Title :
Top-down teaching enables task-relevant classification with competitive learning
Author :
De Sa, Virginia ; Ballard, Dana
Author_Institution :
Dept. of Comput. Sci., Rochester Univ., NY, USA
Abstract :
A method of augmenting the basic competitive learning algorithm with a top-down teaching signal which allows task relevant information to guide the development of synaptic connections is described. This teaching signal removes the restriction inherent in unsupervised learning and allows high-level structuring of the representation while maintaining the speed and biological plausibility of a local Hebbian-style learning algorithm. The function of the teaching input is illustrated geometrically, and examples of the use of this algorithm in small problems are presented. This work supports the hypothesis that cortical back-projections are important for the organization of sensory traces during learning
Keywords :
learning (artificial intelligence); neural nets; biological plausibility; competitive learning; cortical back-projections; high-level structuring; local Hebbian-style learning algorithm; sensory traces; synaptic connections; task relevant information; task-relevant classification; top-down teaching; unsupervised learning; Approximation algorithms; Biological system modeling; Biology; Computer science; Education; Hebbian theory; Neurons; Pattern recognition; Supervised learning; Unsupervised learning;
Conference_Titel :
Neural Networks, 1992. IJCNN., International Joint Conference on
Conference_Location :
Baltimore, MD
Print_ISBN :
0-7803-0559-0
DOI :
10.1109/IJCNN.1992.227147
