The Representation and Learning of Procedures
Carnegie Mellon University
John Anderson (CMU)
This is a SISM NRA2 grant, initially
administered by the Intelligent Systems Project.
Instructions are frequently specified as procedures: lists of steps to be executed in sequence. Experienced users tend to find shortcuts to well-practiced linear procedures, and are often able to repair errors made during execution. Investigators will model human learning and execution of procedures, including prediction and analysis of such shortcuts and repairs. The ACT-R cognitive architecture will be used for this modeling, in the domain of procedures for the Boeing 777 Flight Management System. Elements of the procedures will be triggered by conditions in the operator's environment. Models in ACT-R are full processing models, interacting with experimental software as humans do and simulating behavior down to the level of eye movements and motor responses. The interactions will be compared with predictions of Freed's Apex cognitive modeling system. A particular goal is to be able to simulate the consequences of various instructional programs in terms of learning success and long-term performance.
|
Responsible NASA Official: Joseph C. Coughlan.
Program Support: Kenneth I. Laws.
/ Updated:
03-Oct-2005
Mail Stop 269-3, NASA Ames Research Center, Moffett Field, CA 94035-1000
NASA Privacy Statement.
For Section 508-accessible information, contact access@mail.arc.nasa.gov.
|