Adaptive Sequencing and Course          Generation        Speaker: Wenkai Dai       Tutor: George Goguadze                ...
Motivation             2
Motivation             3
Outline• Motivation• Introduction• Concepts• Pedagogical Objectives• Operator & Method• Dynamic Task• Converting a Plan in...
Introduction• Middle way between above studying ways• Course generator• Content organized by pedagogical principles• Probl...
Introduction• PAIGOS, web-based course generator on math study• Using novel techniques from  • Semantic web  • Artificial ...
IntroductionPedagogical knowledge       Learning goal         Learner’s competencies                        Personally str...
Screenshot of LeActiveMath                             8
Screenshot of LeActiveMath                             9
Course Generation                    10
Concepts• Hierarchical Task Network Planning• An HTN planning problem consists of  • An initial state  • Logical atoms  • ...
Pedagogical Objectives• t = (discover, (def_slope, def_diff))                                          12
General Form of Planning Problem                                   13
Operators and Methods• An assignment expression (assign ?var t) binds ?var to  the term t.• Basic operator (:operator (!in...
Dynamic Tasks• Dilemma: The sequence of contents leading to his goal  has been structured in advance but the assumptions  ...
Dynamic Tasks• Extended the planner in such a way that planning may  stop at the level of specially marked tasks (dynamic ...
Dynamic Items• Dynamic tasks are simulated by (:operator  (!dynamicTask ?ped Obj ?refs) ()()())• No precondition for the O...
Converting a Plan into a Course• After plan found, generate a course, a table of  contents• PAIGOS represents courses usin...
Selecting Exercises & Examples                                 19
Scenario of LearnNew                       20
Generated Course of LearnNew                               21
PerformanceRequired time of course generation vs. increasingamount of concepts                                            ...
PerformanceA plot of the number of concepts vs. time required forcourse generation in milliseconds                        23
Discussion• Including more factors like motivation, meta cognitive  adaptation                                            ...
Question, Thanks
Upcoming SlideShare
Loading in …5
×

E learning

329 views

Published on

wenkai dai

Published in: Education
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
329
On SlideShare
0
From Embeds
0
Number of Embeds
2
Actions
Shares
0
Downloads
3
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

E learning

  1. 1. Adaptive Sequencing and Course Generation Speaker: Wenkai Dai Tutor: George Goguadze 1
  2. 2. Motivation 2
  3. 3. Motivation 3
  4. 4. Outline• Motivation• Introduction• Concepts• Pedagogical Objectives• Operator & Method• Dynamic Task• Converting a Plan into a Course• Performance & Discussion 4
  5. 5. Introduction• Middle way between above studying ways• Course generator• Content organized by pedagogical principles• Problems of existing course generators 5
  6. 6. Introduction• PAIGOS, web-based course generator on math study• Using novel techniques from • Semantic web • Artificial intelligence • Technology-enhanced learning 6
  7. 7. IntroductionPedagogical knowledge Learning goal Learner’s competencies Personally structured courses 7
  8. 8. Screenshot of LeActiveMath 8
  9. 9. Screenshot of LeActiveMath 9
  10. 10. Course Generation 10
  11. 11. Concepts• Hierarchical Task Network Planning• An HTN planning problem consists of • An initial state • Logical atoms • The initial task network • A domain description– Planning operators: (:operator h P D A)– Methods: (:method h L1 T1 L2 T2 . . .Ln Tn) 11
  12. 12. Pedagogical Objectives• t = (discover, (def_slope, def_diff)) 12
  13. 13. General Form of Planning Problem 13
  14. 14. Operators and Methods• An assignment expression (assign ?var t) binds ?var to the term t.• Basic operator (:operator (!insertResource ?r) () () ((inserted ?r))). no precondition and delete list. It adds a logical atom to the world state that describes that a resource ?r was inserted into the course• Insert method 14
  15. 15. Dynamic Tasks• Dilemma: The sequence of contents leading to his goal has been structured in advance but the assumptions about the learner become invalid in real time 15
  16. 16. Dynamic Tasks• Extended the planner in such a way that planning may stop at the level of specially marked tasks (dynamic tasks)• These tasks are inserted into the course like any other reference to a learning object• When the learner first visits a page that contains a dynamic task, this task is passed on to the course generator• Dynamic subtask is not natively supported• Solution: Dynamic Items 16
  17. 17. Dynamic Items• Dynamic tasks are simulated by (:operator (!dynamicTask ?ped Obj ?refs) ()()())• No precondition for the Operator to perform this task• Operator creates a special element called dynamic item when applied• Latter when course is presented, dynamic item on the page will pass the associated dynamic task to the course generator• By this way, some contents of the next page to be viewed are inserted dynamically 17
  18. 18. Converting a Plan into a Course• After plan found, generate a course, a table of contents• PAIGOS represents courses using the element omgroup from OMDoc Standard, a semantic knowledge representation for math documents• But omgroup is independent of math domain, which can be easily map to other data structures• Omgroup consist of metadata information like author and title of the element, references to other OMDoc elements, omgroup elements and dynamic items. 18
  19. 19. Selecting Exercises & Examples 19
  20. 20. Scenario of LearnNew 20
  21. 21. Generated Course of LearnNew 21
  22. 22. PerformanceRequired time of course generation vs. increasingamount of concepts 22
  23. 23. PerformanceA plot of the number of concepts vs. time required forcourse generation in milliseconds 23
  24. 24. Discussion• Including more factors like motivation, meta cognitive adaptation 24
  25. 25. Question, Thanks

×