Demo: Generate Emergent NPC Behaviours With Symbolic Reasoning

Abstract

We propose to use ontologies and declarative symbolic reasoning to generate emergent behaviours of Non-Player Characters (NPCs). The objective is that the game designer only needs to specify the rules of the game and its components in a declarative way, as he would naturally do in a traditional board game. The logic reasoner will then deduce the NPC behaviours that comply with the game designer’s rules without the game designer having to manually specify all the game possibilities by hand. We illustrate this approach on a prototype of the revisited Wumpus World game made on the Unity game engine with a Prolog environment. This approach is combined with the Well-Founded Semantics (WFS) to solve the problem of representation and reasoning despite the lack of NPC knowledge.

Type
Sylvain LAPEYRADE
Sylvain LAPEYRADE
Data Scientist & PhD Student

My interests include Artificial Intelligence, Data Science, Machine Learning and Games.