Fine-Grained Semantic Markup of Descriptive Data for Knowledge Applications in Biodiversity Domains

该奖项是根据2009年美国复苏和再投资法案(公法111-5)资助的。亚利桑那大学被授予一项赠款，用于开发和评估一套算法/软件，以帮助计算机阅读和理解？对植物、动物和其他活的或化石的生物的分类描述。算法/软件的主要功能包括：1)以机器可读的方式注释大量文本描述，以支持各种知识应用，包括为各种分类群生成字符矩阵和识别关键字。该算法不需要大量的训练实例或手工制作的正则表达式或语法规则；这使得该算法易于与不同的文档类型和/或主题一起使用；2)在最少的人工干预下为各种分类群生成交互式标识关键字；3)发现和验证描述中所表示的本体概念和关系的使用，以提高领域本体的覆盖率和文献性；以及4)产生高质量、可重用性的资源(例如词典、基准等)。用于未来的生物信息学研究。这项工作的更广泛影响包括：支持以社区为基础的信息学活动(如本体开发、文件注释和检索)；该项目制作的所有软件以及文件和培训材料都将开放源码；可重复使用的知识实体(词汇、词典等)。由该项目生成的文档将可以通过Web服务方便地访问；带注释的文档集合将可供研究、教育和公众访问。此外，将开发基于网络的教育模块，以激发K-12学生对分类学的兴趣，并向他们介绍收集、观察、描述和识别蚂蚁标本的科学方法，最后，将利用该项目提供的软件和文件集，为对数字信息管理感兴趣的研究生水平的学生创造培训机会和指导计划。有关该项目的更多信息，请访问国际刑警组织的网站：http://sirls.arizona.edu/cui.

This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5).The University of Arizona is awarded a grant to develop and evaluate a set of algorithms/software to help computers to read and ?understand? taxonomic descriptions of plants, animals, and other living or fossil organisms. The major functions of the algorithms/software include 1) annotate large sets of text descriptions in a machine-readable way to support various knowledge applications, including producing character matrices and identification keys for various taxon groups. The algorithms do not need large sets of training examples or hand-crafted regular expression or grammar rules; this makes the algorithms easy to use with different document types and/or subjects; 2) generate interactive identification keys for various taxon groups with minimal human intervention; 3) discover and verify the use of ontological concepts and relationships as represented in the descriptions to improve coverage and literary warrant of domain ontologies; and 4) produce high quality, reusability resources (e.g. lexicons, benchmarks, etc.) for future bioinformatics research. The broader impacts of this work include: support community-based informatics activities (e.g. ontology development, document annotation and retrieval); all software produced by the project will be made open source, along with documentation and training materials; reusable knowledge entities (glossaries, lexicons etc.) generated by the project will be conveniently accessible via Web services; annotated document collections will be accessible for research, education, and the public. In addition, Web-based education modules will be developed to spark interests of K-12 students in taxonomy and introduce them to scientific ways of collecting, observing, describing, and identifying ant specimens, and lastly, training opportunities and a mentoring program for graduate level students interested in digital information management will be created, using the software and document collections provided by the project. Further information about this project may be found at the PI's website at http://sirls.arizona.edu/cui.