SBIR Phase II: Semantically Intelligent Knowledge Hub for Course Authoring

这个小型企业创新研究第二阶段项目旨在开发一个系统，极大地简化教育机构共享、发现、管理和管理课程内容的方式。教育机构中的绝大多数内容都位于多个孤立的存储库中。因此，合规跟踪、使课程与能力保持一致以及编写课程都是手动且昂贵的过程。这个项目旨在打开这些孤岛，使跨多个平台的主题级别的课件搜索、标记和策展变得简单。该技术还提取学生反馈和评估信息，并将其与细粒度学习目标对应起来，为跟踪和改进学习结果提供新的见解。通过解锁困在孤立的大学系统中的课件，该项目促进了一个跨机构的市场，使任何机构都能够分享和/或将精心策划的高质量内容货币化，并反过来获得可能由兼职教师教授的精心策划的课程包，这是一个价值20亿美元的市场机会。简化教师对开放教育资源和其他学院精选学习内容的访问，将有助于显著降低学生课程包成本。该项目的关键技术创新是从多个教育资源库无缝提取学习内容，并在课程管理过程中融入一个大纲驱动的语义智能搜索和推荐引擎。开发一种抽象机制来与不同类型的学习管理系统和其他存储库接口，并使其可移植，而不需要通用的内容格式，这是一项重大的创新。自然语言处理技术与精选的受控词汇表和分类法相结合，用于简化课程的精选，特别是自动主题识别和教学标记。在语义搜索引擎中利用精选数据来改进课程创作过程中的课件推荐。自适应推荐引擎的目标是提高对开放教育资源和其他大学开发的课件的发现和使用，而不需要昂贵的专业服务。该项目的另一个目标是使用IMS定义的标准无缝地提取学生反馈，无论使用哪个学习管理系统来传递内容。然后，将学生反馈、评估数据、成绩和从不同大学系统中提取的其他信息关联起来，以提供丰富的分析和学生结果的可见性，并使课程与能力保持一致。

This Small Business Innovation Research Phase II project aims to develop a system to dramatically simplify how course content is shared, discovered, curated and managed in educational institutions. The vast majority of content in educational institutions is in multiple silo-ed repositories. Hence, compliance tracking, aligning curricula with competencies and course authoring are manual and expensive processes. This project aims to unlock these silos and make it simple to search, tag and curate courseware at topic-level across multiple platforms. The technology also extracts student feedback and assessment information and maps it against granular learning objectives to provide new insights to track and improve learning outcomes. By unlocking courseware stuck in silo-ed college systems, this project facilitates a cross-institutional marketplace that enables any institution to share and/or monetize curated high quality content, and conversely have access to curated course packs that may be taught by adjunct faculty, a $2B market opportunity. Simplifying faculty access to Open Educational Resources and curated learning content from other colleges will help drive down student course-pack costs significantly.The key technical innovation in this project is seamless extraction of learning content from multiple educational repositories coupled with a syllabus-driven semantically intelligent search and recommendation engine woven into the course curation process. Developing an abstracted mechanism to interface with different types of learning management systems and other repositories, and making it portable without requiring a common content format is a significant innovation. Natural Language Processing techniques combined with curated controlled vocabularies and taxonomies are applied to simplify curation of a course, notably automatic topic identification and instructional tagging. The curation data is leveraged in the semantic search engine to improve recommendation of courseware during course authoring. The goal of the adaptive recommendation engine is to improve discovery and usage of both open educational resources and courseware developed in other colleges, without requiring expensive professional services. Another goal of the project is to extract student feedback seamlessly using IMS defined standards regardless of which learning management system is used to deliver the content. The student feedback, assessment data, grades and other information pulled from various college systems are then correlated to provide rich analytics and visibility of student outcomes and alignment of courseware with competencies.