 |
|
|
|
|
|
|
|
|
|
|
|
|
 |
|
|||||||||||
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
|
|
 |
 |
|
||||||||||
 |
|
|||||||||||
 |
|
|||||||||||
 |
 |
|
||||||||||
| Seminar Schedule (Seminar Organizer: Prof. Ziv Bar-Joseph) | |
Machine Learning Lunchtime Chats
|
Spring 2007 Seminar
Date: May 3, 2007 Time: 10:00 AM - 11:00 AM (Refreshments) Location: 3305 Newell-Simon Hall Speaker: Hao Cen Ph.D. Candidate Title: Learning Factors Analysis - A General Method for Cognitive Model Abstract: Cognitive Tutors are an intelligent tutoring system based on cognitive psychology results. By 2006, Cognitive Tutors have been widely used in over 1300 school districts in the U.S. by over 475,000 secondary school students who spend about 40% of their class time using the software. A center piece of the Tutors is cognitive models, which represent the knowledge a student might possess about a given subject. Cognitive Tutors use the cognitive models to assess students' knowledge step by step and present curricula tailored to individual skill levels. However, usually generated by brainstorming and iterative refinement between subject experts, cognitive scientists and programmers, the first pass of cognitive models are best guesses and our experience is that such models can be improved. In this KDD project, we propose a semi-automated method called Learning Factors Analysis for improving cognitive models. Innovatively combining a statistical model, human expertise and a combinatorial search, this method can generate alternative meaningful cognitive models and search for better ones. We use this method to evaluate and improve a cognitive model in Cognitive Geometry Tutor. We present the improved cognitive models and make suggestions for enhancing the intelligent tutor based on those models.