SBIR Phase I: Automated Reusability Analysis of Digital Learning Resources

这项小企业创新研究（SBIR）第一阶段研究项目将调查支持自动化分析和改进数字学习资源可重用性特征所需的技术创新。数字学习资源的可重用性对于增加在线培训和教育的可访问性和降低成本至关重要。最近在可重用性理论、元数据管理和自然语言处理方面的进展使得开发分析和提高可重用性的软件变得可行。这种软件解决更深层次的可重用性问题的能力取决于两个关键研究问题的满意解决：(1)新兴的自动元数据生成（AMG）技术（包括潜在语义分析（LSA）和存储库收集技术）用于为学习内容生成准确的上下文元数据的能力有多好，包括使用学习目标的分类法进行分类？(2)是否有可能自动识别与重用相关的学习资源的语义和结构设计特征？这包括将资源分解为具有单一学习目标的自包含学习对象的能力。阶段1将测试并提供这些技术概念的证明，并将确定如何有效地将自动化可重用性分析集成到学习内容开发工作流和学习内容管理技术中。由此产生的技术将显著提高自动化数字资源分析、元数据生成和权限管理的水平，并将适用于所有类型的网络可交付内容，而不仅仅是学习资源。工业、政府和教育组织正在大力投资于数字学习资源以及用于管理和传播这些资源的网站、存储库和门户。他们希望通过提供获得高质量个性化学习的便利途径，同时降低获取、制作和维护学习材料的成本，从而提高培训和教育的有效性。实现这些目标需要具有良好可重用性特征的资源，即内容开发人员、学习者和讲师可以轻松找到并重用的资源，以响应特定的学习或教学需求。如果第一阶段是成功的，那么自动化的可重用性分析软件将在第二阶段开发。该软件将生成结构化的可重用性报告卡，为可重用性改进提出建议，并在配置时采取纠正措施。它的开发和设计将集成到内容开发、获取、联合和部署工作流程中，并将消除可重用学习资源的可伸缩和实际开发的一些主要障碍。对自动化可重用性分析的需求是直接的，并且代表了一个重要的商业机会。来自企业培训部门和教育数字图书馆的反馈表明，如果这项技术现在可用，他们会使用它。

This Small Business Innovation Research (SBIR)Phase I research project will investigate technology innovations needed to support the automated analysis and improvement of the reusability characteristics of digital learning resources. Reusability of digital learning resources is crucial to increasing access and lowering the cost of online training and education. Recent advances in the theory of reusability, metadata management and natural language processing make it plausible to develop software that analyzes and improves reusability. The ability of such software to address deeper reusability issues depends on the satisfactory resolution of two key research questions: (1) How well can emerging automated metadata generation (AMG) techniques (including latent semantic analysis (LSA) and repository harvesting techniques) be used to generate accurate contextual metadata for learning content, including classifications using taxonomies of learning objectives? (2) Is it possible to automatically recognize semantic and structural design characteristics of learning resource that are germane to reuse? These include the ability to break a resource into self-contained learning objects with single learning objectives. Phase I will test and provide proofs of concept of these techniques and will identify how to effectively integrate automated reusability analysis into learning content development workflows and learning content management technologies. The resulting techniques will significantly advance the state of automated digital resource analysis, metadata generation and rights management and will apply to all types of Web-deliverable content, not just learning resources. Industry, government and educational organizations are investing heavily in digital learning resources and in Web sites, repositories and portals for managing and disseminating these resources. They wish to improve training and educational effectiveness by providing easy access to high quality personalized learning and at the same time to lower the cost of acquiring, producing and maintaining learning materials. Achieving these goals requires resources with good reusability characteristics, i.e., resources that content developers, learners and instructors can easily find and reuse in response to specific learning or instructional needs. If Phase I is successful, then automated reusability analysis software will be developed in Phase 2. This software will produce structured reusability report cards, make recommendations for reusability improvement, and take corrective actions when configured to do so. It will be developed and designed for integration into content development, acquisition, syndication and deployment workflows and will remove some of the chief barriers to the scalable and practical development of reusable learning resources. The need for automated reusability analysis is immediate and represents a significant commercial opportunity. Feedback from corporate training departments and educational digital libraries indicates that they would use the technology if it were available today.