Training method and system for language model

Abstract

The embodiment of the invention provides a training method and system for a language model, wherein the method specifically comprises the following steps: acquiring seed corpuses of all fields, and training the seed models in the corresponding fields based on the seed corpuses of all the fields; screening big data corpuses by utilizing the perplexity of the corpuses on the seed models of all the fields to obtain seed screening corpuses of in the corresponding fields; respectively training by utilizing the seed screening corpuses of all fields to obtain the screening model of the corresponding field; fusing the screening models of all the fields to obtain corresponding screening fusion model. According to the training method and system for the language model disclosed by the embodiment of the invention, the parameter reasonableness of the language model can be improved on the premise of reducing a computation burden and saving time.

Description

technical field [0001] The invention relates to the technical field of natural language processing, in particular to a language model training method and system. Background technique [0002] In the field of natural language processing technology, speech recognition is a technology in which a machine converts speech signals into corresponding text or commands through the process of recognition and understanding. [0003] Speech recognition system is essentially a pattern recognition system, usually including basic units such as feature extraction, pattern matching and reference model. refer to figure 1 , shows a schematic structural diagram of an existing language recognition system, wherein the input speech signal is first analyzed by the feature extraction unit to form a feature vector, and then enters the word-level matching unit, and the word-level matching unit follows the dictionary and the subword model The word model formed by the collection and concatenation recog...

AI Technical Summary

Problems solved by technology

The existing language model training methods usually directly train the big data to obtain the language model. However, due to the large size and scale of the big data, directly training it not only needs to occupy more hard disk and memory consumption, but also requires a long time. Training time, that is, the existing language model training methods have problems such as occupying a large amount of computing resources and being time-consuming

[0005] In order to overcome the above-mentioned problems such as large amount of computing resources occupied and time-consuming, there are still some language model training methods that will cut some big data and then train the cut big data. However, the language model obtained by the above training is easy Lose the original statistical distribution of big data, resulting in a lower language recognition rate

Images