The goal of computer scientists working on Interactive Storytelling is to create computer programs that can improvise stories in real time taking into account input from the audience. While many algorithms have been suggested and systems were designed, almost none of them has managed to convincingly improvise stories in practise. This can be in large parts attributed to the difficulty of capturing the knowledge about the story world that these algorithms need to perform. In this talk I am going to give an overview of my PhD work in which I am trying to solve this knowledge acquisition problem by automatically extracting a large portion of the required knowledge from a set of example stories set within the story world.
Andrew MacVeen: Task-Involved Versus Ego-Involved: Motivating Children to Exercise In a Pervasive Exergame
This talk presents current work on a pervasive health and fitness game (exergame), designed to motivate children to reach their recommended daily exercise goals and facilitate long term behavioural change. I discuss briefly the current work in the area of pervasive exergames and highlight a common theme in the approach they take. Through my study of the relevant psychology literature, I identify a potential problem in this approach - the goal context that the majority of systems adopt. I hypothesize that a 'Task-Involved' rather than an 'Ego- Involved' system would be more suitable at addressing the problem of sedentary childhood behaviour.
Allan Weallans: Distributed Drama Management: Dramatic Intelligence in Emergent Narrative
Abstract: Emergent Narrative is a character-based approach to interactive digital storytelling with the aim of diminishing the effect of the Narrative Paradox - the inversely proportional relationship between authorial control of the story and user freedom within the story. However, Emergent Narrative in its current state is highly simulative: characters will select courses of action that are believable within the context of the story and the character's own personality, but they are not guaranteed to select dramatically interesting actions, much less sequences of dramatically interesting actions occurring in a recognisable dramatic structure. Our work, DISTRIBUTED DRAMA MANAGEMENT, intends to build on existing Emergent Narrative work by incorporating "dramatic intelligence" into our character agents by monitoring the story's dramatic needs and giving agents the ability to co-ordinate and re-prioritise their characters' goals for dramatic purposes. In this way, we can retain the character believability and action contextualisation that are the strengths of Emergent Narrative while allowing the characters to be steered towards fulfilling given narrative concerns such as dramatic tension, structure and theme.
No comments:
Post a Comment