Small sample corpus-based intention classification method and apparatus, and computer device

Abstract

The embodiment of the invention provides an intention classification method and device based on small sample corpora and computer equipment, and the method comprises the steps: constructing a first sample data set which comprises a labeled sample data set and an unlabeled sample data set; obtaining a plurality of initial weak classification models; training each initial weak classification model based on the labeled sample data set to obtain a target weak classification model; inputting a label-free sample data set into each target weak classification model to obtain a first prediction classification label; constructing a second sample data set based on a label-free sample data set and the first prediction classification label; and training the initial intention classification model based on the second sample data set to obtain a target intention classification model, performing intention classification by using the target intention classification model, and outputting a classification result. Through the steps, accurate intention classification can be realized on the basis of a small number of labeled samples and unlabeled samples.

Description

technical field [0001] This application relates to the field of artificial intelligence, in particular to a method, device and computer equipment for intent classification based on small sample corpus. Background technique [0002] Intent classification is an important function of the robot, and the intent classification refers to the need for the robot to identify the user's intent. For example, user A: "Help me find out what the weather will be like tomorrow", user A's intention is "check the weather". User B: "Hi, what's your name?" User B's intention is to chat. Chatbots must be able to accurately identify the user's intentions in order to lay the foundation for subsequent processing (such as word slot extraction) and improve user satisfaction. [0003] Robot intent classification can generally be divided into chat type and task type. In related technologies, intent classification is usually implemented by a method based on supervised learning, which requires a large a...

AI Technical Summary

Problems solved by technology

In practical application scenarios, there are few or no corpus related to the actual business that can be used. With the accumulation of dialogue logs for a period of time, corpus data related to the actual business can be obtained. However, manually labeling the above corpus data will Consume a lot of human, material and financial resources

Images