A multi-example keyword detection method based on a multi-task neural network

Abstract

The invention provides a multi-example keyword detection method based on a multi-task neural network, which belongs to the technical field of voice signal processing. The method comprises the steps of: training a bottleneck deep neural network on a multi-language data set; extracting fbank features from target data set audio frame by frame and extracting bottleneck features of the target data set;based on a training set, training an HMM model for each keyword by using the bottleneck features of the keywords, acquiring a frame-level station tag thereof and training a filler word model by usingthe bottleneck features of all non-keywords; performing multi-task DNN acoustic model training by using the bottleneck features; acquiring the acoustic score of test set audio and obtaining keyword detection results through Viterbi decoding. The multi-task technology can effectively improve low resource conditions and remarkably improve multi-example keyword detection performance.

Description

technical field [0001] The invention belongs to the technical field of speech signal processing, and in particular relates to a multi-example keyword detection method based on a multi-task neural network. Background technique [0002] Voice keyword detection technology, as one of the artificial intelligence technologies for processing massive audio data, provides a solution for people to quickly retrieve predefined keywords from massive voice data. At present, keyword detection can be divided into two categories according to different keyword retrieval objects: text-based keyword detection, keywords are given in the form of text; example-based keyword detection, keywords are given as speech fragments (sample ) form is given. In terms of algorithms, the mainstream technology of text keyword detection is based on LVCSR (Large Vocabulary Continuous Speech Recognition) and text matching; the mainstream technology of sample keyword detection is based on DTW (Dynamic Time Warping...

AI Technical Summary

Problems solved by technology

In actual application scenarios, text-based methods are difficult to achieve good results when dealing with languages ​​with scarce resources or dialects with a narrow range of use, or even unknown languages.

In this low-resource condition, the example-based method is more suitable, but when the number of examples for each keyword increases to tens of orders of magnitude, the performance improvement brought by the template matching method is limited

Images