Using Genetic Algorithm to Improve the Performance of Speech Recognition Based on Artificial Neural Network

The goal of this paper is to apply artificial neural network (ANN) to recognize speech. We use Genetic algorithm (GA) to replace the Steepest Descent Method (SDM) for the training of BPNN such that a global search of optimal weight in neural network can be. Thus, the performance of speech recognition was improved by the proposed method in this paper. The non-specific speaker recognition, which is trained by SDM, The recognition rate achieve up to 91% in this experiment. This paper will show that if BPNN is trained by genetic algorithm, higher recognition rate will be attained. Key words: back-propagation neural network, genetic algorithm and non-specific speaker speech recognition