WPI Worcester Polytechnic Institute

Computer Science Department
------------------------------------------

AIRG Topics - Spring 2013


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


Jan 17
--

Jan 24
--

Jan 31
--

Feb 7
--

Feb 14
Academic Advising Appointment Day

Feb 21
--

Feb 28
--

March 7
Postponed
"Modeling Discussion Topics in Interactions with a Tablet Reading"
Adrian Boteanu & Sonia Chernova
    CloudPrimer is a tablet-based interactive reading primer that aims to foster early literacy skills and shared parent-child reading through user-targeted discussion topic suggestions. The tablet application records discussions between parents and children as they read a story and leverages this information, in combination with a common sense knowledge base, to develop discussion topic models. The long-term goal of the project is to use such models to provide context-sensitive discussion topic suggestions to parents during the shared reading activity in order to enhance the interactive experience and foster parental engagement in literacy education. In this paper, we present a novel approach for using commonsense reasoning to effectively model topics of discussion in unstructured dialog. We introduce a metric for localizing concepts that the users are interested in at a given moment in the dialog and extract a time sequence of words of interest. We then present algorithms for topic modeling and refinement that leverage semantic knowledge acquired from ConceptNet, a commonsense knowledge base. We evaluate the performance of our algorithms using transcriptions of audio recordings of parent-child pairs interacting with a tablet application, and compare the output of our algorithms to human-generated topics. Our results show that words of interest and discussion topics selected by our algorithm closely match those identified by human readers.

Mar 14
--

Mar 21
Joseph E. Beck
"EdRank: finding educationally useful webpages"
    The world wide web, combined with powerful search engines, has been a boon for education. More and more educational content is being put online, ranging from interactive virtual manipulatives to teach fraction arithmetic, to MOOCs teaching advanced topics in artificial intelligence. One limitation is that existing search technology, such as PageRank, is based on finding webpages that are popular, rather than those that are educationally effective. Unfortunately, in education, popularity and efficacy are not necessarily the same thing. Therefore, there is a need for a tool that estimates how educationally effective a webpage is. This work will propose an approach, dubbed "EdRank," that focuses on learning what properties of webpages make them educationally effective. We have a small sample of 7,000 student webpage visits, and exploit the knowledge that we have their performance data from both before and after visiting the webpage, to make a first attempt at constructing a model for predicting how useful a webpage is. This talk is definitely about a work in progress, as although we have initial results, we do not yet have a polished tool.

Mar 28
Yutao Wang
Using Bayesian networks for assessing students' knowledge.
    One of the most popular methods for modeling students’ knowledge is Corbett and Anderson’s Bayesian Knowledge Tracing (KT) model. The original Knowledge Tracing model does not allow for individualization. Several researchers have tried to show the power of individualization. Corbett and Andersen’s presented a method to individualize students’ parameters with a two phase process: first run Knowledge Tracing on all the students and then run a separate regression to learn a set of slip, guess, learning and prior parameters per students. Pardos and Heffernan explored the individualized student prior, but did not learn all of the student parameters and skill parameters in one single model. We presented the SS model, which is elegant in accounting for individual differences (of learning rate, prior knowledge and guess and slip rates). Our simulation showed that we could reliably fit such a model. We evaluated the model in comparison with the traditional knowledge tracing model in both simulated and real world experiments. The new model predicts student responses better than the standard knowledge tracing model when the number of students and the number of skills are large enough and there is enough data for each of them.

Apr 04

Apr 11

Apr 18
Undergraduate Project Presentation Day

Apr 25

Tue Apr 30
End of D term


[Feedback] [Search Our Web] [Help & Index]

[Return to the WPI Homepage] [Return to the CS Homepage]

AIRG Coordinator / Thu Mar 21 18:37:30 EDT 2013