Dichotomies in Ontology-Mediated Querying with the Guarded Fragment
This work addresses the complexity classification problem for ontology-based querying, which is crucial for database and AI applications, though it is incremental in building on existing theoretical frameworks.
The paper investigates the complexity of ontology-mediated querying using the guarded fragment of first-order logic, identifying fragments with a dichotomy between PTime and coNP complexity and showing that others lack such a dichotomy, with implications for real-world ontologies like those in BioPortal.
We study the complexity of ontology-mediated querying when ontologies are formulated in the guarded fragment of first-order logic (GF). Our general aim is to classify the data complexity on the level of ontologies where query evaluation w.r.t. an ontology O is considered to be in PTime if all (unions of conjunctive) queries can be evaluated in PTime w.r.t. O and coNP-hard if at least one query is coNP-hard w.r.t. O. We identify several large and relevant fragments of GF that enjoy a dichotomy between PTime and coNP, some of them additionally admitting a form of counting. In fact, almost all ontologies in the BioPortal repository fall into these fragments or can easily be rewritten to do so. We then establish a variation of Ladner's Theorem on the existence of NP-intermediate problems and use this result to show that for other fragments, there is provably no such dichotomy. Again for other fragments (such as full GF), establishing a dichotomy implies the Feder-Vardi conjecture on the complexity of constraint satisfaction problems. We also link these results to Datalog-rewritability and study the decidability of whether a given ontology enjoys PTime query evaluation, presenting both positive and negative results.