Language Technologies for Lifelong Learning

1,010 views

Published on

Adriana Berlanga, Gaston Burek, Fridolin Wild

Published in: Education, Technology
0 Comments
1 Like
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total views
1,010
On SlideShare
0
From Embeds
0
Number of Embeds
1
Actions
Shares
0
Downloads
14
Comments
0
Likes
1
Embeds 0
No embeds

No notes for slide

Language Technologies for Lifelong Learning

  1. 1. Language Technologies for Lifelong Learning
  2. 2. Lifelong Learning & Language Technologies Adriana Berlanga Open University, NL Gaston Burek University of Tubingen, DE Fridolin Wild Open University, UK European Summer School on Technology Enhanced Learning Terchova, Slovakia. June, 2009
  3. 3. Outline • Lifelong learning • Critical supporting activities • Language Technologies • LTfLL Project: Language Technologies for Lifelong Learning • LTfLL: Positioning of the learner in a domain • Questions • Discussion 3
  4. 4. Lifelong Learning 4
  5. 5. Critical support activities • Assessment – Formative feedback • Answering questions – Routing questions – Formulating personalised answer • Monitoring progress – Drop out prevention • Supporting groups and communities – Selecting and creating groups – Providing overviews 5 Van Rosmalen et al. (2008)
  6. 6. Critical support activities 6
  7. 7. Language Technologies Question mark photo by Leo Reynolds. Licensed under Creative Commons. 7
  8. 8. Natural Language Processing (NLP) Natural Language = Human speech Analyzes spoken or written language Discover terms, generate text, translation, synthesis… – Computational Linguistics – Artificial Intelligence  Language Technologies 8
  9. 9. Working Principle 9 Landauer (2007)
  10. 10. Latent Semantic Analysis (LSA) Input (e.g., documents) term = feature vocabulary = ordered set of features Only the red terms appear in more than one document, so strip the rest. TEXTMATRIX {M}= 10 Deerwester, Dumais, Furnas, Landauer, and Harshman (1990)
  11. 11. How to support these activities in a (semi-) automatic way? 11
  12. 12. Language Technologies for Lifelong Learning
  13. 13. LTfLL – Objective To create a set of next-generation support and advice services that will enhance individual and collaborative building of competences and knowledge creation in educational as well as organizational settings. The project makes extensive use of language technologies and cognitive models in the services. 13
  14. 14. LTfLL - Themes Theme 2 support feedback services Theme 1 Theme 3 position of social and the learner informal in a domain learning LTfLL 14
  15. 15. Positioning • To determine in a (semi-)automatic way learner’s prior knowledge –by analyzing her Portfolio and the domain of study– to recommend learning materials or courses to follow. • To provide formative feedback with regard to the learner’s profile in the domain of study and recommend remedial actions to overcome conceptual gaps. Learner Support & Feedback services • To offer recommendations based on an analysis of interactions in collaborative learning using chats and discussion forums. • To offer recommendations based on the analysis of textual outputs by the learner. Supporting social & informal learning services • To provide recommendations on the basis of the learner’s profile, interests, preferences, network and learning task. This requires implementing a Common Semantic Framework (i.e. an ontology). • To provide a list of search results prioritized and categorized according to the conditions specified by the learner, and the 15 opinions of the learner’s trusted network of contacts.
  16. 16. LTfLL - Themes Theme 2 support feedback services Theme 1 Theme 3 position of social and the learner informal in a domain learning LTfLL 16
  17. 17. Positioning Question mark photo by Leo Reynolds. Licensed under Creative Commons. 17
  18. 18. Positioning • Automatically determine learner’s knowledge in a domain, given the chosen learning goal 18
  19. 19. Positioning To determine in a (semi-) To provide formative feedback automatic way learner’s prior with regard to the learner’s profile knowledge –by analyzing her in the domain of study and Portfolio and the domain of study– recommend remedial actions to to recommend learning overcome conceptual gaps materials or courses to follow Provide formative Locate best suitable feedback and learning materials or recommend courses to follow remedial actions 19
  20. 20. Positioning - Adriana Berlanga & Fridolin Wild FORMATIVE FEEDBACK 20
  21. 21. Formative feedback • Services will offer semi-automatic measurement of conceptual development • Diagnosing conceptual development – Person’s knowledge of a domain by looking on how s/he organizes the concepts of such domain – Novice vs. expert approach 21
  22. 22. The approach: Novice vs. Expert Novices and experts differ in • How they express the concepts underlying a domain • How they discriminate relevant from non- relevant information • And how they use and relate the concepts to one another 22
  23. 23. Exploring the approach SHOWCASE 24
  24. 24. Procedure 1. Ask learners what they know 2. Find out what learners should know “Expert-reference model” 3. Compare 4. Provide feedback and recommendations 25
  25. 25. 1. Ask learners what they know A think-aloud protocol was designed. Learners must explain the relevant concepts in a domain and how they are related Examples • Medical students (U Manchester) had a think- aloud session to explain concepts they just studied • Researchers (OUNL) had a think-aloud session to explain concepts they develop in their work  The think-aloud audio is transcribed: text 26
  26. 26. 2. Find out what learners should know “Expert-reference model” 1. Theoretical expert model, documents of a particular course (e.g., course material, tutor notes, presentations) 2. Archetypical expert model, state-of the art information (e.g., scientific literature) 3. Emerging expert model, concepts and the relations a group of people (co-workers, peers...) use to describe a domain 3. Compare 27
  27. 27. 2. Find out what learners should know “Expert-reference model”= Theoretical expert model Learner Expert Generation of expert and student concept maps Leximancer 28
  28. 28. 2. Find out what learners should know “Expert-reference model”= Archetypical expert model Expert map generated from LSA 29 scientific literature
  29. 29. 2. Find out what learners should know “Expert-reference model”= Emergent expert model Leximancer 30
  30. 30. 4. Provide feedback and recommendations  These are the concepts you mentioned the most  From your peers these are the most mentioned concepts  The differences are:  This means that you might find useful to • Read this material • Do this activity • Contact this person • …  Observations
  31. 31. Positioning To determine in a (semi-) To provide formative feedback automatic way learner’s prior with regard to the learner’s profile knowledge –by analyzing her in the domain of study and Portfolio and the domain of study– recommend remedial actions to to recommend learning overcome conceptual gaps materials or courses to follow Provide formative Locate best suitable feedback and learning materials or recommend courses to follow remedial actions 32
  32. 32. Positioning – Gaston Burek LOCATE BEST SUITABLE LEARNING MATERIALS 33
  33. 33. Positioning Principles Learner Objectives Evidence Materials Automatic Assessors indicated indicated indicated by recommendations decision by learner by learner assessors John S. Topics 1 Self Assessment Course No automatic Accept documents yes/no Topics 2 Certificates Course No automatic Accept documents yes/no Topics 3 Portfolio Materials LSA or Accept Documents or docs knowledge rich yes/no accepted/ approach rejected Topics 4 Formal Assessment LSA or Accept assessment samples knowledge rich yes/no accepted/ approach rejected 34
  34. 34. Objectives  A plan for extending language technologies for positioning • Implementing a showcase and scenarios to investigate and validate the approaches and give input to the design of the first release of the positioning (Wild, Hoisl and Burek, 2009) • To implement a positioning service that assess learner competences and recommend a sequence of learning material according to learning goals. 35
  35. 35. Scope of the analysis (text) • Learning history captured in the learner portfolio can be documented in different formats including text • Texts (learning materials, learner produced docs, etc.) may use long linguistic expressions to describe high level conceptualisations and terms for more precise descriptions • Curricula contain high level descriptions of learning activity materials and goals • Reduced number of samples generate sparse linguistic data • Learning materials and e-portfolios can be in different 36 languages (future work)
  36. 36. Ongoing Analysis • Finding prototypes • Students discussions on safe prescribing: Classified according expected learning outcomes related subtopics topics (A, B, C, D, E, F) Graded by experts (poor, fair, good, excelent) Methodologies • LSA • KNN • Permutation tests 37
  37. 37. Scenario Learning goal: specified by job descriptions Stakeholders objectives: • Unemployed individual: to acquire qualifications needed to find a job according to available vacancies • Individual currently employed: to acquire the competences to be useful for his company and reduce the risk of loosing his/her job • Educational provider: delivering high quality personalised technical education • Employment Governmental agency: to reduce unemployment • Enterprises: to develop employees competences for them to be capable to undertake duties required within the organisation according to predefined expectations 38
  38. 38. uestions? Question mark photo by bureau120. Licensed under Creative Commons.
  39. 39. DISCUSSION 40
  40. 40. Guiding questions Theme 2 support feedback services Theme 1 Theme 3 position of social and the learner informal in a domain learning 1. How can you validate services for these areas? 2. Big challenges? LTfLL 3. Other experiences? 41 4. Who is already working with NLP technologies for TEL?
  41. 41. Workshop A NLP-based experiments in TEL Today @ 16.15 42
  42. 42. Adriana Berlanga adriana.berlanga@ou.nl Gaston Burek gaston.burek@googlemail.com Fridolin Wild f.wild@open.ac.uk www.ltfll-project.org DSpace dspace.ou.nl/simple-search?query=LtfLL 43
  43. 43. References • Berlanga, A. J., Kalz, M., Stoyanov, S., Van Rosmalen, P., Smithies, A., & Braidman, I. (2009). Using Language Technologies to Diagnose Learner´s Conceptual Development. In Proceedings of the 9th IEEE Int. Conference on Advanced Learning Technologies. IEEE, in press. • Deerwester, S., Dumais, S.T., Landauer, T., Furnas, G.W., & Harshman, R.A. (1990): Indexing by Latent Semantic Analysis. In: Journal of the Society for Information Science, 41(6):391-407 • Landauer, T. (2007). LSA as a Theory of Meaning. Handbook of Latent Semantic Analysis, Mahaw: Lawrence Erlbaum Associates, 3-35. • Van Rosmalen, P. Sloep, P.B., Kester, L., Brouns, F., De Croock, Pannekeet, K., & Koper, R. (2008). A learner support model based on peer tutor selection. Journal of Computer Assisted Learning, 24(1), 74-86. • Wild, F., Hoisl, B., & Burek, G. (2009): Positioning for Conceptual Development using Latent Semantic Analysis. In: Basili, Pennacchiotti (Eds.): Proceedings of the EACL 2009 Workshop on GEMS: GEometical Models of Natural Language Semantics, Association for Computational Linguistics, 41–48. 44

×