Collaborative Project: StandardConnection--Mapping NSDL Educational Objects to Content Standards

本研究是与锡拉丘兹大学（Syracuse University, Elizabeth D. Liddy，首席研究员）合作的项目。华盛顿大学信息学院和锡拉丘兹大学自然语言处理中心的研究人员在中部大陆教育与学习研究（McREL）和Achieve公司的帮助下领导了这项工作。研究人员正在开发一种自然语言处理工具（“StandardConnection”），用于自动分配内容标准和基准到国家科学，数学，工程和技术教育数字图书馆（NSDL）和网络上的其他教育资源。这些标准和基准来自McREL制定的《标准和基准汇编》和“实现标准数据库”。作为一般描述性元数据的补充，StandardConnection工具生成的内容标准元数据将使任何州的教师都可以使用NSDL找到合适的教学资源，帮助学生实现该州设定的特定能力。该项目需要获取用于分析的教育资源的“培训”和“测试”集合；培养对人工为资源分配内容标准元数据标签的人类认知过程的复杂理解；在此基础上设计和开发技术；在一组看不见的数据上运行StandardConnection工具；通过迭代分析结果并调整工具，直到产生高度可靠的标签；并聘请一组教师专家来分析该工具将资源映射到标准和基准的质量。本项目是对编号为0085837和0085838的“打破元数据生成瓶颈”项目的合理延伸，在该项目中，研究人员正在处理教育资源的文本，以自动为教育资源的描述和主题方面分配教育材料门户（GEM）元标签。

This is a collaborative project with Award No. 0121543 (Syracuse University; Elizabeth D. Liddy, Principal Investigator). Researchers at the University of Washington's Information School and Syracuse University's Center for Natural Language Processing are leading this effort, with assistance from Mid-continent Research for Education and Learning (McREL) and Achieve, Inc.The investigators are developing a natural language processing tool ("StandardConnection") for the automatic assignment of content standards and benchmarks to educational resources in the collections of the National Science, Mathematics, Engineering, and Technology Education Digital Library (NSDL) and to other educational resources on the Web. The standards and benchmarks come from the Compendium of Standards and Benchmarks developed by McREL, and the Achieve Standards Database. Supplementing general descriptive metadata, the content standards metadata generated by the StandardConnection tool will make it possible for a teacher in any state to use the NSDL to locate appropriate teaching resources for helping students achieve a particular competency set by the state.The project entails acquiring a "training" and "testing" collection of educational resources for analysis; cultivating a sophisticated level of understanding of the human cognitive processes involved in manually assigning content standard metadata tags to resources; designing and developing the technology based on this understanding; running the StandardConnection tool on an unseen set of data; analyzing the results and adjusting the tool, through iterations, until a highly reliable tagging is produced; and employing a group of teacher-experts to analyze the quality of the tool's mappings of resources to standards and benchmarks.This project constitutes a logical extension of work conducted under Award Nos. 0085837 and 0085838, "Breaking the Metadata Generation Bottleneck," in which the investigators are processing the text of educational resources to automatically assign Gateway to Educational Materials (GEM) metatags for the descriptive and subject aspects of educational resources.