CAREER: Evolving and Self-Managing Data Integration Systems

数据集成系统提供了对大量数据源的统一访问。它们有可能彻底改变我们访问数据的方式，并为构建更先进的信息处理架构提供基础。然而，今天这样的系统仍然非常难以建造，维护成本高昂。必须详细地告诉他们如何与数据源交互，并且必须不断修改以处理数据源的更改。为了解决这个问题，该项目设想建立数据集成系统，随着时间的推移学习发展和自我管理，最大限度地减少人为干预。为了实现这一愿景，该项目采用了数据库和人工智能，（特别是机器学习）技术来应对以下核心挑战：（a）有效地自动化关键的劳动密集型任务，包括模式匹配、全局模式创建和重复检测，（B）以最少的人为干预检测由于源处的改变而导致的系统故障，以及（c）通过将负担分散到大量用户上，进一步减轻系统管理员的巨大数据集成负担。该教育计划利用研究为学生和更广泛的社区做好准备，以应对互联网世界提出的新数据管理挑战。就智力价值而言，该项目在数据集成研究方面迈出了合乎逻辑的下一步。它为几乎所有数据集成或共享工作的基本问题带来了概念上新颖的解决方案。该项目的结果具有潜在的并行计算应用。就更广泛的影响而言，该项目将促进数据整合系统的广泛部署，从而为社会带来更有效的信息管理和获取。它在教育下一代专业工作者和研究人员方面发挥着不可或缺的作用。这项研究还将有助于整合伊利诺伊州农村消防队员的数据，并培训他们访问和使用综合信息系统。项目信息将通过出版物、研讨会、教程和网站http://www.cs.wisc.edu/~anhai/projects/career.html传播，其中将包括最终的研究结果、数据和系统工件。

Data integration systems provide a uniform access to a multitude of data sources. They have the potential to revolutionize the way we access data, and provide a basis on which to build even more advanced information processing architectures. However, today such systems are still extremely hard to build and costly to maintain. They must be told in tedious detail how to interact with data sources, and must be constantly modified to deal with changes at the sources. To address this problem, the project envisions building data integration systems that learn to evolve and self-manage over time, with minimal human intervention. To make fundamental contributions toward realizing this vision, the project employs database and artificial intelligence (especially machine learning) techniques to attack the following central challenges: (a) effectively automating key labor-intensive tasks, including schema matching, global schema creation, and duplicate detection, (b) detecting system failures due to changes at the sources, with minimal human intervention, and (c) further reducing the tremendous data integration burden of the system administrators by spreading the burden thinly over the mass of users. The education plan leverages the research to prepare students and the broader community for the novel data management challenges raised by the Internet world. In terms of intellectual merit, the project takes a next logical step in data integration research. It brings conceptually novel solutions to fundamental issues underlying virtually any data integration or sharing efforts. The project results have the potential for autonomic-computing applications. In terms of broader impacts, the project will facilitate the widespread deployment of data integration systems, thus resulting in more effective information management and access for society. It plays an integral part in educating next-generation professional workers and researchers. The research will also help integrate data for rural Illinois fire fighters, and train them in access and use of the integrated information systems. The project information will be disseminated via publications, workshops, tutorials, and the Web site http://www.cs.wisc.edu/~anhai/projects/career.html that will include the resulting research results, data and system artifacts.