Polyphone processing method and device and model training method and device

Abstract

The invention provides a polyphone processing method and device and a model training method and device. The polyphone processing method comprises the steps: receiving a Chinese statement containing polyphones, and determining first position features of the polyphones in the Chinese statement; performing word segmentation on the Chinese statement, and determining part-of-speech features of the polyphone based on words obtained by word segmentation and part-of-speech of the words; obtaining a word containing a polyphone, and determining a second position feature of the polyphone in the word containing the polyphone; inputting the first position feature, the part-of-speech feature and the second position feature into a polyphone disambiguation model to obtain pronunciation corresponding to the polyphone.

Description

technical field [0001] The present application relates to the field of computer technology, in particular to a method and device for processing polyphonic characters, a method and device for model training, a computing device, and a computer-readable storage medium. Background technique [0002] The earliest Chinese polyphone ambiguity disambiguation system was based on rules, which were summarized by linguists and encoded into computers in a computer-understandable manner; however, with the increase of data scale, data-based statistical models have gradually become the The preferred solution for disambiguating phonetic characters, that is, traditional machine learning models (such as decision tree models, maximum entropy models) and deep neural networks have begun to be applied to disambiguating Chinese polyphonic characters. [0003] However, the rule-based disambiguation system for Chinese polyphonic characters has the problem that a polyphonic character may match multipl...

AI Technical Summary

Problems solved by technology

[0003] However, the rule-based disambiguation system for Chinese polyphonic characters has the problem that a polyphonic character may match multiple rules or a rule may match multiple correct pronunciations of a certain polyphonic character, resulting in rule conflicts. The neural network Chinese of the statistical model In the polyphonic character disambiguation system, when the neural network is used as the feature extraction module, there will be a problem of complex feature forms, and the extracted feature dimension is too high, which increases the cost of training and reasoning, and reduces the prediction efficiency. As a feature extraction module, the analysis process will have the problem that the required feature vocabulary is too large and the encoding is not accurate enough

Images