Open AccessOntology-based Knowledge System and Team Verification Tool for Competitive Pokemon
Author(s) -
Daniel Verdear,
Ubbo Visser
Publication year - 2021
Publication title -
proceedings of the ... international florida artificial intelligence research society conference
Language(s) - English
Resource type - Journals
eISSN - 2334-0762
pISSN - 2334-0754
DOI - 10.32473/flairs.v34i1.128544
Subject(s) - computer science , domain (mathematical analysis) , ontology , semantic reasoner , semantics (computer science) , representation (politics) , artificial intelligence , precision and recall , domain knowledge , human–computer interaction , information retrieval , natural language processing , software engineering , machine learning , programming language , mathematical analysis , philosophy , mathematics , epistemology , politics , political science , law
Competitive Pokemon is a domain with rich semantics and complex relationships between its elements. Current research in the domain has focused on developing AI agents to select moves within a match, ignoring the problem of team building. We propose a team verification tool based on ontologies to model the complexities of the domain. A user can input their team into the tool, and the tool uses a description logic reasoner to classify Pokemon into their appropriate roles. The tool exports a visual representation of the team and a score evaluating its competitive viability. The classifications made by the TeamVerify tool have 87.7% precision and 86.0% recall in a multiclass, multilabel domain. We expect such a tool to reduce the learning curve for novice players who have not yet built intuitions on proper team structure.