Speech recognition training and application method and device

Abstract

The present invention discloses speech recognition training and application methods and devices, wherein, a speech recognition training method includes: acquiring a plurality of scenarios preset by developers, and first speech recognition resources corresponding to the plurality of scenarios and related information A plurality of second speech recognition resources corresponding to each scene in the plurality of scenes, wherein the first speech recognition resource is a common part of the plurality of scenes, and the second speech recognition resource is the A unique part of a scene; training a first model corresponding to the first speech recognition resource and a plurality of second models corresponding to the plurality of second speech recognition resources; using the trained first model and the first speech recognition resources are stored in each server, and the trained second models and the plurality of second speech recognition resources are stored in at least one server; thereby reducing speech recognition resources can be achieved Repetitive work during updates reduces maintenance costs.

Description

technical field [0001] The invention belongs to the technical field of speech recognition, in particular to a speech recognition training and application method and device. Background technique [0002] In recent years, with the continuous development of speech recognition technology, speech interaction has been widely used in our life and work, especially in smart home, smart car, smart phone, children's robot, smart customer service and other fields. [0003] In order to support voice interaction in different scenarios, most current voice recognition systems provide a dedicated system for each scenario. In order to ensure the accuracy of recognition, the dedicated speech recognition system for each scene needs to support the interactive commands that may be used in the scene, which will lead to a large speech recognition system for each scene, and there are certain differences between different scenes. duplication, resulting in high resource maintenance costs. In additio...

AI Technical Summary

Problems solved by technology

In order to support speech recognition services in each scene, the fixed recognition server allocation method usually allocates redundant recognition servers to each scene, and cannot be adjusted according to the actual call volume, resulting in waste of server resources and low utilization.

Images