BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Three NCCR Robotics Seminars
DTSTART:20160129T131500
DTEND:20160129T153000
DTSTAMP:20260508T010629Z
UID:a362be1b9bbae331ee5894ee21ecc8363aeb856809a05b45288bce25
CATEGORIES:Conferences - Seminars
DESCRIPTION:Three international experts from the world of Robotics\nSpeake
 rs: Prof. Oliver Brock\, TUB\nOliver Brock is the Alexander von Humboldt P
 rofessor of Robotics in the School of Electrical Engineering and Computer 
 Science at the Technische Universität Berlin in Germany. He received his 
 Diploma in Computer Science in 1993 from the Technische Universität Berli
 n and his Master's and Ph.D. in Computer Science from Stanford University 
 in 1994 and 2000\, respectively. He also held post-doctoral positions at R
 ice University and Stanford University. Starting in 2002\, he was an Assis
 tant Professor and Associate Professor in the Department of Computer Scien
 ce at the University of Massachusetts Amherst\, before to moving back to t
 he Technische Universität Berlin in 2009. The research of Brock's lab\, t
 he Robotics and Biology Laboratory\, focuses on autonomous mobile manipula
 tion\, interactive perception\, grasping\, manipulation\, soft hands\, int
 eractive learning\, motion generation\, and the application of algorithms 
 and concepts from robotics to computational problems in structural molecul
 ar biology. He is also the president of the Robotics: Science and Systems 
 foundation.Prof. Tamim Asfourm\, KIT\nTamim Asfour is is full Professor at
  the Institute for Anthropomatics and Robotics\, Karlsruhe Institute of Te
 chnology (KIT) where he holds the chair of Humanoid Robotics Systems and i
 s head of the High Performance Humanoid Technologies Lab (H2T). His curren
 t research interest is high performance 24/7 humanoid robotics. Specifical
 ly\, his research focuses on engineering humanoid robot systems integratin
 g the mechano-informatics of such systems with the capabilities of predict
 ing\, acting and learning from human demonstration and sensorimotor experi
 ence.  He is developer of the ARMAR humanoid robot family and is leader o
 f the Humanoid Research Group at KIT (2003-now).\nHe is Founding Editor-in
 -Chief of the IEEE-RAS Humanoids Conference Editorial Board\, co-chair of 
 the IEEE RAS Technical Committee on Humanoid Robots (2010-2014)\, Editor o
 f the Robotics and Automation Letters\, Associate Editor of Transactions o
 n Robotics (2010-2014). He is president of the Executive Board of the Germ
 an Association of Robotics (Deutsche Gesellschaft für Robotik\, DGR) and 
 member of the Board of Directors of euRobotics (2013-2015).Prof. José San
 tos-Victor\, IST\nJosé Santos-Victor received the Ph.D. degree in electri
 cal and computer engineering from Instituto Superior Técnico (IST)\, Lisb
 on\, Portugal\, in 1995. He is currently Full Professor at IST\, Director 
 of the Institute for Systems and Robotics (ISR-Lisboa) and head of the Com
 puter and Robot Vision Laboratory (VisLab). He was the scientific responsi
 ble for IST in many European and national research projects in computer vi
 sion and robotics. He is particularly interested in cognitive robotics\, i
 n some cases inspired by biological systems\, or as embodied tools to help
  understanding biological systems. From 2006 and January 2015 he was the I
 ST VIce President for International Affairs.\nNCCR Robotics cordially invi
 tes you to join us for three seminars hosted by the Learning Systems and A
 lgorithms Laboratory (LASA).\n13h15-14h00 - Oliver Brock\, Technische Univ
 ersität BerlinDeep Learning - Stairway to (Robotic) Heaven?\nDeep Learnin
 g is certainly a huge success.  Some believe it might solve robotics (wha
 tever that may mean).  I would like to take a step back and analyze the d
 ifferent types of problems we face in robotics\, speculate about the types
  of problems for which Deep Learning it most promising\, and present some 
 alternative and/or complementary ideas.  Such alternatives include:\n1) e
 xcessive sensing feedback for simple linearized local models\,\n2) using g
 eneral priors to aid learning in non-deep settings\, and\n3) the use of co
 ntextual data to assist learning\, also in non-deep settings. \nAnd then 
 I guess my not-very-surprising conclusion will be two-fold:\nA) Pick the "
 right" tool for the job!\nB) To "solve robotics" we'll need more than one 
 tool!\n14h00-14h45 - Tamim Asfourm\, Karlsruhe Institute of Technology (KI
 T)On the Duality of Grasping and Balancing\nThe talk will discuss the para
 llelism between grasping and whole-body balancing. Inspired by the idea th
 at a stable whole-body configuration of a humanoid robot can be seen as a 
 stable grasp on an object\, we present taxonomy for whole-body poses\, who
 le-body grasps\, and its validation based on human motion capture data.  
 Further\, we show how co-joint object-action representations used for obje
 ct grasping can be extended to associate whole-body actions with affordanc
 es of objects and environmental elements in the scene. We demonstrate how 
 affordance hypotheses are generated through visual exploration and verifie
 d using haptic feedback.\n14h45-15h30 - José Santos-Victor\, Instituto Su
 perior Tecnico (IST)Affordances and motor primed attention for humanoid ro
 bot\nIn this talk I will briefly present a model of affordances that allow
  a robot to learn how to interact with objects in a meaningful way and how
  the model can be extended for grounding language and for learning how to 
 use tools. Affordances depend on the physical characteristics of the of th
 e specific agent. As a related topic i will also describe how motor inform
 ation primes attention in primates and... robots.
LOCATION:CM 1 4 https://plan.epfl.ch/?room==CM%201%204
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR