Structured Data & the Future of Educational Material
Download as PPTX, PDF2 likes1,202 views
Structured data and linked open data standards can help improve educational materials and enable personalized learning. By tagging educational content with metadata about learning objectives, concepts, and relationships, recommender systems can suggest personalized learning paths and assessments. Elsevier is applying these approaches to nursing education by building structured learning objects, mapping them to a taxonomy of student learning objectives, and using recommender algorithms to sequence content and questions for individual students. Linking educational and research content through metadata also allows discovering new ways for academic research to support education.
Structured Data & the Future of Educational Material
- 1. STRUCTURED DATA & THE FUTURE OF EDUCATIONAL MATERIAL Paul Groth | @pgroth | pgroth.com Disruptive Technology Director Elsevier Labs | @elsevierlabs September 28, 2016 - Onderwijs en Linked Data – wat waarom en hoe Thanks to: Mike Lauruhn, Mary Millar , David Kuilman
- 2. ELSEVIER LABS - INTRO WORLD LEADER IN DIGITAL INFO SOLUTIONS 2 Published over 330,000 articles in 2013 Founded over 130 years ago Work with over 30 million Scientists, students, health & information professionals Employ over 7,000 employees in 24 countries Received over 1 million submissions in 2013 Over the last 50 years the majority of Noble Laureates have published with Elsevier Over 53 million items indexed by Scopus Elsevier eBooks, Online Journals, Databases Publishes over 2,200 online journals & over 10,000 e-books SOLUTIONS Elsevier R+D Solutions Elsevier Clinical Solutions Helps corporate researchers, R+D professionals, and engineers improve how they interact with, share, and apply information to solve problems using our digital workflow tools, analytics, and data Provides universities, governments, and research institutions with the resources and insights to improve institutional research strategy, management, and performance. Elsevier Education Helps medical professionals apply trusted data and sophisticated tools to make better clinical decisions, deliver better care, and produce better healthcare outcomes. Helps educate highly-skilled, effective healthcare professionals, using the most advanced pedagogical tools and reference works. Elsevier Research Intelligence CONTENT CAPABILITIESPLATFORMS
- 6. 40 million reactions 75 million compounds 500 million facts
- 7. HOW DO WE ENABLE LEARNERS TO ACQUIRE AND APPLY ALL THIS KNOWLEDGE EFFECTIVELY?
- 8. THREE PROJECTS 1. Nursing education 2. Linked data for learning about linked data 3. Using research resources to facilitate education
- 9. http://www.theatlantic.com/health/archive/2016/02/nursing-shortage/459741/ According to the Bureau of Labor Statistics, 1.2 million vacancies will emerge for registered nurses between 2014 and 2022. By 2025, the shortfall is expected to be “more than twice as large as any nurse shortage experienced since the introduction of Medicare and Medicaid in the mid-1960s,” a team of Vanderbilt University nursing researchers wrote…
- 11. Product Design (Student facing) 11 PROBLEMS ADDRESSED I have too much content to consume. I have limited time to study, so I need to know where to concentrate. 1. PERSONALIZED LEARNING TUTOR Reminders and prioritization assistance to help students optimize available study time and balance test prep with assignment completion. 3. QUIZZING COACH Test prep featuring content recommended based on upcoming goals, time-available, and current areas of weakness. 2. ENGAGING LEARNING EXPERIENCE Media-rich learning instructionally-designed to convey essential information through bite-sized content, quizzes, activities, and feedback.
- 12. WHAT IS A RECOMMENDER? 12 For most businesses, recommenders are essential for surfacing items to consumers. For education, recommenders enable adaptive and enhanced learning. Recommendation engines come in different types: content-based, knowledge-based, collaborative-filtering, and hybrids. Content-based looks for items that users view or purchase together Knowledge-based applies metadata to make inferences about the similarity of items Collaborative-filtering looks for similarity among users to make item recommendations Hybrid systems combine capabilities of different types and reconcile the recommendations from each Recommenders are tools for interacting with large and complex information spaces. They provide a personalized and prioritized view of items in the space. Rec Tech Content + User Data
- 13. Recommender System MACHINE LEARNING RECOMMENDER COMPOSITION OUR RECOMMENDER COMPOSITION 13 Our Recommender System is composed of the recommendation technology (engines, analytics, search, etc.) and content & user data. To make effective recommendations for our use cases, the system needs to be built for the content & user data in our domain. In our own system, technology is built to match the content. Learning Content Bite-Sized,Tagged Reference Structured,XMLContent & User Data Ontologies E.g.,EMMeT SLOs Taxonomyofstudent learningobjectives UserData Storage(datalake) ENGINES NLP, Analysis, Modeling SEARCH Semantic, Filtering ANALYTICS Collection, Processing INTEGRATION APIs, Connections
- 14. Discovery using Recommender algorithms 14 Learning Object Student Learning Objective Student Learning Objective Student Learning Objective edu:teaches edu:requires Bloom concept skos:narrowerMatch skos:broader Content Object ecm:hasRecommendation edu:remediates Question Object edu:assesses edu:teaches ecm:hasRecommendedAssessment Student Learning Objectiveskos:narrower QSEN concept skos:broaderMatch Recommender algorithms can help: - Identify test banks - Alternate content in books - Alternate content for learning objective Manual, expert curation can help: - Set questions to objectives - Set remediation content for a learning object - Set alternative content to learning objective
- 15. QUIZ COACH EXAMPLE QUIZ COACH WORKED EXAMPLE 15 User Flow 1. Student chooses to take 15 questions on their Weak Areas, covering Chain of Infection 2. Quiz Coach service generates a quiz using the Recommender Recommender fills SLO “slots” from a pool of responses (questions) for each SLO. Constraints: 1. Time – Total time available (selected), time per question 2. Depth – Prioritizes breadth over depth 3. Novelty – Unseen, Incorrect, Skipped/Delivered/Unanswered, Correct 4. Difficulty – increasing order or difficulty range 5. Content preference – source of questions
- 16. HOW SHERPATH USES CONTENT 16 Product Content Learning Content is purpose- built to fit our content model (individual learning objects that teach or assess learning objectives) and tagged to SLOs. Our SLOs are a new unified taxonomy of domain learning outcomes that map didactic and clinical objectives. Reference content is available as structured XML for each of our 53 core Nursing titles EMMeT is used to text mine the XML for relevant health concepts to save manual tagging In EOLS, new and legacy content are integrated on every page. Users consume content designed for digital learning, but can also request “More Details” straight from ELS book content when they’re confused. Rec API
- 17. CONTENT MODEL COMPONENTS 17 • Elsevier Enterprise Content Model ontology • 40+ properties • 20 datatypes • 10 Content types • 20 Asset types • Adaptive Learning ontology • Recommendation • Teaching • Assessing • Remediation • SKOS ontology • 3 third-party vocabularies: QSEN, Bloom etc. • QTI 2.1 compliant schema • XHTML5 schema • 50+ data-type attribute definitions • Student Learning Objective ontology • SKOS ontology extended with 2 properties • Multi-media assets incl. Text Time based Markup Language
- 18. THE ROLE OF STRUCTURED DATA • Content needs be organized/tagged at a fine grained level • Content organization helps drive recommendation • Allow for the recommendation of specific pieces of content • Allow for the combination of multiple pieces of content to support a student’s journey 18
- 20. Learner Transit Map courtesy Stuart Sutton
- 24. THE ROLE OF STRUCTURED DATA • Central aggregation enables the location of material • Structured data eases aggregation especially if produced by multiple providers • Structure enables a separation between defining the learning goals/objectives and the variety of content that can support that 24
- 25. What’s the role of academic / research content in education?
- 27. Challenge: how can we use these resources to improve education?
- 28. 28/09/16LinkedUp – Michael Lauruhn 28 Syllabus Discover Application: • Prototype allows user to make connections between a syllabus and a corpus of journal article content • Live demo (with LinkedUp Catalogue Water & Ecology content) available online • Configurable code available on Github
- 29. Clustering & Visualisation Discover Application: • Prototype provides user a visualisation of concepts from a corpus of journal article content based on clustering • Live demo (with LinkedUp Catalogue Water & Ecology content) available online • Configurable code available on Github
- 30. THE ROLE OF STRUCTURED DATA • Establishing structure enables new applications to be built • Allows rethinking the role of existing content with respect to educational apps
- 31. CONCLUSION • Education increasingly needs to be personalized, adaptive, on-demand and current • Technology is central • But technology requires the right kinds of content (bite size, reusable, malleable) • Linked data standards and methods are an important mechanism for delivery Contact: Paul Groth @pgroth
Editor's Notes
- #3: Elsevier Labs in context
- #6: 40 million reactions 75 million compounds 500 million experimental facts ,
- #7: 40 million reactions 75 million compounds 500 million experimental facts ,
- #11: Sherpath is one of Elsevier’s flagship products. It is a solution designed for nursing students and instructors It provides instructionally designed pedagogical content using adaptive and personalization techniques to improve learning experience and outcomes. It includes collection, analysis and reporting based on student and instructor usage and activities. I will be using our Quiz coach feature as an example throughout my presentation today
- #17: Recommender system = tech + content Complication – we don’t have it yet Resolution – Win because world-class content
- #18: Use of open standards
- #26: Elsevier is one of the partners
- #28: Content from new Open Access Water Resources & Industry Contetn from Elsevier Open Achive Scientific datasets from Water Footprint Network