An Adversarial Risk Analysis Model for an Autonomous Imperfect Decision Agent

摘要: Machines that perform intelligent tasks interacting with humans in a seamless manner are becoming reality. A key element their design is ability to make decisions based on reasonable value system, and the perception of surrounding environment, including incumbent persons. In this chapter, we provide model supports decision making process an autonomous agent imperfectly perceives its environment actions performed by person, which shall designate user. The approach has analytic flavour, but includes models forecasting user’s behaviour impact over environment. We describe implementation edutainment robot sensors capture information about world around it, may serve as cognitive personal assistant, be used kids for educational, recreational therapeutic purposes elderly people companion purposes.