BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Continuous Neural Association for Efficient Robot Learning
DTSTART:20140109T110000
DTSTAMP:20260408T101525Z
UID:ffa969111a1bfc7175421d91cbfb7e0941d21886f1890da26cabda6e
CATEGORIES:Conferences - Seminars
DESCRIPTION:Prof. Jochen Steil\, Director COR-Lab\, Bielefeld University\n
 Bio: In 1993\, I received the diploma in mathematics from the Bielefeld Un
 iversity\, and joined the Neuroinformatics Group headed by Prof. H. Ritter
 . In 1995/96 I stayed one year at the St. Petersburg Electrotechnical Univ
 ersity\, Russia\, supported by a German Academic Exchange Foundation (DAAD
 ) grant. In 1999\, I received the PhD. Degree (Dr. rer. nat) with a Disser
 tation on "Input-Output Stability of Recurrent Neural Networks". In 2002\,
  I was appointed tenured senior researcher (Akad. Rat) and in 2006 my habi
 litation was accepted by the Faculty of Technology  and I was assigned th
 e venia legendi in Neuroinformatics. From March-July 2006 I was on leave a
 s principal scientist at the Honda Research Institute Europe\, Offenbach. 
 In 2008\, I was appointed apl. Professor for Neuroinformatics at the Facul
 ty of Technology. Since April 2007\, I have been managing director of the 
 Research Institute for Cognition and Robotics\, CoR-Lab and since Nov. 200
 7 a member of the scientific board of the center of excellence in Cognitiv
 e Interaction Technology (CITEC). Also in 2007\, I started the teutolab-ro
 botics with workshops for school pupils to experience robotics for one aft
 ernoon. Since 2010\, I am coordinator of the FP7-EU project AMARSi - Adapt
 ive Modular Architectures for Rich Motor Skills. In the academic year 2012
 -13 I participate in the ZiF research group on "Competition and Priority C
 ontrol in Mind and Brain". From Sept-Oct. 2013 I have been on leave as a t
 emporary professor to Oxford Brookes University.\nHow to represent movemen
 t skills in complex behavioral architectures ? This is a persistant resear
 ch question in cognitive robotics which we tackle through a neuro-robotics
  approach based on the principle of continous neural association. The latt
 er combines ideas from reservoir computing\, input-driven dynamical system
 s and classical associative memories in a coherent framework. The goal is 
 to learn to bind together sensori-motor data\, parametric representations 
 like dynamic movement primitives and  low-dimensional  embeddings of tas
 k-specifying parameters like e.g. via points in a multi-level and multi-sc
 ale skill memory. It is further shown that this paramteric skill memory is
  highly beneficial for application of modern trajectory based learning bas
 ed on roll-outs and reward-weighted averaging\, which we demonstrate in ap
 plications to velocity field learning\, inverse kinematics control\, and s
 kill optimization for humanoid robots like the iCub.
LOCATION:MEB110 http://plan.epfl.ch/?lang=en&room=MEB110
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR