One approach to axiom pinpointing (AP) in description logics is its reduction to the enumeration of minimal unsatisfiable subformulas, allowing for the deployment of highly optimized methods from SAT solving. Exploiting the properties of AP, we further optimize incremental SAT solving, resulting in speedups of several orders of magnitude: through persistent incremental solving the solver state is updated lazily when adding clauses or assumptions. This adaptation consistently improves the runtime of the tool by an average factor of 3.8, and a maximum of 38. SATPin, our system, was tested over large biomedical ontologies and performed competitively.
Manthey, N., Penaloza, R., Rudolph, S. (2020). SATPin: Axiom Pinpointing for Lightweight Description Logics Through Incremental SAT. KI - KÜNSTLICHE INTELLIGENZ, 34(3), 389-394 [10.1007/s13218-020-00669-4].
SATPin: Axiom Pinpointing for Lightweight Description Logics Through Incremental SAT
Penaloza Rafael.
;
2020
Abstract
One approach to axiom pinpointing (AP) in description logics is its reduction to the enumeration of minimal unsatisfiable subformulas, allowing for the deployment of highly optimized methods from SAT solving. Exploiting the properties of AP, we further optimize incremental SAT solving, resulting in speedups of several orders of magnitude: through persistent incremental solving the solver state is updated lazily when adding clauses or assumptions. This adaptation consistently improves the runtime of the tool by an average factor of 3.8, and a maximum of 38. SATPin, our system, was tested over large biomedical ontologies and performed competitively.File | Dimensione | Formato | |
---|---|---|---|
KIJ-2019v3.pdf
accesso aperto
Tipologia di allegato:
Submitted Version (Pre-print)
Dimensione
351.79 kB
Formato
Adobe PDF
|
351.79 kB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.