SS1 : New Approaches to Improving and Measuring Believability in Games

Using state of the art Artificial Intelligence techniques we already build intelligent agents able to surpass human capabilities in many ways, for instance playing chess. However, in the domain of games, intelligence is not necessarily a synonym of believability. In modern games, Non-Player Characters (NPCs) are expected not only to show intelligence, but to appear human in the way they behave. To artificially generate such a realistic human-like behavior is both a scientific and technical challenge.

The aim of this special session is to analyze current approaches and propose new research lines in the field of believability from the point of view of NPCs development and also from the point of view of human-like behavior assessment. Topics of interest include, but are not limited to, believability applied to games in the following aspects:

  • Design and Implementation of Believable NPCs.
  • Believable Content Generation.
  • On-Line Learning in Believable NPCs.
  • Human-Like Learning in Games.
  • Game Play Experience and Believability Assessment.
  • Behavior Pattern Analysis in Games.
  • Measures of Human-Like Intelligence in Games.
  • User Modelling in Games.
  • Practical Applications of believable NPCs.
  • Algorithms for the Generation of Human-Like Behavior.
  • Believability versus Playability.
  • Future Directions in Human-Like Intelligence.
  • Believability in Social Games.
  • Believable Emotions in Game Characters.
  • Emergent Behavior in NPCs.
  • Player Behavior Prediction.
  • Human-Like Motion Planning.
  • Social Path Planning.
  • Human-Like Combat Reasoning.
  • Machiavellian Intelligence in Games.
  • Human-NPC Social Interaction.
  • Player Cognitive Modelling.
  • Restricted Turing Tests in Games.


Juan Peralta Donate (Centre de Visió per Computador / UAB)

Manuel G. Bedia (Universidad de Zaragoza)

Raúl Arrabales (Conscious Machines Lab. U-tad).

[Print Version]