METHOD AND APPARATUS FOR SPEECH ENDPOINT DETECTION BASED ON NEURAL NETWORK, DEVICE, AND MEDIUM

A method and an apparatus for speech endpoint detection based on a neural network, a device, and a storage medium, relating to the technical field of speech detection. The method comprises performing acoustic feature extraction on sample speech files (S1), performing distribution of context feature information to speech features of each frame to obtain a feature matrix (S2), by means of a neural network model, performing feature processing on the feature matrix to obtain a one-dimensional feature vector, and performing learning processing of sequence information of the speech frames on the first feature vector to obtain a predicted value (S3); calculating a loss function value between the predicted value and an actual speech value, and on the basis of the loss function value, updating network parameters of the neural network model (S4), and by means of the trained neural network model, outputting a prediction result for the speech file for prediction (S6). The method further relates to blockchain technology,