SimStudent is an interactive learner, who applies inductive logic
programming to learn cognitive skills from examples. SimStudent is
customizable (or "programmable") so that we can test various hypothesis to
identify factors that affect learning and test when and how such factors
affect learning. The learning outcome (i.e., learned cognitive skills) is
represented as a set of production rules, which is also human readable.
The above characteristics provides us unique opportunities to conduct
studies on machine- and human-learning. Currently, we have three research
areas as listed below:
Teachable Peer Learner
What if we ask human students to teach SimStudent? It is well known that students learn by teaching others. We can build an on-line learning environment where SimStudent acts as a peer tutoree and have human students tutor SimStudent. Such a learning environment also provides us opportunities to study more about the theories of tutor learning and other social and motivational factors of learning. See our REESE page at ARC to study more about our findings. An article at Research.gov introducing our study.
Computational Model of Learning
SimStudents can be used as simulated students with which researchers can
conduct various controlled studies to explore theories of learning. For
example, we have compared different learning strategies, learning from
worked-out examples vs. learning by being tutored. We have also studied how
differences in prior knowledge affect learning by comparing learning
outcomes from “good” SimStudent, who has stronger prior
knowledge of prerequisite concepts, and “poor” SimStudent, who has
weaker, more perceptually-oriented prior knowledge. See PSLC Theory Wiki
for details of the
Error Analysis study.
Intelligent Authoring
Use of SimStudent within the Cognitive Tutor Authoring Tools (CTAT)
allows authors to build a Cognitive Tutor (especially the so-called expert
model that represents domain principles) simply by teaching SimStudent how to
solve problems. SimStudent is available as a built-in component for CTAT,
which can be downloaded at the CTAT
project web page.
Last Updated
January 24, 2010
|
|