Intention recognition method, system and device based on deep learning and storage medium

Abstract

The invention provides an intention recognition method, system and device based on deep learning and a storage medium. The method comprises the steps of acquiring and preprocessing the statement information to be analyzed; performing word segmentation on the statement information by using a word segmentation tool; weighting all the segmented words by using a pre-trained word vector model; obtaining a mood classification corresponding to the statement information by using a pre-trained first neural network; obtaining sentence pattern classification corresponding to the sentence information by using a pre-trained second neural network; obtaining expression classifications corresponding to the statement information by using a pre-trained third neural network; and predicting the intention category of the statement information by using a pre-trained fourth neural network according to the mood classification, sentence pattern classification and expression classification of the statement information. According to the method, the feature information of the statement text information in the corresponding field does not need to be extracted, and the intention type to be expressed by the statement can be accurately determined from the statement text information by analyzing the word vector of each segmented word in the statement text information.

Description

technical field [0001] The present invention relates to the field of language recognition, in particular, to a method, system, device and storage medium for intent recognition based on deep learning. Background technique [0002] Dialogue System is a human-computer interaction system based on natural language. Through the dialogue system, people can use natural language to interact with the computer for multiple rounds to complete specific tasks, such as information query, service acquisition, etc. The dialogue system provides a more natural and convenient way of human-computer interaction, and is widely used in vehicle, home, customer service and other scenarios. [0003] Among them, Natural Language Understanding (Natural Language Understanding) is the core module in the dialogue system. The goal of natural language understanding is to convert the text information of natural language into a semantic representation (Semantic Representation) that can be processed by a comp...

AI Technical Summary

Problems solved by technology

[0006] (1) The setting of rules usually relies on domain knowledge and needs professionals to set, which consumes manpower

[0007] (2) Due to the diversity of natural language expressions, a large number of rules need to be set to cover different expressions, but even so, it is impossible to exhaust all the rules

[0008] (3) Rule-based methods cannot continue to evolve based on actual data for self-learning. When encountering problems, they can only be solved by continuously setting new rules.

[0009] (4) If there are more and more rules, there may even be conflicts between the rules

First of all, the quality of features will affect the effect of the model, and feature design depends on domain knowledge, so professionals need to invest a lot of energy in feature design

Secondly, because most of the features are related to domain knowledge, the same feature that is effective in solving problems in one domain does not mean that the feature is still effective in solving problems in another domain, which makes the method less capable of cross-domain transfer. weak

Images