Title:
SELF-ORGANIZING FEATURE MAP WITH IMPROVED PERFORMANCE BY NON-MONOTONIC VARIATION OF THE LEARNING RATE
Document Type and Number:
WIPO Patent Application WO2003063017
Kind Code:
A3
Abstract:
The learning rate used for updating the weights of a self-ordering feature map is determined by a process that injects some type of perturbation into the value so that it is not simply monotonically decreased with each training epoch. For example, the learning rate may be generated according to a pseudorandom process. The result is faster convergence of the synaptic weights.
Inventors:
GUTTA SRINIVAS V R
PHILOMIN VASANTH
TRAJKOVIC MIROSLAV
PHILOMIN VASANTH
TRAJKOVIC MIROSLAV
Application Number:
PCT/IB2003/000170
Publication Date:
July 01, 2004
Filing Date:
January 21, 2003
Export Citation:
Assignee:
KONINKL PHILIPS ELECTRONICS NV (NL)
International Classes:
G06N3/00; G06K9/62; G06N3/08; (IPC1-7): G06N3/08
Foreign References:
| US5809490A | 1998-09-15 | |||
| FR2668625A1 | 1992-04-30 | |||
| US5398302A | 1995-03-14 |
Other References:
SCHNITMAN L ET AL: "An efficient implementation of a learning method for Mamdani fuzzy models", PROCEEDINGS SIXTH BRAZILIAN SYMPOSIUM ON NEURAL NETWORKS, RIO DE JANEIRO, BRAZIL, 22-25 NOV 2000, LOS ALAMITOS, CA, USA, IEEE COMPUT. SOC, USA, 2000, pages 38 - 43, XP010527545, ISBN: 0-7695-0856-1
WEISHUI WAN ET AL: "A new method to prune the neural network", PROCEEDINGS OF THE IEEE-INNS-ENNS INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS, COMO, ITALY, 24-27 JULY 2000, LOS ALAMITOS, CA, USA, IEEE COMPUT. SOC, USA, 2000, pages 449 - 454 VOL.6, XP010505028, ISBN: 0-7695-0619-4
WEISHUI WAN ET AL: "A new method to prune the neural network", PROCEEDINGS OF THE IEEE-INNS-ENNS INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS, COMO, ITALY, 24-27 JULY 2000, LOS ALAMITOS, CA, USA, IEEE COMPUT. SOC, USA, 2000, pages 449 - 454 VOL.6, XP010505028, ISBN: 0-7695-0619-4
Download PDF:
Previous Patent: SYSTEM, METHOD AND INTERFACE FOR CONTROLLING SERVER LIFECYCLE
Next Patent: FUNCTIONAL PIPELINES
Next Patent: FUNCTIONAL PIPELINES