PSOM+ : Parametrized Self-Organizing Maps for noisy and incomplete data

We present an extension to the Parametrized Self-Organizing Map that allows the construction of continuous manifolds from noisy, incomplete and not necessarily grid- organized training data. All three problems are tackled by minimizing the overall smoothness of a PSOM manifold. For this, we introduce a matrix which defines a metric in the space of PSOM weights, depending only on the underlying grid layout. We demonstrate the method with several examples, including the kinematics of a PA10 robot arm.

[1]  Aluizio F. R. Araújo,et al.  Modeling and Production of Robot Trajectories Using the Temporal Parametrized Self Organizing Maps , 2003, Int. J. Neural Syst..

[2]  Helge Ritter,et al.  Local PSOMs and Chebyshev PSOMs -Improving the Parametrised Self-Organizing Maps , 1995 .

[3]  Helge J. Ritter,et al.  Parametrized SOMs for hand posture reconstruction , 2000, Proceedings of the IEEE-INNS-ENNS International Joint Conference on Neural Networks. IJCNN 2000. Neural Computing: New Challenges and Perspectives for the New Millennium.

[4]  Helge Ritter,et al.  Parametrized Self-Organizing Maps , 1993 .

[5]  Carme Torras,et al.  Learning Inverse Kinematics via Cross-Point Function Decomposition , 2002, ICANN.

[6]  Jörg A. Walter.  PSOM network: learning with few examples , 1998, Proceedings. 1998 IEEE International Conference on Robotics and Automation (Cat. No.98CH36146).

[7]  Teuvo Kohonen,et al.  Self-Organizing Maps , 2010 .

[8]  Helge J. Ritter,et al.  Parametrized SOMs for Object Recognition and Pose Estimation , 2002, ICANN.