0


实体、事件和事实的提取和解析能力

Extraction and Resolution Capabilities for Entities, Events and Facts
课程网址: http://videolectures.net/estc08_jackson_perce/  
主讲教师: Peter Jackson
开课单位: 路透社
开课时间: 信息不详。欢迎您在右侧留言补充。
课程语种: 英语
中文简介:
我们讨论了当前关于使用机器提取和解析来传递我们所称的“智能信息”的一些想法。这些信息是为用户的基础任务量身定制的,必须从基础内容源自动派生。我们描述了一个名为ClearForest标签的平台,它执行最先进的实体、事实和事件提取,并用一些商业用例说明了它的实用性。我们还描述了一个名为dexter/concord的实体解析工具包,它能够将命名的实体映射到授权文件,并合并或更新数据库记录。我们提供了这些完全实现和部署的系统的一些技术细节,并描述了更紧密地集成它们的计划。
课程简介: We discuss some current thinking around the use of machine extraction and resolution to deliver what we call 'Intelligent Information'. Such information is tailored to a user's underlying task, and must be derived automatically from underlying content sources. We describe a platform called ClearForest Tags, which performs state of the art entity, fact and event extraction, and illustrate its utility with some commercial use cases. We also describe an entity resolution toolkit called Dexter/Concord, which is capable of both mapping named entities to authority files and merging or updating database records. We supply some technical details of these fully implemented and deployed systems, as well as describing plans to more tightly integrate them.
关 键 词: 机器提取; 智能信息; 实体解析工具
课程来源: 视频讲座网
最后编审: 2019-12-10:cwx
阅读次数: 33