Ai Speech Co., Ltd.

Weisi Shi

Shuai Fan

Hongbo Song

Chengya Zhu

Disclosed are a method and apparatus for testing a full-duplex speech interaction system. The method includes: determining a scene mixed corpus set by mixing a valid corpus set related to a test scene with an invalid corpus set unrelated to the test scene; playing each corpus audio in the scene mixed corpus set to a speech interaction device under test equipped with the full-duplex speech interaction system; acquiring a work log of the speech interaction device under test, the work log including at least a first log and a second log; and obtaining number of false responses by counting number of log entries which have false response result in the second log, and determining a false response rate based on the number of false responses and a total number of corpus audios played. End-to-end testing of the full-duplex speech interaction system is realized.

Method and apparatus for testing full-duplex speech interaction system

A server-side processing method for implementing an active initiation of a dialogue is disclosed, comprising: establishing a communication connection with a voice client, in response to a received request for establishing a connection from the voice client; receiving an information stream sent by the voice client through the communication connection; performing a dialogue decision-making process according to the information stream, obtaining and outputting an adapted dialogue content to the voice client upon determining that it is an active dialogue scenario. A server and a system for implementing an active initiation of a dialogue are also provided. The disclosed solutions realize intelligent decision-making for voice interaction, and can actively initiate a dialogue based on server-side decision-making, improving interaction experience and realizing intelligent interaction.

Server-side processing method and server for actively initiating dialogue, and voice interaction system capable of initiating dialogue

The present disclosure discloses a man-machine dialogue mode switching method, which is applicable to an electronic device. The method includes receiving a current user sentence spoken by a current user; determining whether a dialogue field to which the current user sentence belongs is a preset dialogue field; if yes, switching the current dialogue mode to a full-duplex dialogue mode; and if not, switching the current dialogue mode to a half-duplex dialogue mode. In the present disclosure, the dialogue mode is switched by determining whether the dialogue field to which the current user sentence belongs is the preset dialogue field, and the dialogue mode can be automatically switched and adjusted according to the difference of the dialogue fields, such that the man-machine dialogue is always in the most suitable dialogue mode and can be realized smoothly.

Man-machine dialogue mode switching method

A skill dispatching method for a speech dialogue platform including: receiving, by a central control dispatching service, a semantic result of recognizing a user's voice sent by a data distribution service; dispatching, by the central control dispatching service, a plurality of skill services related to the semantic result in parallel, and obtaining skill parsing results from the plurality of skill services; sorting the skill parsing results based on priorities of the skill services, and exporting a result with the highest priority to a skill realization discrimination service; when failure in realization, selecting a result with the highest priority among the rest of skill parsing results and exporting the same to the skill realization discrimination service, and when success in realization, sending the result with the highest priority to the data distribution service for feedback to the user. The method improves skill dispatching efficiency, reduces delay, and improves user experience.

Skill dispatching method and apparatus for speech dialogue platform

An embodiment of the present invention provides a method of man-machine interaction, including: receiving first audio uploaded by a user through a client end, marking a start time and an end time of the first audio, and generating a first recognition result of the first audio using an audio decoder; determining whether the first audio is a short speech based on the start time and end time thereof, and in case of a short speech, generating a second recognition result of the second audio using the audio decoder upon receiving the second audio uploaded by the client end within a preset heartbeat protection time range, sending at least the first recognition result and the second recognition result to a language prediction model; and if it is determined that a combination of the recognition results constitutes a sentence, generating an answering instruction corresponding to the sentence, and sending the answering instruction together with a feedback time mark of the answering instruction to the client end. Unreasonable sentence segmentation in a full-duplex dialogue scenario and redundant replies in the dialogue can thereby be avoided.

Growing community of inventors

Suzhou, China

Weisi Shi