jsteil's picture
Name: apl. Prof. Dr. Jochen J. Steil
Position: Managing Director CoR-Lab
Phone: +49 (0) 521 106-67109
Fax: +49 (0) 521 106-89042
Office: N4-122
E-Mail: jsteil[AT]cor-lab.uni-bielefeld.de

The CoR-Lab: The Research Institute for Cognition and Robotics (CoR-Lab) has been founded at the Bielefeld University in July 2007 to face the enormous challenges implied by creating intelligent systems and robots. CoR-Lab investigates artifical cognition, computer vision, neural networks, and robot learning and runs a Graduate School in Cognitions and Robotics, jointly with the strategic partner Honda Research Institute Europe. More that 70 researchers on all levels are involved in CoR-Lab projects. The CoR-Lab is headed by a directorium (apl. Prof. Steil, Prof. Ritter, Prof. Sagerer, Prof. Körner) and administred by J. Steil (Managing Director) and Dr. Carola Haumann (Manging Director Graduate School). Within CoR-Lab, J. Steil is  the head of a research group "cognitive robotics and learning".

Vita: In 1993, I received the diploma in mathematics from the Bielefeld University, and joined the Neuroinformatics Group headed by Prof. H. Ritter. In 1995/96 I stayed one year at the St. Petersburg Electrotechnical University, Russia, supported by a German Academic Exchange Foundation (DAAD) grant. In 1999, I received the PhD. Degree (Dr. rer. nat) with a Dissertation on "Input-Output Stability of Recurrent Neural Networks". In 2002, I was appointed tenured senior researcher (Akad. Rat) and in 2006 my habilitation was accepted by the Faculty of Technology  and I was assigned the venia legendi in Neuroinformatics. From March-July 2006 I was on leave as principal scientist at the Honda Research Institute Europe, Offenbach. In 2008, I have been appointed apl. Professor for Neuroinformatics at the Faculty of Technology. Since April 2007, I am managing director of the Research Institute for Cognition and Robotics, CoR-Lab.

Research Interests: Learning in artificial cognitive systems, learning and stability in recurrent networks, reservoir networks and nonlinear dynamics, representation in dynamical systems, robot learning, imitation learning, robot learning and system architectures, generative learning

Selected  Publication: The following paper introduces the online-learning algorithm BPDC for reservoir computing and shows that it naturally arises from a constraint optimisation of the standard quadratic error function.

Steil, Jochen J.. "Backpropagation-Decorrelation: online recurrent learning with O(N) complexity." Proc. IJCNN. Vol. 1., 2004843-848.  Download: Steil_BPDC_IJCNN2004.pdf (128.47 KB)

Research Projects: