Dialogue intention recognition method and device

Abstract

The invention discloses a dialogue intention recognition method and device. The dialogue intention recognition method comprises the steps that statistics is conducted on the occurrence probability ofeach historical user intention in historical dialogue data to serve as a priori knowledge sequence; predicting the probability q1 that the currently input text belongs to the user intention ai according to a preset classification model; calculating the probability q2 that the currently input text belongs to the user intention ai by utilizing a priori knowledge sequence; and according to the probability q1 and the probability q2, selecting the intention with the maximum probability as the current intention of the currently input text. Aiming at an existing dialogue intention recognition method,the current intention of the user is predicted by inserting the priori knowledge sequence into the current input text on the basis of predicting the probability that the current input text belongs tothe intention of a certain user in the training model; the problem that in the prior art, only text information input by a user at the current moment is used for judging the intention of the user, but the current intention of the user cannot be accurately judged is effectively solved, and the experience feeling of the user is enhanced.

Description

technical field [0001] The present invention relates to the technical field of artificial intelligence, in particular to a method and device for identifying dialogue intentions Background technique [0002] With the development of search engine technology, what modern search engines, question answering systems and dialogue robots need is no longer simple correlation information retrieval, but a deep understanding of users' information needs in order to accurately provide users with the services they need. Correctly identifying user intent is a critical step in achieving this goal. [0003] The dialogue intent recognition methods in the prior art mainly include methods based on keywords, based on template matching and based on machine learning models. However, these methods only use the text information input by the user in the dialogue at the current moment to determine the user's intention, and cannot accurately determine the user's current dialogue intention. For the user...

AI Technical Summary

Problems solved by technology

However, these methods only use the text information entered by the user at the current moment in the dialogue to judge the user's intention, and cannot accurately judge the user's current dialogue intention. For the user, the experience is extremely poor.

Images