skip to main content


Title: Exploring Regression-Based Narrative Planning
A valid and believable narrative plan must often meet at least two requirements: the author’s goal must be satisfied by the end, and every action taken must make sense based on the goals and beliefs of the characters who take them. Many narrative planners are based on progression, or forward search through the space of possible states. When reasoning about goals and beliefs, progression can be wasteful, because either the planner needs to satisfy the author’s goal first and then explain actions, backtracking when an explanation cannot be found, or explain actions as they are taken, which may waste effort explaining actions that are not relevant to the author’s goal. We propose that regression, or backward search from goals, can address this problem. Regression ensures that every action sequence is intentional and only reasons about the agent beliefs needed for a plan to make sense.  more » « less
Award ID(s):
1911053
NSF-PAR ID:
10197962
Author(s) / Creator(s):
;
Date Published:
Journal Name:
Proceedings of the 12th Intelligent Narrative Technologies workshop at the 16th AAAI international conference on Artificial Intelligence and Interactive Digital Entertainment
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. Endriss, U. ; Nowé, A. ; Dignum, F. ; Lomuscio, A. (Ed.)
    Sabre is a narrative planner—a centralized, omniscient decision maker that solves a multi-agent storytelling problem. The planner has an author goal it must achieve, but every action taken by an agent must make sense according to that agent’s individual intentions and limited, possibly wrong beliefs. This paper describes the implementation of Sabre, which supports a rich action syntax and imposes no arbitrary limit on the depth of theory of mind. We present a search procedure for generating plans that achieve the author goals while ensuring all agent actions are explained, and we report the system’s performance on several narrative planning benchmark problems. 
    more » « less
  2. Thue, David ; Ware, Stephen G. (Ed.)
    Sabre is a narrative planner—a centralized, omniscient decision maker that solves a multi-agent storytelling problem. The planner has an author goal it must achieve, but every action taken by an agent must make sense according to that agent’s individual intentions and limited, possibly wrong beliefs. This paper describes the implementation of Sabre, which supports a rich action syntax and imposes no arbitrary limit on the depth of theory of mind. We present a search procedure for generating plans that achieve the author goals while ensuring all agent actions are explained, and we report the system’s performance on several narrative planning benchmark problems. 
    more » « less
  3. Narrative planning generates a sequence of actions which must achieve the author's goal for the story and must be composed only of actions that make sense for the characters who take them. A causally necessary action is one that would make the plan impossible to execute if it were left out. We hypothesize that action sequences which are solutions to narrative planning problems are more likely to feature causally necessary actions than those which are not solutions. In this paper, we show that prioritizing sequences with more causally necessary actions can lead to solutions faster in ten benchmark story planning problems. 
    more » « less
  4. In this paper we present two studies supporting a plan-based model of narrative generation that reasons about both intentionality and belief. First we compare the believability of agent plans taken from the spaces of valid classical plans, intentional plans, and belief plans. We show that the plans that make the most sense to humans are those in the overlapping regions of the intentionality and belief spaces. Second, we validate the model’s approach to representing anticipation, where characters form plans that involve actions they expect other characters to take. Using a short interactive scenario we demonstrate that players not only find it believable when NPCs anticipate their actions, but sometimes actively anticipate the actions of NPCs in a way that is consistent with the model. 
    more » « less
  5. This review presents a theoretical account of the development of possibility beliefs in childhood through two developmental pathways, centered around the experience and understanding of our intentional, goal-directed actions. Pathway 1 (Naïve Optimism to Calibrated Realism) can be seen as early as the first year, as increased coordination of action through motor experience leads infants to a graded notion of what is possible and how much effort is required to achieve goals. Infants also incorporate social information into their earliest possibility beliefs, referencing caregivers to guide them in uncertain situations and learning from role models to effectively calibrate effort. Pathway 2 (Naïve Pessimism to Creative Transcendence) emerges from ages 4 to 7. At first, preschoolers correctly distinguish possible and impossible actions but are overly pessimistic about limits on possibility. With age, children use their imaginations to overcome hypothetical limits. This account suggests that realistic beliefs about what we can possibly do are in place in early childhood, preceding later developmental milestones in self-concept, identity, self-efficacy, achievement-orientation, and self-goals. This leaves open questions about mechanisms of change, how possibility beliefs contribute to later self-beliefs, and whether interventions that combine action experience with creative idea generation can increase the sense of the possible in children and adults.

     
    more » « less