Natural language information data processing method and device

Abstract

The embodiment of the invention provides a natural language information data processing method and device. The natural language information data processing method includes: acquiring a first corpus and a second corpus; generating natural language information according to the second corpus; determining a second characteristic word in the natural language information; extracting a first characteristic word from the first corpus; and using the first characteristic word to replace the second characteristic word meeting a preset rule. According to the embodiment of the invention, more reasonable and effective first characteristic words can be provided through determination of the rationality of the replaced natural language information, and the second characteristic word is replaced; the natural language information can achieve reasonable matching and fluent statements; overfitting phenomenon can be accurately and reasonably eliminated in a generation process of a natural language; high quality natural language information can be generated; the matching rationality and the statement fluency of the generated natural language information can be ensured; and the statement creativity of the natural language information can be improved.

Description

technical field [0001] The present invention relates to the technical field of information processing, in particular to a data processing method for natural language information and a data processing device for natural language information. Background technique [0002] In the prior art, deep learning is a new field in machine learning research. Its motivation is to establish and simulate the neural network of human brain for analysis and learning. It imitates the mechanism of human brain to explain data, such as image information, text information and natural language information. [0003] With the application of deep learning in natural language information processing, rule-based natural language information processing is gradually being eliminated. Deep learning-based natural language information processing no longer requires cumbersome rule configuration, continuous addition and maintenance of rules, and syntactic analysis , Analyzing semantics, computers can create and...

AI Technical Summary

Problems solved by technology

However, in the process of generating natural language information, the computer's creative ability depends on the training corpus. When the corpus is not large enough, it will easily lead to overfitting of the generated content.

At the same time, the generated content is only based on the training corpus, the generated content is empty and single, and lacks creativity

Images