摘要 |
A method and device for human-machine voice interaction. The method for human-machine voice interaction comprises: while a terminal is voice broadcasting a broadcast result, receiving a voice recognition result transmitted by a voice recognition server (101); transmitting the voice recognition result to a QU server for context comprehension, receiving and storing a context comprehension result (102); determining, on the basis of the context comprehension result stored, the intention of voice inputted by a user, generating a broadcast result on the basis of the intention (103); and, transmitting the broadcast result to the voice recognition server to allow the voice recognition server to transmit the broadcast result to the terminal for voice broadcasting (104). This allows voice broadcasting and user voice input to be implemented concurrently in a human-machine voice interaction process, thus obviating the need for repeated switchovers between a recording state and a broadcasting state in the human-machine interaction process, and allowing increased coherence for multiple rounds of dialogue. |