Designs for an Intelligent Tutoring System for Students Learning to Program in Java

1
Designs for an Intelligent Tutoring System for
Students Learning to Program in Java

• Ed Sykes

2
Outline

• Introduction
• Intelligent Tutoring Systems
• User Interface Design
• JFlex CUP
• Error Recovery Techniques
• Demo
• Questions

3
Introduction

• Intelligent Tutoring Systems (ITS) are software
  systems that tutor students in specific domain.
• They have had significant achievements
• Human expert tutors have increased student
  achievement by 2.0 SD (thats 2 letter grades)
• Idea
• ? Model human tutoring techniques within ITS

4
Introduction

• Intelligent Tutoring System Success
• 0.4 to 1.0 SD improvement in performance
• Self-paced learning
• Students are engaged (rich-interaction)
• Current research is for Distance Education

5
Research Goal

• To design, and construct an ITS for tutoring
  post-secondary students learning the Java
  programming language.

6
Rationale

• Java programming language
• most popular language for Internet development
• millions of Java programmers world wide
• many post-secondary institution offer Java
  courses
• little research done in ITS for tutoring
  programming

7
User Interface Design
8
(No Transcript)
9
Example

• Problem
• Write a program that adds all integer numbers
  from 1 to a specified number (N). For example,
  if N were assigned the value 10, then the sum of
  the numbers from 1 to 10 is 55.
• Program specifications
• This program requires the use of a for-loop
  structure. Fill in the code to complete this
  program.
• OUTPUTgtSum 55

10
Demonstration

• Java Intelligent Tutoring System demo

11
Issues

• Issues
• Student will make mistakes (syntax and logic
  errors)
• Goal
• Intelligently recognize the intent of the
  student
• Need to analyze the students code submission
• auto-correct where appropriate
• produce a modified code that will compile and
  meet program specification
• ? feedback to the student

12
JFlex CUP

• To intelligently recognize the intent of the
  student
• Need to analyze the students code submission
• JFlex Lexical analyzer generator (in Java)
• CUP Parser generator (in Java)
• JFlex ?? Lex
• CUP ?? YACC

13
JFlex CUP

• 1st task Test Identifiers against keywords
• (e.g., For, fro, FOR, forr should
  all as the keyword for not identifiers)
• Error-recovery Action Techniques
• Deleting an extraneous character
• Inserting a missing character
• Transposing two adjacent characters
• Replacing an incorrect character by a correct
  character

14
JFlex CUP

• Demonstration
• Test_Identifiers program
• JFlex file java.flex

15
JFlex CUP

• 2nd task Perform error recovery during parsing
• e.g.,
• for ( int i0 ilt10 i ) // missing
  semi-colon
• sum sum i
• for ( int i0 ilt10 i // missing )
• sum sum i

16
Error Recovery Techniques

• Basic error recovery technique in YACC and CUP
• An error occurs --gt
• munch tokens until a reasonable starting point is
  found
• munch tokens until 3 in succession can be
  recognized
• Usually results in unsuccessful recovery
• Many error messages are produced.
• Historical reason behind this approach

17
Error Recovery Techniques

• Goal for Java Intelligent Tutoring System
• Only 1 error message is produced.
• Parsing stops
• Student addresses the 1 problem and corrects it
• Reduces student from getting overwhelmed with
  error messages
• Focuses the student on the error statement

18
Error Recovery Techniques

• Demonstration
• Java12.cup file

19
Questions