METHOD FOR CALCULATING HIGH-SPEED INTRAMUSCULAR NEURAL NETWORK USING HIGH-RANK N-BEST NORMALIZATION

摘要

The present invention relates to a method for calculating a high-speed intramuscular neural network for increasing a recognition speed through a high-rank N-BEST normalization method, regarding intramuscular neural network-based technology for recognizing a voice. According to one aspect of the present invention, the method for calculating a high-speed intramuscular neural network using the high-rank N-BEST normalization comprises the following steps of: selecting an activated node; calculating the node output with respect to the selected node; and calculating the normalized approximate probability of the intramuscular neural network with respect to the selected node.;COPYRIGHT KIPO 2017