DC-Train 4.0 Simulator and Trainer for Crisis Decision Making
This talk presents the DC-Train 4.0 simulator and trainer for crisis decision making in the domain of ship damage control.
DC-Train simulates a ship and the personnel on a ship, and allows a decision maker through an immerse interface to solve problems that involve fire, smoke, flooding, rupture and the like. The innovative feature of this project from a reasoning standpoint is its method of knowledge representation of the domain knowledge, which consists of Declarative Graph modification Operators (DGMOs) composed of G-Clauses for representing static knowledge and a Causal Story Graph (CGG) for representing dynamic knowledge. This method of knowledge representation supports four very important and very diverse uses of the same knowledge. The knowledge can be inspected for correctness by domain experts, and it can be interpreted by a program to achieve the DC-Train expert model, the student critiquing model, and the tutoring model. A description will be given of ongoing extensions to DC-Train for spoken dialogue tutoring in conjunction with a Stanford-Illinois MURI Award. The DC-Train project received an AAAI/IAAI Deployed Innovative Application Award.
David Wilkins has been on the faculty of the University of Illinois
at Urbana-Champaign for fifteen years. He is Director of the
Knowledge Based Systems Group and holds faculty appointments in the Aviation Institute, Beckman Institute, and Department of Electrical and Computer Engineering. He received his Ph.D. from the University of Michigan in 1987 and his dissertation research was carried out in the Department of Computer Science at Stanford University from 1982-1987. He has published over 70 refereed articles in the areas of artificial intelligence, human-computer interaction, and intelligent tutoring systems. His primary research areas are machine learning and expert critiquing systems. More information can be found at http://www-kbs.ai.uiuc.edu.