Simulation-based training has proven to be an efficient and effective approach to training cognitive decision making. Militaries around the globe have adopted simulation-based training into their curriculum (as evidenced by I/ITSEC) and it is being seen more and more in the corporate world.
At the end of his blog, “Psychological Research Guides Next Generation Computing“, Irving Wladawsky-Berger describes a vision of next generation computing taking advantage of cognitive science research such as described in Daniel Kahneman‘s 2011 bestseller Thinking, Fast and Slow. That book is certainly one place in cognitive science and psychology research to look for inspiration – there are many others as well. I am particularly inspired by Bonnie John, Janet Kolodner, Gary Klein, Mica Ensley, B. Chandrasekaran, Allen Newell, and Benjamin Bloom.
The good news is that we do not have to wait for the next generation of computing – we are already there. Discovery Machine methodology and technology enables the visual modeling of cognitive decision making and the ability to execute the visual models as software. Discovery Machine executable models – we call gears – can be run in simulation environments so that avatars and entities will exhibit complex decision making.
Kahneman discussed two type of thinking – fast reactions based on recognized patterns and slow based on more methodical analysis. Both types of thinking are represented in Discovery Machine cognitive models. The models of expertise represent situation awareness, mission, and reaction. The situation awareness collects information about the environment, understands the patterns, and when possible predicts possible outcomes. The mission is similar to Kahneman’s concept of slow thinking. The avatar will have a set of steps to execute to accomplish a goal. The reaction is similar to Kahneman’s concept of fast thinking. If someone is trying to hurt me I will get out of the way quickly.
We represent cognitive models the good old-fashioned way. We talk with the experts themselves. We find that with our cognitive architecture and methodology, we can very efficiently capture needed decision making by working directly with the subject matter expert. That does not preclude us from leveraging data analytics, but for our current applications, the data does not exist. All of the knowledge is the head of the subject matter expert. By leveraging a solid foundation of cognitive science principles, we are able to help the subject matter expert go through the introspection and articulation of expertise in well scoped domains. We simulate the expertise under relevant conditions to verify and validate the expertise. The models are changing consistently as new evidence and revelations are made by the subject matter experts.
In the past this type of knowledge acquisition has been cost and time prohibitive, but one of the differentiators of Discovery Machine is the fact that we can reduce the knowledge engineering bottleneck and get the expertise working for the organization quickly.
With the expertise modeled in a simulation environment, rich scenarios can be developed to transfer that knowledge to other members of the organization. Leadership, tactics, communication, and cooperation are a few of the examples of skills that can be transferred in a simulation-based training event. Providing the right intelligent avatars enables the simulation-based training to be adopted on a broader scale and with more complex scenarios without a comparable increase in cost.