Method and system for enhancing features by aid of bidirectional long-term and short-term memory recurrent neural networks

Abstract

The invention discloses a method and a system for enhancing features by the aid of bidirectional long-term and short-term memory recurrent neural networks. The method includes steps of training inputted features with noise and corresponding features without noise by the aid of bidirectional long-term and short-term memory recurrent neural network models; enhancing the features by the aid of the bidirectional long-term and short-term memory recurrent neural network models. The method and the system have the advantages that the long-term and short-term memory recurrent neural network models can be built in two directions by the aid of the inputted features, and accordingly models of contexts of current frames can be effectively built; long-term and short-term memory cells and certain control variables are introduced into the neural networks, and accordingly the models can be built in the dependence on the long contexts; excellent enhancement effects can be realized by the models owing to long-term dependence when the features are enhanced by the aid of the models, and accordingly the performance of voice recognition systems and audio event classification systems can be improved.

Description

technical field [0001] The present invention relates to speech and audio information technology, in particular, the present invention relates to features for enhancing the performance of speech recognition systems, audio event detection and classification systems. Background technique [0002] With the development of computing technology and information technology, especially the proposal of deep machine learning method and its successful application in voice, image, video and other fields, the performance of speech recognition system has been greatly improved and improved, and speech recognition technology has been advanced by leaps and bounds. At the same time, there are some commercial recognition systems and application software, such as Google Voice Search, Bing Voice Search, Siri Voice Assistant, Baidu Voice Assistant, Sogou Voice Input, Xunfei Language Point, etc. Voice as a human-computer interaction tool is gradually changing The way people interact with different d...

AI Technical Summary

Problems solved by technology

This kind of feature transformation using a fixed time window or span cannot dynamically use the time window or span to perform feature transformation according to the context information, and cannot reflect the dynamic characteristics of the system well.

Images