Intelligent Tutoring Systems and Affective Computing

1
Intelligent Tutoring Systems and Affective
Computing
2
What makes a good teacher?
3
What makes a good teacher?

• Mastery of the subject matter
• Good communication skills
• An understanding of what students get and dont
  get
• Genuine care and concern for the students

4
Intelligent Tutoring

• Mastery of the subject matter
• Good communication skills
• An understanding of what students get and dont
  get
• Genuine care and concern for the students

5
A Problem

• How can we more effectively help novices learn to
  program?
• What errors do they most commonly make?
• How can we intervene?

6
A Possible Approach

• Educational framework
• Examine the programming syllabus
• Make a set of exercises
• Make a set of all possible solutions to the
  exercises

• Programming
• Create a middle layer that will capture all
  compilation errors and their corresponding source
  files
• Analysis
• Compute for distances between the wrong solutions
  and the set of solutions

7
What makes a good teacher?

• Mastery of the subject matter
• Good communication skills
• An understanding of what students get and dont
  get
• Genuine care and concern for the students

8
Affective Computing

• Mastery of the subject matter
• Good communication skills
• An understanding of what students get and dont
  get
• Genuine care and concern for the students

9
Affective Computing

• Concerned with the development of systems that
  can detect, respond to, or exhibit human emotions
• Does not assert that computers have feelings
• Built on models of emotions that are
  computationally tractable

10
Current Research

• Studies the relationships between affect and user
  behaviors while using games, simulation
  environments, and intelligent tutors

11
A Problem

• What computer-detectible behaviors correspond to
  what emotions?
• What behaviors indicate boredom, frustration,
  delight, surprise, engagement, etc.

12
A Possible Approach

• Human observation to judge emotions
• Automated interaction and/or biometric sensor
  logging to record behaviors
• Statistics and data mining to find patterns and
  correlations

13
Related tools

• Synchronization tools of logs
• Video
• Human observations
• Biometric sensors
• Automatic interaction logging
• Validation tools
• Playback and labeling tools

14
Related work

• Gaming the system an attempt to advance
  through the curriculum without actually doing the
  work
• Intelligent agents now that you have the model,
  how do you fold it into the application?

15
More info?

• mrodrigo_at_ateneo.edu

16
For Networks

• Research Websitehttp//cng.ateneo.edu/cng/wyu/cl
  asses/cs197/Presentation Materialshttp//cng.a
  teneo.edu/cng/wyu/classes/cs197/presentation/discs
  -fmc-20070117.pdf