Stefan Klanke Some time ago, I left academia. I now work in industry and live in good old Germany again. In my last academic job, I worked as a scientific programmer / PostDoc at the Donders Centre for Cognitive Neuroimaging in Nijmegen (NL), where my main task was to implement real-time streaming of fMRI data. Previously I worked as a postdoctoral research fellow with the Statistical Machine Learning and Motor Control Group at the Institute of Perception, Action and Behaviour, University of Edinburgh. I studied computer science at the University of Bielefeld, where I also did my PhD with the Neuroinformatics Group, Faculty of Technology under the supervision of Helge Ritter. My research interests are fairly broad and include non-parametric regression, dimensionality reduction and manifold learning, robotics, and learning for control. If you'd like to contact me, please use sk@sklanke·de.

Software

If you are looking for an incremental online regression algorithm, you might want to try Locally Weighted Projection Regression (LWPR), which fits multiple local linear models to your training data. As one of the deliverables of the SENSOPAC project, I have implemented that algorithm in plain C, with comfortable bindings and wrappers for Matlab, Python, and C++. The software package was one of the first to be accepted for the new JMLR track on open source (MLOSS). You can get it from the project page or directly from sourceforge.net.

While working in Nijmegen, I contributed various bits of code to Fieldtrip, a free Matlab toolkit for analysing neurophysiological data.

During my PhD studies I have written a Matlab toolbox for Unsupervised Kernel Regression (UKR), a method for nonlinear dimension reduction and manifold learning. If you want to give it a try, please use the following link.

UKR toolbox Version 2b (2007/04/05)

I've tested the software with Matlab v6.5 running on Windows and Linux. Feedback, comments, and bug reports are much appreciated.

Publications

Journal Papers

• Online decoding of object-based attention using real-time fMRI
  Adnan M. Niazi, Philip L. C. van den Broek, Stefan Klanke, Markus Barth, Mannes Poel, Peter Desain, Marcel A. J. van Gerven
  European Journal of Neuroscience, doi:10.1111/ejn.12405, 2013
• Using brain-computer interfaces and brain-state dependent stimulation as tools in cognitive neuroscience
  Ole Jensen, Ali Bahramisharif, Robert Oostenveld, Stefan Klanke, Avgis Hadjipapas, Yuka Okazaki, Marcel van Gerven Marcel
  Frontiers in Psychology 2/2011, doi:10.3389/fpsyg.2011.00100, 2011
• A Computational Model of Limb Impedance Control based on Principles of Internal Model Uncertainty
  Djordje Mitrovic, Stefan Klanke, Rieko Osu, Mitsuo Kawato, Sethu Vijayakumar
  PLoS ONE, 5(10): e13601. doi:10.1371/journal.pone.0013601, 2010
• Learning impedance control of antagonistic systems based on stochastic optimisation principles
  Djordje Mitrovic, Stefan Klanke, Sethu Vijayakumar
  International Journal of Robotic Research, doi:10.1177/0278364910387653, 2010
• A novel method for learning policies from variable constraint data
  Matthew Howard, Stefan Klanke, Michael Gienger, Christian Görick, Sethu Vijayakumar
  Autonomous Robots 27:105-121, doi:10.1007/s10514-009-9129-8, 2009
• Behaviour Generation in Humanoids by Learning Potential-based Policies from Constrained Motion
  Matthew Howard, Stefan Klanke, Michael Gienger, Christian Görick, Sethu Vijayakumar
  Applied Bionics and Biomechanics, Vol. 5, No. 4, pp. 195-211, 2008
• A Library for Locally Weighted Projection Regression
  Stefan Klanke, Sethu Vijayakumar, Stefan Schaal
  Journal of Machine Learning Research 9(Apr):623-626, 2008 pdf
• Variants of Unsupervised Kernel Regression: General Cost Functions
  Stefan Klanke, Helge Ritter
  Neurocomputing 70(7-9), pp. 1289-1303, 2007 view online
  
• Principal Surfaces from Unsupervised Kernel Regression
  Peter Meinicke, Stefan Klanke, Roland Memisevic, Helge Ritter
  IEEE Trans. on Pattern Analysis and Machine Intelligence, 27, (9), 1379-1391, Sep. 2005
  

Conference Proceedings

• Exploiting Sensorimotor Stochasticity for Learning Control of Variable Impedance Actuators
  Djordje Mitrovic, Stefan Klanke, Matthew Howard, Sethu Vijayakumar
  Proc. 2010 IEEE-RAS International Conference on Humanoid Robots, Nashville, TN, USA, 2010
• Optimal Feedback Control for Anthropomorphic Manipulators
  Djordje Mitrovic, Sho Nagashima, Stefan Klanke, Takamitsu Matsubara, Sethu Vijayakumar
  IEEE International Conference on Robotics and Automation (ICRA), 2010
• Robust Constraint-consistent Learning
  Matthew Howard, Stefan Klanke, Michael Gienger, Christian Goerick, Sethu Vijayakumar
  IEEE International Conference on Intelligent Robots and Systems (IROS), 2009
• Towards Semi-supervised Manifold Learning: UKR with Structural Hints
  Jan Steffen, Stefan Klanke, Sethu Vijayakumar, Helge Ritter
  7th Int. Workshop on Self-Organizing Maps (WSOM), 2009
• A Novel Method for Learning Policies from Constrained Motion
  Matthew Howard, Stefan Klanke, Michael Gienger, Christian Goerick, Sethu Vijayakumar
  IEEE International Conference on Robotics and Automation (ICRA), 2009
• Does dimensionality reduction improve the quality of motion interpolation?
  Sebastian Bitzer, Stefan Klanke, Sethu Vijayakumar
  Proc. 17th European Symposium on Artificial Neural Networks (ESANN), pp. 71-76, 2009
• Multi-task Gaussian Process Learning of Robot Inverse Dynamics
  Kian Ming Chai, Chris Williams, Stefan Klanke, Sethu Vijayakumar
  Advances in Neural Information Processing Systems (NIPS) 21, pp. 265--272, 2008
• Learning Potential-based Policies from Constrained Motion
  Matthew Howard, Stefan Klanke, Michael Gienger, Christian Görick, Sethu Vijayakumar
  IEEE Int. Conf. on Humanoid Robots 2008
• Adaptive Optimal Control for Redundantly Actuated Arms
  Djordje Mitrovic, Stefan Klanke, Sethu Vijayakumar
  From Animals To Animats 10, pp. 93-102, Springer, 2008
• Optimal Control with Adaptive Internal Dynamics Models
  Djordje Mitrovic, Stefan Klanke, Sethu Vijayakumar
  Proc. Intl. Conf. on Informatics in Control, Automation and Robotics (ICINCO), 2008
• Dynamic Path Planning for a 7-DOF Robot Arm
  Stefan Klanke, Dmitry Lebedev, Robert Haschke, Jochen J. Steil, Helge Ritter
  Proc. International Conference on Intelligent Robots and Systems (IROS), 3879-3884, Oct. 2006
• Variants of Unsupervised Kernel Regression: General Cost Functions
  Stefan Klanke, Helge Ritter
  Proc. European Symposium on Artificial Neural Networks (ESANN), 581-586, Apr. 2006
  
• A Leave-K-Out Cross-Validation Scheme for Unsupervised Kernel Regression
  Stefan Klanke, Helge Ritter
  Artificial Neural Networks -- ICANN 2006, LCNS, 4132, 427-436, Springer, Sep. 2006
• PSOM⁺: Parametrized Self-Organizing Maps for noisy and incomplete data
  Stefan Klanke, Helge Ritter
  Proc. of the 5th Workshop on Self-Organizing Maps (WSOM '05), Paris, France, Sep. 2005

Book Chapters

• Methods for Learning Control Policies from Variable Constraint Demonstrations
  Matthew Howard, Stefan Klanke, Michael Gienger, Christian Goerick, Sethu Vijayakumar
  In: O. Sigaud and J. Peters (eds.): From Motor Learning to Interaction Learning in Robots, SCI 264, pp. 253-291, Springer-Verlag, 2010
• Adaptive Optimal Feedback Control with Learned Internal Dynamics Models
  Djordje Mitrovic, Stefan Klanke, Sethu Vijayakumar
  In: O. Sigaud and J. Peters (eds.): From Motor Learning to Interaction Learning in Robots, SCI 264, pp. 65-84, Springer-Verlag, 2010

Workshop Contributions and Posters

• A Theory of Impedance Control based on Internal Model Uncertainty
  Djordje Mitrovic, Stefan Klanke, Adrian Haith, Sethu Vijayakumar
  Computational principles of sensorimotor learning, Kloster Irsee, Germany, 2009
• Realising Dextrous Manipulation with Structured Manifolds using Unsupervised Kernel Regression with Structural Hints
  Jan Steffen, Stefan Klanke, Sethu Vijayakumar, Helge Ritter
  ICRA Workshop: Approaches to Sensorimotor Learning on Humanoid Robots, 2009
• Optimal control with adaptive internal dynamics models
  Djordje Mitrovic, Stefan Klanke, Sethu Vijayakumar
  NIPS Workshop: Robotics Challenges for Machine Learning, 2007

PhD and Diploma Thesis

• Learning manifolds with the Parametrized Self-Organizing Map and Unsupervised Kernel Regression
  Stefan Klanke
  PhD thesis, Bielefeld University, Faculty of Technology, 2007 view
• Nichtlineare Dimensionsreduktion zur Resynthese von Klängen
  Stefan Klanke
  Diploma thesis, Bielefeld University, Faculty of Technology, 2003