Computational Influence for Training and Entertainment

David L. Roberts

Abstract: An interactive narrative is an education, training, or entertainment experience. There are two qualities that make an experience an interactive narrative. To understand those, let us clarify the meanings of interactive and narrative. Interactive: capable of acting on or influencing each other; and Narrative: presentation of events in a purposeful sequence. Thus, the qualities that define an interactive narrative are autonomy for players to act and intent of the system for the player to experience a narrative prescribed by an author. This conception of interactive narrative is broad and encompasses many types of entertainment and training experiences. One characteristic that sets it apart from other experiences and common games like Chess is that authorial intent often requires the player’s experience to be dramatic or adhere to some aesthetic. There is a tension between the systematic control required to ensure the intent of the narrative and the player autonomy required for interactivity— the player-driven exploration that results in an interactive quality is a potential threat to the narrative. Therefore, AI researchers in the field of interactive narrative are interested in balancing the conflicting requirements of autonomy and authorial intent. It is in this area that both I and many interactive narrative researchers have focused their efforts (Roberts and Isbell 2008). While many of the existing research projects have met with success in their own right, they have also uncovered limitations in the state of the art. I plan to complete my dissertation by addressing some of these limitations. The tools of artificial intelligence and machine learning have often been applied in domains where the limits of human ability are stretched. For example, sophisticated AI algorithms retrieve information from the internet, filter spam from email inboxes, and aid doctors in diagnosing illnesses. In each of these cases, a technically-minded AI expert has created a technique to solve a class of problems and has put the power of that technology in the hands of a practitioner. Similarly, I plan to develop AI technologies for computerbased gaming and simulation that when given to game designers will ease their authorial burden and increase their expressive power. The fundamental question I plan to an-