Notice
This item was automatically migrated from a legacy system. It's data has not been checked and might not meet the quality criteria of the present system.
Baldazzi, T., Bellomarini, L., Sallinger, E., & Atzeni, P. (2021). Eliminating Harmful Joins in Warded Datalog+/−. In Rules and Reasoning (pp. 267–275). https://doi.org/10.1007/978-3-030-91167-6_18
E192-02 - Forschungsbereich Databases and Artificial Intelligence E192 - Institut für Logic and Computation
-
Published in:
Rules and Reasoning
-
Date (published):
2021
-
Event name:
RuleML+RR 2021
-
Event date:
8-Sep-2021 - 15-Sep-2021
-
Event place:
Leuven, Belgium, Belgium
-
Number of Pages:
9
-
Peer reviewed:
Yes
-
Abstract:
We provide a rewriting technique of Warded Datalog+/− settings to sustain decidability and data tractability of reasoning tasks in the presence of existential quantification and recursion. To achieve this behaviour in practice, reasoners implement specialized strategies which exploit the theoretical bases of the language to control the effects of recursion, ensuring reasoning termination...
We provide a rewriting technique of Warded Datalog+/− settings to sustain decidability and data tractability of reasoning tasks in the presence of existential quantification and recursion. To achieve this behaviour in practice, reasoners implement specialized strategies which exploit the theoretical bases of the language to control the effects of recursion, ensuring reasoning termination with small memory footprint. However, as a necessary condition for such exploitation, the setting is required to be in a "normalized form", essentially without joins on variables affected by existential quantification. We present the Harmful Join Elimination, a normalization algorithm of Warded Datalog+/− that removes such "harmful" joins, supporting the tractability of the reasoning task as well as the full expressive power of the language. The algorithm is integrated in the Vadalog system, a Warded Datalog+/− -based reasoner that performs ontological reasoning in complex scenarios.
en
Project title:
Scalable Reasoning in Knowledge Graphs (WWTF Wiener Wissenschafts-, Forschu und Technologiefonds)
