Advanced Interactive Learning Environments 20034 UG4MSc

1
Advanced Interactive Learning Environments
2003/4(UG4/MSc)

• Helen Pain, HCRC/ICCS,
• Informatics, Edinburgh
• Room 7, 3rd floor left
• 2 Buccleuch Place
• 650 8485, helen_at_inf.ed.ac.uk

2
AILE Lecture 1Introduction
3
Contents

• An exercise...
• Goals, methods and domains
• What skills might we teach?
• What sorts of tools and systems?
• Course Overview

4
1. An exercise
5
Example problem subtraction

• a. 73 b. 32 c. 164 d. 187 e.
  19763
• -11 -16 - 37 - 99
  -16824
• How do you do each of them?
• What methods do you use?
• Do you use the same method for all of them?
• How did you learn to do it?
• How would you teach someone else to do it?
• What would they need to know to do so?
• What would you need to know to teach them?

6
What if their answers were

• a. 73 b. 32 c. 164 d. 187
• -11 -16 - 37 - 99
• S1 62 24 133 112
• S2 62 26 137 198
• S3 62 24 214 817
• S4 61 14 130
  89
• or no response at all....

7
Modelling learners

• trying to find out what the student knows,
  believes, can do
• looking for evidence that user fails to exploit
  some knowledge
• looking for inconsistent beliefs, differences
  between student and domain models
• teach accordingly

8
Diagnosing Student Models

• If the teacher believes a student has a different
  model from their own (correct) one
• make list of common errrors and match to it
• reason about what student would believe in order
  to exhibit behaviour indicating this
• Representation of student's current state of
  knowledge STUDENT MODEL
• Inferring the Student Model DIAGNOSIS

9
Possible Diagnoses

• a. 73 b. 32 c. 164 d. 187
• -11 -16 - 37 - 99
• S1 62 24 133 112
• take higher from lower
• S2 62 26 137 198
• give 10 but don't pay back
• S3 62 24 214 817
• work l to r, higher from lower
• S4 61 14 130
  89
• guess

10
What would you do?

• Work out patterns to error
• Hypothesize what misconceptions and why
• How would you help the student?
• - Ask child to talk through it
• - Re-teach
• - Give explanation of what you think is wrong
• - Lead child to work out what wrong
• - Give calculator
• Could you use AI to help the student?
• - Give child feedback on their errors
• - Give child explicit instruction
• - Do it for child
• - Get the child to teach the system

11
Could you use AI to help you?

• Build a tool, with errors, to test your theories?
• Build a tool to teach with?
• What would system need to know?
• - How to do subtraction
• - How to teach subtraction
• - How to diagnose errors
• - What the student knows already
• - How the student learns as they do it
• - How to talk to student
• - When to (or not to) interupt
• - How to represent all these

12
What to represent? (McCalla)
13
2. Goals, Methods and Domains
14
Helping learners to improve their learning and
communication

• through observing and analysing their
  difficulties
• modelling individual students difficulties and
  misconceptions
• designing, developing and evaluating computer
  based learning support and communication tools

15
Methods

• Include
• consulting teachers and domain experts
• investigating literature and previous research
• observing teaching interactions
• theories and models of learning and communication
• design and formative evaluation of systems
• summative evaluation of effectiveness

16
Human One-to-one Tutoring

• Human tutoring is the most effective form of
  instruction
• Tutors maintain delicate balance
• - students do as much of the work as possible
• - tutors provide just enough guidance to keep
  students from becoming frustrated or confused
• students maintain a feeling of control
• Todays intelligent tutoring systems show
  learning gains that are half that of human
  tutoring

17
Dialogue-based Learning Environments

• Intelligent Tutoring Systems are effective, but
  NOT as effective as human tutors.
• The question is why not?
• Dialogue is the key - observing, analysing and
  modelling in educational contexts
• Natural language offers indirect techniques for
• signalling disagreement or uncertainty,
  suggesting solutions, etc
• - switching topic
• - taking or relinquishing initiative

18
Edinburgh Experience in this.

• Mathematics (Logo, LeActivemaths)
• Physics (dynamics, electricity)
• Language learning and Communication
• typing, spelling, reading, story writing,
  syntax, argument, discussion, humour
• Second language learning/ESL
• Special needs (cognitive and physical)
• e.g. dyslexia, hearing, typing
• Music (harmony teaching, drumming)
• Programming (Prolog, Lisp)
• Ecological modelling, Knowledge Modelling

19
Various possible names for this area

• Intelligent Tutoring Systems
• Intelligent Learning Environments
• Adaptive Learning Environments
• Knowledge Based Learning Environments
• Educational Informatics
• Pedagogic Informatics
• Learning Sciences
• Computational Mathetics
• .......
• A.I. and Education
• Advanced Interactive Learning Environments

20
3. What skills might we teach?
21
Language learning example

• e.g. neiz -gt

22
Identifying and Correcting Errors

• e.g. neiz -gt knees/niece
• wen -gt

23
Identifying and Correcting Errors

• e.g. neiz -gt knees/niece
• wen -gt when/went/we/win
• fiknusiz -gt

24
Identifying and Correcting Errors

• e.g. neiz -gt knees/niece
• wen -gt when/went/we/win
• fiknusiz -gt thicknesses
• thhhee fdsooog rrrrraaanm -gt

25
Identifying and Correcting Errors

• e.g. neiz -gt knees/niece
• wen -gt when/went/we/win
• fiknusiz -gt thicknesses
• thhhee fdsooog rrrrraaanm -gt the dog ran
• John is teacher.
• Sandy is pig.
• I am doctor.
• John is a good man.

26
Providing feedback

• e.g. neiz -gt knees/niece
• wen -gt when/went/we/win
• fiknusiz -gt thicknesses
• thhhee fdsooog rrrrraaanm -gt the dog ran
• John is teacher.
• Sandy is pig.
• I am doctor.
• John is a good man.
• You seem to use no article instead of a or an
  before a singular count noun and after the verb
  to be

27
4. What sorts of tools and systems?
28
How Do We Teach? Teaching Strategies

• What method of teaching?
• ? drill and practice
• ? expository teaching
• ? mastery/apprenticeship
• ? learning by examples
• ? guided coaching
• ? learning by doing

29
Distinctions

• DIDACTIC V DISCOVERY LEARNING
• focus on system goals focus on learner goals
• OPPORTUNISTIC V DELIBERATED TEACHING
• exploits situation agenda of material
• DIAGNOSTIC V EFFECTIVE TEACHING
• goal is to ASSESS goal is to CHANGE
• current knowledge or knowledge or skill of
  student
• skill of student
• EXPOSITORY V PROCEDURAL TUTORS
• factual knowledge, skills and procedures,
• inferential skills examples and exercises

30
What Is An AILE?

• Tutoring and training systems which mimic tasks
  traditionally done by teachers
• A fairly broad definition, to include
• Intelligent Learning Environments,
• Intelligent Tutoring System
• Adaptive Learning Environments,
• Intelligent Computer Assisted Instruction,
• and other intelligent interactive
  teaching/learning tools...
• Blur distinction between tutoring and training
• use the terms tutoring, teaching,
  education
• and training interchangably....

31
Distinguishing Features

• 1. Can react intelligently to changing events in
  the tutorial session, that is, it is adaptive
• Has sufficiently explicit representations of its
  own knowledge that it can decide FOR ITSELF how
  to react.
• Might be expected to do a number of things, such
  as
• answer students questions
• adapt explanation to knowledge of learner
• choose appropriate problems
• adapt to learner's learning style
• maintain focus and direction in tutorial...
• Any one system may not do all of these things -
  but most will do a significant number of them.

32
Hartleys (1973) Framework

• 1.Representation of the knowledge of skill that
  is to be taught Domain Knowledge
• eg the ability to solve problems in the domain,
  to judge/comment on student's answer, or answer
  questions posed by the student.
• 2. Teaching actions Teaching Strategies and
  Tactics
• eg making positive/negative comments, providing
  examples, setting problems, asking the student to
  explain, find counter-examples, step through
  examples
• 3. Model of the student's current state Student
  Model
• ie history, capabilities, knowledge, beliefs,
  goals and motivation
• 4. Interface and Communication
• eg discourse between student and system, choice
  of interface - WIMP, graphics, text, speech, VR

33
Hartley's Framework Provisos..

• does not imply decomposition into modules, but
  that these should all be reflected in system
• we do not assume a teacher-directed view i.e.
  it can be mixed initiative
• student model should change over time as student
  learns
• mislearning should also be modelled
• We will use this traditional 4 part framework to
  structure the course, but we will not stick
  rigidly to it - often we cannot cleanly separate
  them.

34
Categorisation of Systems

• Relates to Engineering v. Cognitive Science
• Aim to create systems which achieve intelligent
  behaviour by any means
• (i.e. fruitful educational interaction)
• Systems which achieve intelligence in manner
  modelled on human information processors
• (i.e. educational interaction based on cognitive
  models of users)
• c. Systems as means of testing educational theory.

35
Current Educational Tools

• It is increasingly important for us to
  incorporate in our educational tools
• the capacity for dialogue
• to personalise them to the user.
• to consider where and how to include social
  skills in our pedagogical tools and agents

36
5. Course Overview
37
Aims and Objectives

• to equip students with the skills to design and
  evaluate intelligent and adaptive educational
  tools
• to better understand the role of communication in
  learning
• to enable students to better understand the
  relationship between informatics and education
• to understand the role that Informatics can play
  in testing educational theory.

38
Contributions to Programme Outcomes Knowledge
Understanding

• broad knowledge of theoretical basis for
  pedagogical tools
• understanding of role of Informatics in
  developing and testing pedagogical theory
• knowledge of previous work in developing
  intelligent educational tools and environments
• understanding of methodology for designing,
  building and evaluating educational tools and
  environments
• appreciation of difficulties of developing and
  testing effective pedagogical software.

39
Contributions to Programme Outcomes Intellectual
Skills

• Students will learn
• to apply pedagogical theory to the design of
  pedagogical software, and
• to understand the difference between software
  that is based on theoretical motivated design and
  that that has no underlying pedagogical
  principles.
• They will also learn techniques for evaluating
  the design and effectiveness of educational
  tools.

40
Contributions to Programme Outcomes

• Practical Skills
• Students will learn how various informatics
  techniques can be used in the design and
  implementation of pedagogical software, and will
  carry out evaluation of a number of existing
  systems.
• Design tasks will include interface and
  interaction design as well as educational/instruct
  ional design.
• Transferable Skills
• Critical evaluation skills. Ability to design and
  evaluate software. Writing reports on empirical
  studies.

41
Syllabus issues addressed..

• History of teaching systems and tools
• Relationship between informatics and education
• Tutors, tools and learning environments
• Methodology empirically informed and user
  centred design
• Pedagogical Issues
• Theoretical and Educational basis of teaching
  tools
• Using Pedagogical Agents
• Collaborative learning, peer tutoring and
  teaching simulated students
• Metacognitive Skills

42
Syllabus issues (continued)

• Modelling and simulating domain knowledge
• Qualititative modelling
• Authoring tools instructional planning.
• Virtual, multi-media and multi-modal interfaces
• Modelling the user
• Diagnosis, errors and misconceptions
• Open learner modelling
• Models of interaction and communication
• Educational dialogue
• Evaluating the design and effectiveness
• State of the Art problems and limitations
• Conclusions

43
Teaching Activities

• Lectures providing the background to the
  methodology and issues in AILE
• Some reading based seminar style sessions
  exploring specific issues
• Hands-on sessions introducing various tools and
  systems.

44
Assessment

• Two coursework assignments
• one will be designing a pedagogical system (or
  component of one), using knowledge or previous
  systems and empirical data to inform the design
• the other will be a practical one involving
  aspects of system implementation
• Examination worth 70

45
Materials

• Readings in George Sq Library and on-line
• (lot lost in fire)
• Will recommend papers
• Slides of lectures on web page (not ready yet)
• Plus other resources
• No set text.