The document discusses using an ePortfolio called the ActiveMath Workbook to capture student learning and reflection within the ActiveMath learning environment. The Workbook allows students to assemble their own books, maintain a learning log, and formulate hypotheses with annotated beliefs. Semantics are captured through the Workbook's structure, questions, and ability to relate entries to learning objects. The architecture involves a math knowledge base, learner model server, and clients connected to a web and database server. Future work includes more evaluation and enhancing collaboration and interaction with ActiveMath services.

1. ePortfolio in Martin Homik, Erica Melis -- ActiveMath Group –- www.activemath.org ePortfolio 2005, Cambridge German Research Center for Artificial Intelligence (DFKI GmbH) Saarland University

2. Outline Motivation ActiveMath Workbook (ActiveMath’s ePortfolio) Capturing Semantics in Workbook Architecture Conclusion

3. Motivation Current ePortfolios: standalone applications Integration into large systems, in general? Integration into learning environments? Definition: “… collection of evidences…” How to capture reflection on processes? How to turn implicit knowledge into explicit one? Function: store and provide access How to support users?

4. ActiveMath Web-based User-adaptive Interactive Semantic KR Service-based

5. ActiveMath Web-based User-adaptive Interactive Semantic KR Service-based

6. ePortfolio in ActiveMath (Workbook)

7. Assembly Tool Students are “book smart” Assemble own books Supports meta-cognitive skills Store locally/upload

8. Learning log Self-managed documentation of learning objectives Base on experience with paper logs Reflection on: Content Process Meta-Cognition

9. Hypotheses Formulation and justification Annotated by beliefs Supports: Reasoning Problem solving Argumentation

10. Views and Queries List entries: all/specific type/specific period … related to a specified learning object … related to LOs a learner does not understand … related to LOs a learner did not understand in first place, but has learned meanwhile. … .

11. Capturing Semantics By structure We know what we ask By multiple choice questions We can interpret answers By template answers We can interpret answers without full understanding By drag&drop We can relate entries to learning objects

12. Capturing Semantics: Structure Questions related to: General data Content Process Meta-cognition Label Privacy Date What was the lesson about? Write a summary. What did you (not) understand? Formulate questions to issues you did not understand. What kind of tools or resources did you use for solving? Who helped you? What difficulties arose during problem solving? What learning objects were most helpful to you? How would you rate your performance? Did you achieve your goals? If not, why? Who can help you?

13. Capturing Semantics: Drag&drop http://demo.activemath.org:8080/... mbase://LeAM_calculus/diffquot/def_diff_quot

14. Architecture Math Knowledge Base Learner Model Server Clients Web Server Manager retrieve JNLP (http) Learning log Database

15. Outlook Evaluation: Currently in seminar with 15 university students Planned for LeActiveMath EU-Project Investigate learning logs (types, semantics, …) Enhance collaboration Increase interaction with ActiveMath’s services Compare learning logs with learner model Integration of more reflection-supporting tools (Examine) Export in IMS ePortfolio

16. Conclusion ePortfolios in learning environments Not restricted to collection of evidences Store reflection processes Explicit reflection supports meta-cognition Capturing semantics enables provision of user-adapted support Still a lot of interesting work to do.

17. Thank you. www.activemath.org