BEGIN:VCALENDAR
PRODID:-//eluceo/ical//2.0/EN
VERSION:2.0
CALSCALE:GREGORIAN
BEGIN:VEVENT
UID:73ca76faf825d50bb691cd2f4b69a5a7
DTSTAMP:20260613T032243Z
SUMMARY:Control\, Inference and Learning
DESCRIPTION:Abstract:\nIntelligent systems\, whether natural or artificial\
 , must act in a world that is highly unpredictable. To plan actions with u
 ncertainty is a stochastic optimal control problem. However\, there are tw
 o fundamental problems: the optimal control solution is intractable to com
 pute and intractable to represent due the non-trivial state dependence of 
 the optimal control. This has prevented large scale application of stochas
 tic optimal control theory sofar. The path integral control theory describ
 es a class of control problems whose solution can be computed as an infere
 nce computation. In this presentation we formalize the intuitive notion th
 at the efficiency of the inference computation is related to the proximity
  of the sampling control to the optimal control. Secondly\, we show new re
 sults that allow approximate computation of state dependent optimal contro
 ls using the cross entropy method. These two ingredients together suggest 
 a novel adaptive sampling procedure\, called PICE (path integral cross ent
 ropy method)\, that learns a controller based on self-generated data. |The
  adaptive sampling procedure can be used to efficiently compute optimal co
 ntrols but can also be used to accelerate other Monte Carlo computations. 
 We illustrate the results on a few examples in robotics and time series.\n
 Bio:\nBert Kappen studied particle physics in Groningen\, the Netherlands 
 and completed his PhD in this field in 1987 at the Rockefeller University 
 in New York. From 1987 until 1989 he worked as a scientist at the Philips 
 Research Laboratories in Eindhoven\, the Netherlands. Since 1989\, he is c
 onducting research on neural networks at the laboratory for biophysics of 
 the University of Nijmegen\, the Netherlands. Since 1997 he is associate p
 rofessor and since 2004 full professor at this university.His group consis
 ts of 10 people and is involved in research on Bayesian machine learning\,
  stochastic control theory\, computational neuroscience and several appli
 cations in collaboration with industry. His research was awarded in 1997 
 the prestigious national PIONIER research subsidy. He co-founded in 1998 t
 he company Smart Research\, that sells forensic software for DNA matching 
 (www.bonaparte-dvi.com) and Big4Data (big4data.nl)\, which develops custom
  made machine learning tools. He is director of the Dutch Foundation for N
 eural Networks (SNN)\, which coordinates research on neural networks and m
 achine learning in the Netherlands.  Since 2009\, he is honorary faculty 
 at UCL’s Gatsby Computational Neuroscience Unit in London.
URL:https://www.imperial.ac.uk/events/102900/control-inference-and-learning
 /
DTSTART;TZID=Europe/London:20160128T140000
DTEND;TZID=Europe/London:20160128T150000
LOCATION:United Kingdom
END:VEVENT
BEGIN:VTIMEZONE
TZID:Europe/London
BEGIN:STANDARD
DTSTART:20160128T140000
TZNAME:GMT
TZOFFSETTO:+0000
TZOFFSETFROM:+0000
END:STANDARD
END:VTIMEZONE
END:VCALENDAR