et 805 ats domain module ii
play

ET-805 ATS Domain Module - II Ramkumar.Rajendran@iitb.ac.in From - PowerPoint PPT Presentation

ET-805 ATS Domain Module - II Ramkumar.Rajendran@iitb.ac.in From Last Class Blackbox Domain Model Glassbox Domain Model Muddy Points - Structure of MYCIN - Differences in interaction of SOPHIE and GUIDON 2 Activity - Each one


  1. ET-805 ATS Domain Module - II Ramkumar.Rajendran@iitb.ac.in

  2. From Last Class Blackbox Domain Model ● Glassbox Domain Model ● Muddy Points - Structure of MYCIN - Differences in interaction of SOPHIE and GUIDON 2

  3. Activity - Each one say one What is the knowledge base or structure of Domain Model in MYCIN Share to the class Class Response - Knowledge tree - some condition - Limited to set of rules given by experts 3

  4. MYCIN Domain Model - If then rule: Production rules - Knowledge as set of goals and hypothesis Interactions in MYCIN: - Generated from knowledge base that is developed by Experts 4

  5. GUIDON Sample Interaction Student: HOW IS THE FACT THAT THE PATIENT IS BURNED RELEVANT? {One option is to ask how a piece of data is relevant to the current goal.} GUIDON: The fact that patient-326 is seriously burned will be used to conclude that Pseudomonas aeruginosa is one of the organisms which might be present (in addition to those seen on cultures or smears) [Rule 578]. Student: HELP {The "help" option is a request for assistance: the student asks "where do I go from here?"} Try to determine the type of the infection: bacterial, fungal, viral, or TB.

  6. Activity - Each one say one Think: What is the difference between interaction of SOPHIE and GUIDON! List down your answers (2 minutes) Share to the class (5 minutes) Class Response: - SOPHIE don’t provide response to What and Why - SOPHIE lacks Production Rules - SOPHIE cannot provide help 6

  7. SOPHIE and GUIDON Questions in SOPHIE - From expert knowledge base: What questions the experts will ask to solve the problem. - Are the terminals open? - Which terminals are open? GUIDON - Generates responses to students or helps them based on knowledge base that are developed by experts.

  8. Activity - Motivating Rule-based Model Think: Consider single digit fraction addition for example 3/4 + 9/8. What are the steps involved to solve fraction addition? (2 minutes) Share to the class (7 minutes) 8

  9. Steps to Solve Fraction Addition Subgoals of Fraction addition problem: Example 3/4 + 9/8 - Fractions should have common denominator - Convert the Fractions. 8,16, 24, 32 , 40… - Add the fractions. 6 + 9 = 15/8, 60/32 - If needed, reduce the fractions to lowest possible denominator. 60/32 = 15/8 Characteristics - Subgoals should follow order - Within subgoal order of steps is not enforced Aleven, V. (2010). Rule-based cognitive modeling for intelligent tutoring systems. In Advances in intelligent tutoring systems (pp. 33- 9 62). Springer, Berlin, Heidelberg.

  10. Rule-Based Model - Each Facts or Chunks contains attributes or slots. Which can be another fact or values - Also, contains the subgoals

  11. Rules for Convert Denominator - The subgoals are generated by the rules while solving the fraction problem DETERMINE-LCD IF there are no subgoals and D1 is the denominator of the first given fraction and D2 is the denominator of the second given fraction THEN Set LCD to the least common denominator of D1 and D2 Add subgoals to convert the fractions to denominator LCD Add a subgoal to add the two converted fractions - Does not model how students determine common multiples but calls the function

  12. Rules for Fraction Addition CONVERT-DENOMINATOR IF there is a subgoal to convert fraction F so that the denominator is D And the denominator for the converted fraction has not been entered yet THEN Write D as the denominator of the converted fraction And make a note that the denominator has been taken care of CONVERT-NUMERATOR IF there is a subgoal to convert fraction F so that the denominator is D And the numerator for the converted fraction has not been entered yet And the (original) numerator of F is NUM And the (original) denominator of F is DENOM THEN Write NUM * (D / DENOM) as the numerator of the converted fraction And make a note that the numerator has been taken care of

  13. Activity - PI Consider the Fraction addition Problem ( ⅙ + ⅜ ). Suppose the Rules are implemented and the student answers 14 as Numerator. What feedback the tutor will give to the student? A. Marks Correct B. Looks likes you added denominator C. You should find a common multiple D. Marks Incorrect Vote Pair (2 mins) Vote Again

  14. Activity - PI Consider the Fraction addition Problem ( ⅙ + ⅜ ). Suppose the Rules are implemented and the student answers 9 as denominator. What feedback the tutor will give to the student? A. Marks Correct B. Looks likes you added denominator C. You should find a common multiple D. Marks Incorrect Vote Pair (2 mins) Vote Again

  15. Tutor Feedback( ⅙ + ⅜) Student 1: Enters 24 as denominator for second fraction ⅙ Tutor: Marks correct Student 2: enters 14 as numerator Tutor: “It looks like you are adding the denominators, but you need a common multiple.” Student 3: Enters 9 as a denominator Tutor: Marks Incorrect. No message

  16. Activity - TPS Think: Write two limitations of Rule-based model? (3 minutes) Pair: Copy your neighbors answers. Together, rank all your answers in the order of most severe limitation to least. (3 minutes) Share to the class (5 minutes) Class Response: - Rule are not exhaustive - Subgoals are should be in order - It cannot account the silly mistakes 16 - Not good non-math domain

  17. Drawbacks of Rule-Based Model - Complex of developing the production set - Might not be possible to create production for domain or task that is ill-defined - Creating error database that contains what mistakes students can do is time-consuming as the space of incorrect knowledge is vast Mitrovic, A. (2010). Modeling domains and students with constraint-based modeling. In Advances in intelligent tutoring systems (pp. 63-80). Springer, Berlin, Heidelberg.

  18. Cognitive Tutor Aleven, V. (2010). Rule-based cognitive modeling for intelligent tutoring systems. In Advances in intelligent tutoring systems (pp. 33-62). Springer, Berlin, Heidelberg.

  19. Next class Assimilation Quiz Make sure you read the paper!

  20. Activity - TPS Think: You have to borrow a book from Institute Library. What are the Rules? List down 2 answers (2 mins) Pair: Together, List down constraints for the above task. (3 minutes) Share to the class (5 minutes) Class Response: - Valid member, Library Timing, maximum number of books allowed to issue, book should be available to borrow, 20

  21. Constraint Based Model - Proposed by Ohlsson in 1992 AIED conference - Captures domain principle that every correct solution should follow - Constraint is an ordered pair (Cr, Cs), where Cr is the relevance condition and Cs is the satisfaction condition - If you are driving in New Zealand, Then you better be on the left side of the road Mitrovic, A. (2010). Modeling domains and students with constraint-based modeling. In Advances in intelligent tutoring systems (pp. 63-80). Springer, Berlin, Heidelberg.

  22. Activity - TPS Think: Consider the same fraction addition problem 3/4 + 9/8. Write down the relevance conditions? (2 minutes) Pair with your Neighbour: Identify the common relevance conditions and together identify satisfaction condition(3 minutes) Share to the class (7 minutes) 22

  23. Activity - Class Response - Denominator should be equal - Denominator cannot be Zero

  24. CBM for Fraction Addition If the current problem is a/b + c/d, and the student’s solution is (a+c)/n, then it had better be the case that n=b=d

  25. SQL Tutor Constraints (Syntax)

  26. SQL Tutor - Semantic Constraints

  27. Feedback in CBM Levels of Feedback: correct/incorrect, error flag, hint, all errors, partial solution First Attempt: only correct/incorrect Second Attempt: Point where the solution is incorrect Third Attempt: Provides a messages attached to violated constraints Further levels are provided based on students’ request

  28. Activity - Each one say one Think: What is the different between rule-based model and constraint based model in ITS? Each one write down one point. (3 minutes) Share to the class Class Response - Varying level of feedback by CBM - Lot of emphasis on procedural knowledge - You cannot violate the rule not constraint 28

  29. Cognitive Models - Rule based vs Constraint Based Rule Based Model Models student thinking to solve problem and capture ● knowledge that required in generating solution step-by step Constraint Based Model - Models all the constraints and requirements that should be satisfied in generating solution

  30. In today’s class... Rule-Based Domain Models ● Constraint Based Domain Models ● 30

  31. Next class ... Learner Modeling in ITS ● Overlay Model ○ IRT ○ BKT ○

  32. Assignment 1 - Reading Assignment Pick an ITS form the below list (You can choose others too ) - Autotutor, Metatutor, Betty’s Brain, Wayang Outpost, Crystal Island, PAT, Cognitive Tutor and SQL Tutor - Describe the Domain and Learner Model of the ITS you picked Rules: - Only 10 minutes for a speaker - Report your tutor and the related paper by 21-Jan - Moodle - Presentation in Class on 28th Jan

  33. Last Activity - Muddy Points List down - two important and - two least clear (muddy) points from today’s class - https://tinyurl.com/et8 05mp 33

Recommend


More recommend