Training method and system for thumbnail sentence understanding model

Abstract

The embodiment of the invention provides a training method for a thumbnail sentence understanding model. The method comprises the following steps: receiving a dialogue training data set; constructinga thumbnail sentence understanding model based on an encoder-decoder framework; taking the first round of complete sentences and the second round of abbreviated sentences as input of an encoder-codinglayer, and determining a second sentence feature vector of the second round of abbreviated sentences through a self-attention mechanism; based on the first word feature vector and the second sentencefeature vector, determining a relationship feature vector between the second round of abbreviated sentences and the first round of complete sentences; and enabling the decoder-decoding layer to generate a simulated complete statement of the second round of abbreviation sentence based on the relationship feature vector, and train a thumbnail sentence understanding model based on the simulated complete statement and the target complete statement. The embodiment of the invention further provides a training system of the thumbnail sentence understanding model. According to the embodiment of the invention, the model is generated in the neural network, the omitted sentence is restored into the complete sentence, the understanding effect of the thumbnail sentence understanding model is improved,and the reply satisfaction rate of the user to the dialogue system is effectively improved.

Description

technical field [0001] The invention relates to the field of natural language processing, in particular to a training method and system for an abbreviated sentence understanding model. Background technique [0002] Natural language human-computer interaction is a hot field in the development of artificial intelligence at present, and it is widely used in our lives, such as "companion robot", "vehicle voice navigation", "smart home appliances" and so on. In the process of human-computer interaction, users often give a lot of abbreviated sentences due to habit, such as "change it to something else", "what do you think", "is it so" and so on. [0003] In the process of realizing the present invention, the inventors have found that there are at least the following problems in the related art: [0004] Most of the current dialogue systems are single-turn dialogue systems, which can hardly understand the real intention of the user correctly. At present, there are mainly two solu...

AI Technical Summary

Problems solved by technology

[0007] In order to at least solve the problem that the abbreviated sentence understanding model in the prior art is only used in a closed field, has limitations, and has poor understanding effect

Images