SBIR Phase II: Building a Large-Scale, Effective, Self-Maintainable and Customizable News Metasearch System

小型企业创新研究(SBIR)第二阶段项目开发了专为新闻搜索而设计的元搜索能力。搜索是互联网上仅次于电子邮件的第二大最受欢迎的活动，它已经拥有数十亿美元的广告市场。新闻搜索在所有搜索中占了很大比例。新闻项目可以从大量的在线来源获得，但当前的新闻搜索技术不能扩展到及时有效地涵盖所有这些来源。本项目旨在通过构建一个大规模、高效、可自我维护、可定制的新闻元搜索引擎来解决这一问题。通过为每个用户查询自动选择要访问的最合适的搜索引擎并通过有效地识别每个查询中的术语的正确含义来实现高效率。通过使用高度自动化的技术来整合搜索引擎，该系统可以自动适应对连接的搜索引擎所做的更改，用户可以通过添加他们最喜欢的新闻搜索引擎来进行定制。这里采用的高度自动化的解决方案降低了开发和维护的劳动力成本，这转化为更低的广告成本，并使在线广告对包括小型本地媒体网站、个人和小公司在内的“小参与者”来说更负担得起。该项目推进了大规模信息集成、大规模分布式信息检索、信息抽取、系统自动自维护和按需定制。拟议的技术使普通用户能够从大量新闻来源中搜索更相关和更新的新闻项目。它还使他们能够定制搜索系统，以满足他们的信息需求。

The Small Business Innovation Research (SBIR) Phase II project develops a metasearch capability engineered for news searching. Searching is the second most popular activity on the Internet behind emailing and it already has a multibillion dollar advertising market. News searching accounts for a major percentage of all searches. News items are available from a large number of online sources but the current technologies for news search are not scalable to effectively cover all of these sources in a timely manner. This project is to develop a new technology to tackle this problem via constructing a large-scale, highly effective, self-maintainable and customizable news metasearch engine. High effectiveness is achieved by automatically selecting the most appropriate search engines to access for each user query and by effectively identifying the correct meanings of the terms in each query. By employing highly automated techniques to incorporate search engines, this system can automatically adapt to changes that are made to the connected search engines and users can customize by adding their favorite news search engines.Highly automated solutions employed herein reduce labor costs for development and maintenance, which translate to lower advertising costs and make online advertising more affordable for "small players", including small, local media Websites, individuals and small companies. This project advances large-scale information integration, large-scale distributed information retrieval, information extraction, automatic system self-maintenance, and customization on demand. The proposed technology empowers ordinary users in their search for more relevant and more up-to-date news items from a large number of news sources. It also empowers them to customize the search system to suit their information needs.