Affordable Cognitive Modeling Authoring Tools using HCI Methods: Carnegie Mellon University Portion

摘要

We were able to show that our tools can be used to build many different rule-based cognitive models, such as a Warrior Simulator at Fort Benning (Livak & Heffernan, 2004), a Logic Tutor for internal use at CMU, and a Genetics tutor. In preliminary controlled experiments involving basic Cognitive Tutor development tasks, we found efficiency gains due to CTAT of 1.4 to 2 times faster (Aleven, McLaren, Sewall & Koedinger, 2006). We also demonstrated across 4 different behavioral modeling projects that our tools created example-tracing tutors that drastically reduced modeling costs (Koedinger, Al even, Heffernan, McLaren, & Hockenberry, 2004, Heffernan, Turner, Lourenco, Macasek, Nuzzo-Jones, Koedinger, 2006). Not only did we reduce the time dramatically (averaging a reduction of over a factor 5), we also reduced the experience level modelers needed. Finally, we pushed the state of the art in data mining and machine learning support to help rule-writers (McLaren et al 2005; Harrer et al, 2005; McLaren et al, 2004b; McLaren et al, 2004a; Jarvis, Nuzzo-Jones, & Heffernan, 2004; Matsuda, Cohen, and Koedinger, 2005a; 2005b; 2005c).