The Application of Improved Genetic Algorithm on the Training of Neural Network for Speech Recognition

This paper adopted artificial neural network (ANN) to recognize Mandarin digit speech. The nonspecific speaker speech recognition was the target of this paper. Genetic algorithm (GA) was first used to replace steepest descent method (SDM) and make a global search of optimal weight in neural network. The improved GA is then used to train the ANN. We can find that the performance of speech recognition was improved by the later method. The experiment in this paper will show that if BPNN is trained by GAs, higher recognition rate will be attained.