Programme
All talks will be held in lecture hall H6 at the western end of the main hall.
Keynote: Marc van Hulle
Wednesday, 2007-09-05 - from 09:00 to 10:00
Kernel-based topographic maps
Abstract:
Rather than developing topographic maps with disjoint and uniform activation regions (Voronoi tessellation), as in the case of the popular SOM algorithm, and its adapted versions, algorithms have been introduced that can accommodate neurons with overlapping activation regions, usually in the form of kernel functions, such as Gaussians. For these kernel-based topographic maps, or kernel topographic maps, as they are called, several learning principles and -schemes have been proposed. The purpose of this talk is to review the theory behind the kernel-based topographic map algorithms, to indicate some of the difficulties, and to show a number of applications.