Title :
Analyzing learning dynamics: How to average?
Author :
Goerick, Christian
Author_Institution :
Inst. fur Neuroinf., Ruhr-Univ., Bochum, Germany
Abstract :
Pattern-based learning processes are usually analyzed by means of probability density functions of the weights or moments thereof. During the derivation of these equations, some averaging has to be performed. We show that the manner of averaging is crucial for the results of the analysis. We do this by comparing two types of analysis (Langevin type and discrete-time moments) for one learning system
Keywords :
Markov processes; learning (artificial intelligence); neural nets; probability; Langevin type analysis; averaging; discrete-time moments analysis; learning dynamics; pattern-based learning processes; probability density functions; Difference equations; Differential equations; Evolution (biology); Information analysis; Markov processes; Nonlinear dynamical systems; Nonlinear equations; Pattern analysis; Probability density function; Stochastic processes;
Conference_Titel :
Neural Networks, 2000. IJCNN 2000, Proceedings of the IEEE-INNS-ENNS International Joint Conference on
Conference_Location :
Como
Print_ISBN :
0-7695-0619-4
DOI :
10.1109/IJCNN.2000.857896
