Deep Learning in Automatic Speech Recognition (ASR): A Review

In today's big data era, many traditional machine learning algorithms for processing large amount of raw unlabeled speech data are no longer applicable.At the same time, deep learning models, with their powerful modeling capability for massive data, are able to process unlabeled data directly and have become a hot research topic in the field of automatic speech recognition.The paper gives an overview of the application of deep learning in speech recognition.The research results of deep learning in the field of speech recognition in recent years are introduced, the correlation between traditional speech recognition models and the current deep learning models is discussed, the development trend of deep learning in the field of speech recognition is analyzed, and it is pointed out that deep learning models need to absorb the ideas of traditional speech recognition models in order to better build a speech recognition system based on deep learning models.