EP 1934895 A2 20080625 - METHOD FOR THE COMPUTER-AIDED STUDYING OF A NEURAL NETWORK, AND NEURAL NETWORK
Title (en)
METHOD FOR THE COMPUTER-AIDED STUDYING OF A NEURAL NETWORK, AND NEURAL NETWORK
Title (de)
VERFAHREN ZUM RECHNERGESTÜTZTEN LERNEN EINES NEURONALEN NETZES UND NEURONALES NETZ
Title (fr)
PROCEDE D'APPRENTISSAGE ASSISTE PAR ORDINATEUR D'UN RESEAU NEURONAL, ET RESEAU NEURONAL CORRESPONDANT
Publication
Application
Priority
- EP 2006066523 W 20060920
- DE 102005046747 A 20050929
Abstract (en)
[origin: DE102005046747B3] The method involves dividing the neurons of a neural network (1) into two interlaced layers (L1,L2). The first layer stores feature instances represented by the input data. The second layer stores majority of categories covering the neurons in the second layer. Each category in second layer is assigned to each feature instance in the first layer. The varying state of the neural network is determined based on the categories with assigned feature instances after entering the input data. The cross-links between first and second layer is changed based on the determined state of the neural network. An independent claim is included for the neural network.
IPC 8 full level
G06N 3/08 (2006.01)
CPC (source: EP US)
G06F 18/29 (2023.01 - EP US); G06N 3/08 (2013.01 - EP US)
Citation (search report)
See references of WO 2007036465A2
Designated contracting state (EPC)
AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IS IT LI LT LU LV MC NL PL PT RO SE SI SK TR
Designated extension state (EPC)
AL BA HR MK RS
DOCDB simple family (publication)
DE 102005046747 B3 20070301; EP 1934895 A2 20080625; US 2010088263 A1 20100408; US 8423490 B2 20130416; WO 2007036465 A2 20070405; WO 2007036465 A3 20071221
DOCDB simple family (application)
DE 102005046747 A 20050929; EP 06806783 A 20060920; EP 2006066523 W 20060920; US 99278506 A 20060920