R. Kamimura and O. Uchida (Japan)
: mutual information maximization, competi tive learning, winner-take-all, Minkovsky distance
In this paper, we propose a new network-growing method to accelerate learning and to extract explicit features in complex input patterns. We have so far proposed a new type of network-growing algorithm called greedy network growing algorithm[8],[9]. By this algorithm, a network can grow gradually by maximizing information on input patterns. In the algorithm, the inverse of the square of the ordinary Euclidean distance between input patterns and connection weights is used to produce competitive unit outputs. When applied to some problems, the method has shown slow learning, and sometimes the method can not produce a state where information is large enough to produce explicit internal representations. To remedy this shortcoming, we here introduce Minkovsky distance be tween input patterns and connection weights used to pro duce competitive unit outputs. When the parameter for Minkovsky distance is larger, some detailed parts in in put patterns can be eliminated, which enables networks to converge faster and to extract main parts of input patterns. We applied our new method to the famous dipole problem and the actual computer guidance questionnaire analysis. In both experiments, results confirm that a new method with Minkovsky can significantly accelerate learning, and clearer features can be extracted.
Important Links:
Go Back
