公开(公告)号：WO2017111646A1

公开(公告)日：2017-06-29

申请号：PCT/RU2015/000919

申请日：2015-12-23

Applicant: MOTOROLA SOLUTIONS, INC.

Inventor： MARIASKIN, Evgenii Leonidovich ,  GENERALOVA, Valeriia Antonovna

IPC: G06N5/00

CPC classification number: G06N5/04 ,  G06F17/2775 ,  G06F17/30 ,  G10L2015/223 ,  G10L2015/228

Abstract: A method and apparatus are provided for analyzing conversational statements and providing feedback in real-time. Audio stream data is recorded by a communication device and converted into a logical analysis format representation. Logical analysis is performed automatically on the logical analysis format representation of the natural-language statement. The logical analysis identifies one or more candidate output resolutions and identifies supplemental data that can confirm the one or more candidate output resolutions. The identified supplemental data is accessed and the one or more candidate output resolutions are evaluated based at least in part on the supplemental data. A feedback message is generated and displayed on the mobile communication device. The feedback message is indicative of the evaluation of the one or more candidate output resolutions and may provide a recommended follow-up question to be asked by the user of the communication device in the current conversation.

Abstract translation: 提供了一种用于分析会话语句并实时提供反馈的方法和装置。 音频流数据由通信设备记录并转换成逻辑分析格式表示。 逻辑分析自动执行自然语言语句的逻辑分析格式表示。 逻辑分析标识一个或多个候选输出分辨率并标识可以确认一个或多个候选输出分辨率的补充数据。 识别的补充数据被访问，并且至少部分地基于补充数据来评估一个或多个候选输出分辨率。 反馈消息被生成并显示在移动通信设备上。 反馈消息表示对一个或多个候选输出分辨率的评估，并且可以提供要在当前对话中由通信设备的用户询问的推荐后续问题。