to re
 plicate desired pathologies. We will use reinforcement and imitation learn
 ing to learn to tune. The availability of a neuro-muscular simulator that 
 can be tuned at will, will not only allow us to generate a limitless quant
 ity of training data, but more important it will allow us to explore what-
 if treatment scenarios directly on the simulator.
END:VEVENT
END:VCALENDAR