Man-machine dialogue method and system

Abstract

An embodiment of the present invention provides a method for man-machine dialogue. The method includes: receiving the first audio uploaded by the user through the client, marking the start time point and the end time point of the first audio, using an audio decoder to generate a first recognition result of the first audio; according to the start time point of the first audio and the end time point, determine whether the first audio is a short sentence, and when it is a short sentence, if the second audio uploaded by the client is received within the preset heartbeat protection time range, the audio decoder is used to generate the second audio The second recognition result: send at least the combination of the first recognition result and the second recognition result to the language prediction model, when it is a sentence, generate an answer instruction corresponding to the combined sentence, and send the answer instruction together with the feedback time stamp of the answer instruction to client. The embodiment of the present invention also provides a man-machine dialogue system. The embodiment of the present invention solves the problems of unreasonable sentence segmentation in the full-duplex dialogue scene and redundant responses in the dialogue.

Description

technical field [0001] The invention relates to the field of intelligent voice dialogue, in particular to a method and system for man-machine dialogue. Background technique [0002] In a general question-and-answer system, ask one sentence and answer one sentence, or use multiple rounds of interaction. The effect that full-duplex interaction hopes to achieve is similar to making a phone call between people. Say more, and then the robot will answer; even the robot can actively ask questions to help interact, and use rhythm control technology to adjust the amount of its own words according to the size and content of the user's words. [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] 1. The answer is irrelevant. According to the response results of the existing equipment, the sentences arriving at the client will be broadcast continuously. When the network delay or th...

AI Technical Summary

Problems solved by technology

[0006] In order to at least solve the problem in the prior art that due to the loss of timeliness of the answer, the input and output content do not correspond, and the discontinuity of the upper and lower audio leads to unreasonable sentence segmentation, so that the response content cannot be reasonably decided.

Images