UnityIIS - Symbolic Reasoning in Unity

UnityIIS is a package for the Unity Game Engine for implementing intelligent interactive systems (IIS). In particular, systems that integrates symbolic knowledge bases for reasoning, planning and rational decision-making in interactions with humans.

The UnityIIS framework is divided into modules to integrate varying agent-modelling and knowledge representation tools. The main modules are 1) the OWL Unity Package to enable ontology reasoning, and 2) the ASP Unity Package to enable ASP-based action reasoning and planning.

Prerequisites

Unity Game Engine: https://unity3d.com/get-unity/download

Installation

• Install Unity
• Create a new Unity 2D/3D project
• Import UnityIIS

Ontology Unity Package

• Integration of an Ontology reasoner in Unity through a web API.
• The Ontology Unity Package comes with a set of helper functions in C# to do operations with ontologies in Unity.
• This allows Unity-ontology projects to be deployed on any device that has internet access.
• Ontology Unity Package is licensed for academic and research use only (See license).

The integrated Ontology reasoner

• Ontology reasoner: BaseVISor
• BaseVISor 2.0.2 is licensed for academic and research use free of charge; all other uses require a commercial license at VIStology, Inc.
• Supported syntaxes: RDF/XML, OWL/XML, All OWL API
• Supported reasoning services: realisation, classification, satisfiability, conjunctive query answering, entailment, consistency.
• Creators of BaseVISor: VIStology, Inc.
• More info: https://www.vistology.com/products/

Getting started with the Ontology Unity Package

• Copy your .owl file to folder Assets/OntologyModule/encodings/
• In file Assets/OntologyModule/encodings/main.bvr, do the following:
  • Make sure that relativePath="ontology.owl" is your ontology name.
• In file Assets/OntologyModule/OntologyWebservice.cs, do the following:
  • Change 'project_key' to your project key, or use the key 'test'
  • Change 'ontology_file_name' to your owl file name, or rename your owl file to 'ontology.owl'

Start your project

• In file Assets/OntologyModule/encodings/main.bvr, do the following:
  • Take a look at how example rules are defined to print the results of the OWL inferencing
  • Define the rules you need for your ontology
• In Unity, create a new GameObject
• In Unity, add the script Assets/OntologyModule/OntologyWebservice.cs to a GameObject
• Run the game and check the output from the reasoner in the Unity Console

Using and adapting the code

• In the script Assets/OntologyModule/OntologyWebservice.cs
  • Parse the text response from the ontology reasoner
  • Define an approrpiate data structure for the response
  • Do application specific operations
    • Function: public void UpdateBVRFile(string input)
      • The function makes procedural changes to main.bvr, such as adding new rules at runtime
      • Any procedural changes are put in the section at the end of main.bvr
      • The input string is in RFD format (see example rules in main.bvr)
    • Function: public void UpdateOntologyFile(string input)
      • The function makes procedural changes to ontology.owl, such as adding new facts at runtime
      • Any procedural changes are put in the section at the end of ontology.owl
      • The input string is in RFD format (see example facts in ontology.owl)

Answer Set Programming Unity Package

The ASP Unity Package integrates an Answer Set Programming (ASP) solver and grounder, Clingo, to the Unity platform through a Web API. ASP defines a problem in terms of a logic program to enable its logical models (answer sets) to represent the solutions of the original problem. ASP is a declarative programming language to solve difficult, typically NP-hard, search problems, and is particularly useful in for non-monotonic reasoning, knowledge representation, planning, and different action reasoning formalisms.

UnityIIS comes with a set of helper functions to do operations (e.g., updating, querying) on ASP programs.

Cite as

Brännström, A., & Nieves, J. C. (2021). UnityIIS: Interactive Intelligent Systems in Unity (Version 1.0) [Computer software]. https://git.io/JMpzr

@software{Brannstrom_UnityIIS_Interactive_Intelligent_2021,
author = {Brännström, Andreas and Nieves, Juan Carlos},
month = {12},
title = {{UnityIIS: Interactive Intelligent Systems in Unity}},
url = {https://git.io/JMpzr},
version = {1.0},
year = {2021}
}

Authors

• Andreas Brännström {andreasb@cs.umu.se} Homepage
• Juan Carlos Nieves {jcnieves@cs.umu.se} Homepage

Department of Computing Science
Umeå university
SE-901 87, Umeå, Sweden