Back to Wayne Gray's home page / To Research Interests Index / To Downloadable Pubs / The CogWorks Lab / Cognitive Science Department


 

Approach -- Formal Methods

My approach to research is informed by formal methods. Indeed, my earliest work at ARI-Monterey attempted to provide a semi-formal notation for guiding the implementation of training programs (Gray, 1986). My work at CMU (Gray & Anderson, 1987; Gray et al., 1988) depended on the development of a goal-structure hierarchy. At NYNEX I mastered CPM-GOMS (e.g., Gray, et al., 1992; 1993) as well as the then current production system language for writing tutoring systems. In 1993 I became a devotee of ACT-R and have followed its development closely. Although ACT-R remains my favorite cognitive architecture, in recent years our research has run into limits of ACT-R with the result that we have joined the many researchers who are attempting to build the next generation "ACT-R" (their number is legion and includes Anderson himself). Recent efforts have turned towards mathematical modeling with Erik Altmann, Bayesian modeling with Wai-Tat Fu, Reinforcement Learning with Chris Sims, and many attempts to understanding the task environment better by formally asking "what is the best performance possible" in a given task environment (see the work with Hans Neth).

GOMS

Many tutorials, including:

And a few papers, including:

Task Analysis

Best non-GOMS example of this is

ACT-R

Examples include:

Bayesian, Mathematical, Reinforcement Learning, and other approaches

Examples include:


Back to Wayne Gray's home page / To Research Interests Index / To Downloadable Pubs / The CogWorks Lab / Cognitive Science Department

 

Last changed: 2006-09-05 wdg