AIRG Topics - Fall 2007

Our group meets on Thursdays at 11:00 a.m., FL 246, Beckett Conf. Room.

Dates and topics for this semester are as follows:

Sept 6

AIRG Organizational Meeting (Coordinator: Chuck Rich)

{DCB away}

Sept 13

No Meeting

Sept 20

Joseph Beck

"Does help help students learn? An approach using Bayesian networks"

An increasing number of students are using educational software as part of their learning experience. Most educational software allows students to ask the system for assistance. A surprisingly complex question is whether such assistance actually helps the students. In this talk I discuss a variety of obvious metrics for measuring the effects of such assistance and show that each has clear drawbacks. I then present an approach using Dynamic Bayesian Networks (DBN) that controls for many of the confounds present in simpler approaches. By using a DBN as an analysis tool it is possible to split the effects of help into two components: assistance on the immediate problem and actual long-term learning by the student. Analyzing students' reading of 2 million words within a computer tutor for reading shows that that system's help improved student performance in both the short- and long-term.

Sept 27

Charles Rich

"DiamondHelp: A Generic Collaborative Task Guidance System"

DiamondHelp is a generic collaborative task guidance system motivated by the current usability crisis in high-tech home products. It combines an application-independent conversational interface (adapted from online chat programs) with an application-specific direct manipulation interfaces. DiamondHelp is implemented in Java and uses Collagen for representing and using task models.
[Paper with Candace Sidner, AI Magazine, Vol. 28, No. 2, Summer 2007]

Oct 4

No Meeting

Oct 11

No Meeting

Oct 18

Ugrad break - No meeting

Oct 25

No Meeting

Nov 1

No Meeting

Nov 8

No Meeting

Nov 15

Zachary A. Pardos

"The effect of model granularity on student performance prediction using Bayesian networks"

A standing question in the field of Intelligent Tutoring Systems and User Modeling in general is what is the appropriate level of model granularity (how many skills to model) and how is that granularity derived? This research will explore varying levels of skill generality and measure the accuracy of these models by predicting student performance within the tutoring system called ASSISTment as well as student performance on the Massachusetts standardized state test. Predicting students' state test scores will serve as a particularly stringent real-world test of the utility of fine-grained modeling. The use of Bayes nets is employed to model user knowledge and for prediction of student responses. The ASSISTment online tutoring system was used by over 600 students during the school year 2004-2005 with each student using the system 1-2 times per month through out the year. Each student answered over 100 state test based items and was tutored by the system when they got a state test item incorrect. The models that will be evaluated are skill models of grain sizes containing 1, 5, 39 and 106 skills. The results of this research show that the finer the granularity of the skill model, the better student performance for the online data can be predicted. However, for the standardized test data, it was the 39 skill model that performed the best. The results support the use of fine-grained models in student knowledge assessment and prediction.

Nov 22

Thanksgiving Break - No Meeting

Nov 29

Dec 6

Dec 13

End of Semester - No Meeting