Gray, W. D. (2008). Cognitive modeling for cognitive engineering. In R. Sun (Ed.), The Cambridge handbook of computational psychology (pp. 565-588). New York: Cambridge University Press.
Cognitive modeling for cognitive engineering
FROM THE INTRODUCTION OF THE CHAPTER
Cognitive engineering is the application of cognitive science theories to human factors practice. As this description suggests, there are strong symbioses between cognitive engineering and cognitive science, but there are also strong differences.
Symbiosis implies a mutual influence, and the history of cognitive engineering supports this characterization in two key areas: the development of cognitive theory and the development of computational modeling software. For theory development, a stringent test of our understanding of cognitive processes is whether we can apply our knowledge to real-world problems. The degree to which we succeed at this task is the degree to which we have developed robust and powerful theories. The degree to which we fail at this task is the degree to which more research and stronger theories are required (Gray, Schoelles, & Myers, 2004).
The development of the production-system-based architectures most strongly associated with cognitive engineering [ACT-R (Anderson, 1993), EPIC (Kieras & Meyer, 1997), and Soar (Newell, 1990)] was motivated by the desire to explore basic cognitive processes. However, each has been strongly influenced by a formalism for cognitive task analysis that was developed explicitly for the application of cognitive science to human-computer interaction (Card, Moran, & Newell, 1980a, 1980b, 1983). Indeed, it can be argued that the modern form of ACT-R (Anderson et al., 2004) and the development of EPIC (Kieras & Meyer, 1997) with their modules that run in parallel owes a great intellectual debt to the development of CPM-GOMS (Gray & Boehm-Davis, 2000; John, 1988, 1993). It is definitely the case that the potential of these architectures for application has long been recognized (Elkind, Card, Hochberg, & Huey, 1989; Pew, 2007; Pew & Mavor, 1998) and that much recent development of these basic research architectures has been funded at least partly because of their potential in tutoring systems (S. F. Chipman, personal communication, 2007-04-02), human-computer interaction (Chipman & Kieras, 2004; Freed, Matessa, Remington, & Vera, 2003; Williams, 2000), or human-system integration (Gluck & Pew, 2005; Gray & Pew, 2004). On the other hand, the engineering enterprise of building systems that are in some way directly relevant to real world problems is fundamentally different from the basic research enterprise of developing or elaborating cognitive theory. Cognitive science and cognitive engineering can be viewed as differing along five dimensions. Although these differences do not imply a dichotomy, they can be viewed as capturing some of the characteristic differences of these two endeavors.
First is the nature of the problems picked. As an applied discipline, the problems addressed by cognitive engineering are often not picked by the researcher, but are defined for the researcher in terms of safety, workload, design, operational need, or financial impact.
Second is the amount of prior study of the task and task domain. Many of our best models of cognitive theory rest on years of exploring a small number of experimental paradigms within a well-specific domain. Great examples of this would be models of reasoning (Johnson-Laird, 1993; Rips, 1994), models of category learning (Love, Medin, & Gureckis, 2004; Nosofsky & Palmeri, 1997; Shepard, Hovland, & Jenkins, 1961), as well as models of memory retrieval (Anderson & Schooler, 1991; Hintzman, 2005). In contrast, many computational models for cognitive engineering tend to be first-generation attempts in that little, if any, prior empirical or modeling work exists. Two examples that are discussed in this chapter are Byrne and Kirlik’s (2005) work on modeling the taxiing behavior of commercial airline pilots and Gluck’s work on modeling uninhabited air vehicle operators (Gluck, Ball, & Krusmark, 2007).
Third, many but not all computational models for cognitive engineering entail domain-specific expertise. This characterization applies to both the development of tutoring systems for the training of novices as well as to the modeling of expert performance. It is definitely the case that much has been learned about basic cognitive processes by studying the acquisition or execution of expertise (Chi, Feltovich, & Glaser, 1981). It is also the case that there is a vast middle ground of educational research in which the distinction between basic versus domain-specific work is often blurred (Anderson, Conrad, & Corbett, 1989; Corbett & Anderson, 1988; Singley & Anderson, 1989). However, at the further extreme are the attempts to model rare forms of expertise, such as that possessed by Submarine Commanders (Ehret, Gray, & Kirschenbaum, 2000; Gray, Kirschenbaum, & Ehret, 1997; Gray & Kirschenbaum, 2000; Kirschenbaum & Gray, 2000), uninhabited air vehicle (UAV) operators (Gluck et al., 2007), or airline pilots (Byrne & Kirlik, 2005). Although, arguably, insights and progress into basic research issues have emerged from these studies, it is undoubtedly true that the motivation and funding to study and the particular expertise of such small populations stems from the need to solve very important applied problems.
Fourth, computational modeling for cognitive engineering operates in an arena where the demand for answers is more important than the demand for understanding. Newell warned us about such arenas (Newell & Card, 1985); if another discipline can reduce human errors, increase productivity, and in general augment cognition then who cares if those advances rely on an in-depth understanding of the human cognitive architecture? The issue for cognitive science is one of relevance.
Fifth, whereas many of our best cognitive science models focus on the distilled essence of a cognitive functionality such as memory or categorization, cognitive engineering models are called on to predict performance in task environments that entail many cognitive functionalities. Hence, the particular challenge of computational modeling for cognitive engineering is to model not just the pieces but also the control of an integrated cognitive system (Gray, 2007b).
These characteristic differences between basic and applied computational cognitive modeling are not meant as dichotomies, but rather to illustrate the different sets of challenges faced by cognitive engineering. To some degree these challenges can be seen as challenges for the basic science; especially the need for cognitive engineering to model the control of integrated cognitive systems (the last item on my list). Unfortunately, neither the list nor the efforts that instantiate it are tidy.
The next section reviews the seminal work of Card, Moran, and Newell (Card et al., 1983) from the modern perspective. We then jump to the 2000s to discuss the issues and applications of cognitive engineering, first for the broad category of complex systems and then for the classic area of human-computer interaction, with a focus on human interaction with quantitative information, that is, visual analytics . The chapter ends with a summary and discussion of cognitive engineering.
ABOUT THE VOLUME
This book is a definitive reference source for the growing, increasingly more important, and interdisciplinary field of computational cognitive modeling, that is, computational psychology. It combines breadth of coverage with definitive statements by leading scientists in this field. Research in computational cognitive modeling explores the essence of cognition through developing detailed, process-based understanding by specifying computational mechanisms, structures, and processes. Computational models provide both conceptual clarity and precision at the same time. This book substantiates this approach through overviews and many examples.
Please note that the copyright of this article is owned by the publisher, Cambridge University Press.Back to Home << Publications Visitors since 2004.12.08: