首页> 外国专利> METHOD AND DEVICE FOR AUTOMATIC ELECTION OF INTERACTIVE TELEVISION CHANNELS BASED ON SPEAKING COMPREHENSION.

METHOD AND DEVICE FOR AUTOMATIC ELECTION OF INTERACTIVE TELEVISION CHANNELS BASED ON SPEAKING COMPREHENSION.

机译:基于语音理解自动选择交互式电视频道的方法和装置。

摘要

Speech understanding system that serves to receive a pronounced request from a user and processes the request by comparing it with a knowledge base that contains programming information to automatically select a television program, consisting of: a knowledge extractor (16) to receive the information contained in the electronic programming guide (EPG) and process the information to create a program database (18); a speech recognition device (32) to receive the pronounced request and convert it into a text string formed by a plurality of words; a natural language processor (34) that receives the text string and processes the words to solve the semantic content of the request made by the user, the natural language processor (34) locates the meaning of the words in a task framework (60) which includes a plurality of boxes with keywords (70); and a dialog manager (40) that analyzes the task framework (60) to determine whether a sufficient number of keyword boxes (70) have been filled in and request the user additional information to fill in the empty boxes (70), the Dialog manager performs a search in the program database using the keywords that the system has placed in the task frame (60) to select a program, and generates a signal to select the television channel associated with the program in question, characterized in that said dialog manager (40) automatically updates a historical database (54) from the recordings of previous pronounced requests and dialogues, deduces user preferences based on the historical database (54), and classify the questions to ask the user according to their preferences.
机译:语音理解系统,用于接收来自用户的明确请求,并通过将其与包含节目信息以自动选择电视节目的知识库进行比较来处理该请求,该系统包括:知识提取器(16),用于接收包含在其中的信息电子节目指南(EPG)并处理该信息以创建节目数据库(18);语音识别装置(32)接收发音的请求并将其转换成由多个单词组成的文本字符串;自然语言处理器(34)接收文本字符串并处理单词以解决用户提出的请求的语义内容,自然语言处理器(34)在任务框架(60)中定位单词的含义,包括多个带有关键词的盒子(70);以及对话管理器(40),其分析任务框架(60)以确定是否已经填充了足够数量的关键词框(70)并请求用户附加信息来填充空框(70),该对话管理器使用系统已放置在任务框架(60)中的关键字在节目数据库中执行搜索以选择节目,并生成信号以选择与所讨论节目相关的电视频道,其特征在于,所述对话管理器( 40)从先前宣布的请求和对话的记录中自动更新历史数据库(54),基于历史数据库(54)推导出用户偏好,并根据他们的偏好对问题进行分类。

著录项

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号