Adaptive Web-Based Leveling Courses

Shunichi Toida, Chris Wild, M. Zubair

Li Li, Chunxiang Xu Computer Science Department

Old Dominion University

OutlineMotivation and backgroundObjectivesSystem Overview

functional requirementsimplementation

StatusCourse structure JtreeArtificial intelligence in discrete mathStudent/peer awareness

Future WorkConclusions

NeedsNon-traditional Student

Second CareerTransferSecond Major

Non-traditional DeliveryAt Work/Home - AnywhereEvenings?weekends – AnytimeLess expensive

TechnologyInexpensive/Ubiquitous Multi-media PCsImproving Communications (internet)

Effective Utilization will requireLearning modelsMethods of organization and deliveryMotivational mechanisms

BackgroundODU CS Dept TechEd initiative

BS degree for AA graduatesTarget non-traditional studentsWeb-centric delivery of course material

BackgroundODU CS Dept TechEd initiative

BS degree for AA graduatesTarget non-traditional studentsWeb-centric delivery of course material

Problem: Diverse backgrounds of entering students

BackgroundODU CS Dept TechEd initiative

BS degree for AA graduatesTarget non-traditional studentsWeb-centric delivery of course material

Problem: Diverse backgrounds of entering students

Solution: Leveling courses in discrete math and programming

ObjectivesTo develop courses that are

adaptiveweb basedleveling supported by AI technologiesmanaged

System Overview

Use Case Summary

Instructor

C reate C ourseD escription

C reate C ourseC ontent

A na lyze andC ategorize S tuden ts

U pda te S tuden tsD ynam ic P ro file

BrowseC ourse

W ork on Q uizzesand T ests

V iew Progress

R egiser and SetProfile

Student

Adaptive Tutoring System

Functional RequirementsStudents

Navigate the course based on his profile and progressGet status on his/her progress and his relative performanceImmediate feedback where possible

InstructorSpecify courses structureClassify course contentsMonitor students performanceTrouble Alerts

Architectural FeaturesCourse description including pre-requisite structure (Oracle)

IEEE Learning Objects Metadata Standard

Student profile and progress (Oracle)Browsing support for course structure using appletContent access based on student progress

CourseDescription andStudents Profile

ContentRepository

Course navigatorand Pre-requisiteEnforcer Module

Student Servle t

HTTP Server/Request Handler

S tudentsRegistration

and S tatic P rofileModule

Students ProgressModule

Assignm ent, Q uiz,and Test Module

Course DescriptionModule

Instructor Servle t

Content CreationModule

Analyze S tudentsProfile

EvaluationModule

AI InterfaceCourse Description and

Students Profile InterfaceContent Interface

AISystem

NavigationApplet

EvaluationApplet

V iew ProgressApplet

In ternet

W eb Browser

S tudent In terface Instructor In terface

CourseDescription

Applet

Analyze ProfileApplet

S tudentEvaluate

AppletF irst T ier

SecondTier

ThirdTier

Status

Query-based content selection

Student/Peer AwarenessProblem: motivating in a self-paced courseShow progress relative to peersShow current class averages in assessment material

Artificial Intelligence in Discrete Math

Theorem prover and symbolic computation are used for exercises on:

English to logic translationChecking inferencesChecking induction proofs

Proving Equivalences of Natural Language to Logic

Translate the following sentence into predicate calculus using “likes(x,y)” predicate

“Nobody likes JOHN”There are multiple correct answers

Proving Equivalences of Natural Language to Logic

Translate the following sentence into predicate calculus using “likes(x,y)” predicate“Nobody likes JOHN”

)),(( JOHNxlikesx

)),(( JOHNxlikesx

Handling Multiple Solutions

Restrict response to unique canonical formCompare student response to “all” correct/obvious answersProve equivalence of student response to any correct answer

Handling Multiple Solutions

Restrict response to unique canonical formCompare student response to “all” correct/obvious answersProve equivalence of student response to any correct answerTPS: Theorem Proving System

Induction ProofsBuilt on the MAPLE symbolic computation system of MATLAB

Example1+2+… + n = n(n+1)/2

Example of Jtree/and some content

On-going and Future Work

Continue development of course materials (adaptability, exercises)Integrate piecesDefine evaluation metrics (market, effectiveness)Run assessment

ConclusionsNeed to serve non-traditional studentsNeed to adapt to diverse backgroundsNeed learning environment architectures and technologiesNeed effective learning strategies which leverage the potential of web connectivity

End of Presentation

Student Profile

<?xml version="1.0"?><!DOCTYPE STUDENT PROFILE "profile.dtd"><course title="cs381 course" student=”John Smith”> <block title="Propositional Logic" status="U"> <block title="Proposition" status="U"> <lesson title="What Is Proposition" href="course=cs381,block=cs381-1- block1.2,lesson=cs381-lesson01"> </lesson> </block> </block></course>

Course NavigationJava applet navigation of high level course structureAccess controlled by student profile

Course DevelopmentXML Course Mark-up LanguageCustomized for course structuree.g. course, block, lesson (marks)

Web-based Development ToolsServlet (Tomcat)Java Server Page (Tomcat)Java

O racleS erver

In ternet

C ontentR eposito ry

RequestHandler andPre-requisite

Enforcer

OracleInterface

S ervle t

HTTP Serverwith

Tom Cat ServletEngine

W eb B rowser

Information Flow

1. send a request to oracle in o rder to get the prerequ is itein form ation o f th is lesson2. ge t the prerequ is ite in fo rm ation from oracle3 . send a request to oracle in o rder to get the studentsta tus in form ation based on prerequ is ite .4 . ge t the student in fo rm ation and check the sta tus

The student who sa tis fies the prerequ is ite .5 . send a request to oracle to ge t the rea l sourcelocation .6 . ge t the location7. d isp lay the content on the web

The student who does not sa tisfy the prerequ is ite .5 . d isp lay the prerequ is ite tha t he does not sa tisfies andte ll h im to study such content