Culturally Aware Agents for Training Environments (CAATE)

The U.S. Army is increasingly conducting non-kinetic missions that bring Soldiers into close contact with indigenous populations, requiring them to interact and communicate effectively with people whose culture, language, lifestyle, and beliefs are very different from their own. It has become increasingly clear that strategy, operational plans, and implementing procedures must occur in the context of these local cultural constraints in order to be effective. The U.S. Army is exploring simulation-based training tools for training the skills needed for success in culturally situated and sensitive missions. The modeling and simulation of culturally believable agents for such applications is a difficult and expensive proposition. Consequently, we propose to develop and evaluate Culturally Aware Agents for Training Environments (CAATE), a modeling and simulation tool for affordably creating software agents that embody local culture and belief sets to drive realistic, situationally relevant and culturally believable behaviors during simulated training missions. By focusing on the social and cultural elements that most affect mission success, we believe we can ensure problem tractability and solution feasibility. CAATE embodies sound didactic principles, uses social network models for modeling the social and cultural situation, and supports the development of human behavior models that display appropriate cultural behaviors.