4200 Lorcom Lane
Arlington, VA 22207
Phone: (703)524-4331
Fax: (703)527-6417
Overview
Dr. Cohen's work combines methods from cognitive psychology,
inference theory, artificial intelligence, and decision analysis
in a variety of research, training, system development, and
consulting applications. These include the development and
testing of methods for training cognitive skills; the design,
implementation, and testing of computer-based decision support
systems for both military and commercial applications; formal
research on methods for representing and reasoning about uncertainty;
and experimental research on processes of human decision making,
elicitation of expert knowledge, and design principles for
high-level aids which adapt to their users while guarding
against weaknesses in unaided cognitive processes.
Human-Computer Cognitive Interface Design
Dr. Cohen has developed a framework for decision
aid design, called Personalized and Prescriptive Aiding, which
uses cognitive theory to design aids that are compatible with
user processing strategies, and uses normative theory to mitigate
potential errors. That framework has been applied and tested
in a number of projects managed by Dr. Cohen: methods for adapting
display and analysis systems to user-preferred modes of representing
knowledge and making decisions (ONR), techniques for manipulating
and displaying uncertainty in expert systems (RADC), the design
of cognitively compatible cockpit aids for commercial airline
pilots (NASA) and for Air Force pilots (WPAFB), research on
the optimal allocation of cognitive tasks between users and
computer aids (ARI), and the experimental testing of human factors
principles for adaptive decision aiding (ONR/NUSC). In other
projects, Dr. Cohen has developed techniques for assessing the
value of competing information sets for inclusion in large data
base systems and for the intelligent real-time selection of
data for display (DARPA), managed a human factors review of
submarine advanced combat system designs (NUSC), and managed
a human factors review of a surface weapons system for the Navy
(NSWC). Dr. Cohen taught a graduate-level course at George Washington
University on the design of user-friendly interfaces for interactive
computer systems. The course included design options in displays,
input devices, and decision aiding; cognitive, linguistic, perceptual,
and motor considerations in the selection of design options;
and system testing and evaluation.
Decision Aid Development
Dr. Cohen has managed projects in which computer-based
decision aiding systems advanced from conceptualization through
stages of design, quantitative validation, and testing with
potential users. Two tactical decision aids developed by Dr.
Cohen for attack submarine command staff are under active testing
by a Navy laboratory (NUSC) and a Navy prime contractor. Other
systems developed in projects initiated and directed by Dr.
Cohen include a route planning aid for tactical air force mission
planners (RADC), a pilot's aid for in-flight situation reassessment
(WPAFB), an intelligent assistant for making selections from
a large database of options (ONR), a database system to support
inferences about uncertainty by intelligence analysts, and a
generic expert-system inference shell that utilizes principles
of non-monotonic reasoning to revise assumptions in a numerical
uncertainty model (ETL). More recently, Dr. Cohen has focused
on the link between training and decision aids. In the Navy's
Tactical Decision Making under Stress program (NAWCTSD), he
has developed concepts for the support of critical thinking
by a decision aid, and has developed and tested critical thinking
training for users of the aid. For the Army's Rotorcraft Pilot's
Associate program (AATD), Dr. Cohen has developed, applied,
and tested a methodology for training decision aid users.
Inference and Artificial Intelligence
Dr. Cohen directrd a project at CTI on the development
of a hybrid connectionist-symbolic architecture for learning
higher-level decision making skills (Office of Naval Research).
In the past, Dr. Cohen has directed a number of research projects
on the development of alternative representations of uncertainty
and processes of inference for AI systems. Dr. Cohen developed
an inference system, called the Non-Monotonic Probabilist, for
resolving conflict among different items of evidence or lines
of reasoning. NMP embeds a numerical belief calculus within
a higher-order process of reasoning about the structure and
assumptions of the model. That concept has been applied in an
expert system for image understanding (ETL), in an information
management system for intelligence analysts, and in an in-flight
pilot decision aid (WPAFB). Dr. Cohen managed a 2-week course
on decision making and inference for Intelligence Community
Staff; and was an invited panel member in a 3-day symposium
on Bayesian inference and artificial intelligence in intelligence
analysis.
Positions held
Cognitive Technologies, Inc., Arlington, Virginia (1990-)
- President, Senior Principal Scientist.
National Academy of Sciences, National Research Council
(1994-present) - Member of the Panel on Human Factors in Air
Traffic Control Automation.
Decision Science Consortium, Inc., Reston, VA (1979-1982;
1983-1990) - Vice President, Senior Principal Scientist, Director
of Cognitive Science & Decision Systems.
National Academy of Sciences, National Research Council (1985-1986)
- Member of the Committee on Tactical Battle Management of
the Air Force Studies Board.
The George Washington University, Graduate School of Arts
and Sciences, Washington, D.C. (1983-1984) - Assistant Professorial
Lecturer.
The MAXIMA Corporation, Bethesda, MD (1982-1983) - Program
Manager. Fee Control Group, Ltd., Washington, D.C. (l983)
- Consultant, partner.
Harvard University, Department of Psychology, Cambridge,
MA (1971-1979) - Research Psychologist, Teaching Fellow.