Multi-round dialogue method and system for voice dialogue platform cross-skill

Abstract

An embodiment of the present invention provides a multi-round dialogue method for cross-skilled voice dialogue platforms. The method includes: identifying that the user's input sentence hits the first skill, feeding back the first output sentence based on the first skill, and at least storing the keywords included in the first output sentence and the general types corresponding to the keywords in the multi-round dialogue keyword table ; When the subsequent input sentence jumps to hit the second skill, if a certain keyword is missing, obtain the historical keyword of the same type as a certain keyword from the multi-round dialogue keyword table; according to the acquired historical The keywords supplement the subsequent input sentence, and the second output sentence is fed back based on the second skill. The embodiment of the present invention also provides a multi-round dialogue system for voice dialogue platform cross-skill. The embodiment of the present invention extracts the semantic result of the hit skill, uses a common type and saves it in the multi-round dialogue keyword table, and realizes the context that can be used across skills in the voice dialogue through a series of processing.

Description

technical field [0001] The invention relates to the field of intelligent voice dialogue, in particular to a multi-round dialogue method and system for a voice dialogue platform across skills. Background technique [0002] Multi-round dialogue In human-computer interaction, the user sends a dialogue request (which can be regarded as an intention), and after the user's intention is initially clarified, the necessary information is obtained to finally obtain a clear user instruction, and then the dialogue service responds to the user according to the skill customization Give feedback. [0003] Even in the communication between people, the information contained in a sentence itself only accounts for a part of the total amount of information transmitted, and the other part of the information comes from a series of information such as the speaker's identity, current time, and location. For example, a user is going on a business trip to Suzhou for a week and asks the smart assista...

AI Technical Summary

Problems solved by technology

[0006] In order to at least solve the problem in the prior art that due to the different data structures defined by each skill, it is impossible to use the context to complete the utterances input by the user in the voice dialogue caused by cross-skills

Images