RETRACTED: Research on the application of speech recognition in computer network technology in the era of big data (Retracted Article)

After the emergence of big data, more and more complex deep models have revealed the huge amount of information carried by millions of data. In speech recognition, the coordination and availability of training data are the main factors to improve system performance. Big data is a fuzzy image of a phenomenon, usually defined as the need for new processing models to achieve better policy design, as well as the ability to identify and optimize processes that adapt to the abundance, growth, and diversification of information products. With the emergence of the concept of deep learning, various types of neural networks have gradually begun to be applied to language models. The most common is the Linked Deep Neural Network (DNN), which is connected with the Hybrid Neural Network (CNN) and Recurrent Neural Network (RNN) to improve their models (LSTM). Now, compared with traditional methods, speech acoustics and language models based on deep neural networks have significantly improved performance and have begun to become the mainstream in practice. But the speech recognition of deep neural networks still needs to be improved. With the advent of big data, we are getting more and more data. Therefore, how to effectively use deep neural network speech recognition system is a challenge. In addition, sound and speech signals are long-term signals, which are very important for the design of acoustic models and language models. RNN or LSTM can represent long-term related success models. But training is more complicated. Therefore, it is very important to develop a complete and very effective neural model with long-term modeling functions for big data.