Title :
An unsupervised learning of a layered network and its application to a motion acquisition
Author :
Nishikawa, Ikuko ; Matsunaga, Kentaro
Author_Institution :
Dept. of Comput. Sci., Ritsumeikan Univ., Shiga, Japan
fDate :
6/24/1905 12:00:00 AM
Abstract :
An unsupervised learning for a layered network is applied to a motion acquisition of an autonomous agent. A basic algorithm is extended in the following two ways for temporal series learning. One is a temporal reward assignment, and the other is a network with temporal integration units. Several simulations show a successful learning of collision avoidance and a capture of both static and moving targets
Keywords :
motion estimation; multilayer perceptrons; time series; unsupervised learning; autonomous agent; collision avoidance learning; layered network; motion acquisition; moving targets; multilayer neural network; static targets; temporal integration units; temporal reward assignment; temporal series learning; unsupervised learning; Application software; Collision avoidance; Computer science; Educational robots; Entropy; Neural networks; Neurons; Stochastic processes; Supervised learning; Unsupervised learning;
Conference_Titel :
Neural Networks, 2002. IJCNN '02. Proceedings of the 2002 International Joint Conference on
Conference_Location :
Honolulu, HI
Print_ISBN :
0-7803-7278-6
DOI :
10.1109/IJCNN.2002.1007768
