1 / 16

A Proof Theory for DL-Lite

A Proof Theory for DL-Lite. Diego Calvanese, Evgeny Kharlamov, Werner Nutt Free University of Bozen-Bolzano June 2007. DL-Lite is a “Nice” Logic. 1. Covers basic constructs of UML, ER. Concepts:. Assertions: . DL-Lite is a “Nice” Logic. 2. Answering Conjunctive Queries (CQs):

cybill
Télécharger la présentation

A Proof Theory for DL-Lite

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. A Proof Theory for DL-Lite Diego Calvanese, Evgeny Kharlamov, Werner Nutt Free University of Bozen-Bolzano June 2007

  2. DL-Lite is a “Nice” Logic 1.Covers basic constructs of UML, ER Concepts: Assertions:

  3. DL-Lite is a “Nice” Logic 2.AnsweringConjunctive Queries (CQs): • Data complexity: LogSpace in • Combined complexity : PTime in • Query complexity: NP in Concepts: Assertions:

  4. DL-Lite is a “Nice” Logic 3.Rewritingtechniquesallow one to exploit current DBMS for CQ answering Concepts: Assertions:

  5. Conjunctive Queries over Knowledge Bases Knowledge Base : Conjunctive Query: Certain Answers:

  6. Perfect Query Reformulation[Calvanese et al. 05] Rewriting rules: Inclusion assertions: Procedure: apply a rewriting rule to a query and possibly factorize the result

  7. Perfect Query Reformulation[Calvanese et al. 05] Rewriting rules: Original CQ: ... ... ... ...

  8. Answering Conjunctive Queries • Store the ABox as RDB:RDB = { Person(Bob), HasFather(Bob,John) } • Query the RDB with the queries, obtained with the Perfect Query Reformulation:

  9. Questions • Which are the characteristics ofthat make it so nice for answering CQs? • Do these characteristics allow for alternative approaches to answering CQs?

  10. DL-Lite is a Fragment of an Extended Horn Logic (EHL) • Extended Horn Clause: Assertions: EHL Clauses:

  11. Properties of Extended Horn Logic • An EHL program has (at least one) Universal Model (UM), which can be homomorphically embedded into any other model • UMs are unique up to homomorphic equivalence • Answering CQs over all modelsis equivalent to answering CQs over any UM

  12. Resolution-Based Calculus forDL-Lite • Resolution (with facts and rules) • Factorization • -resolution (with rules containing )

  13. The Calculus at Work Goal: • Calculus Rules: • Resolution ( ), • Factorization ( ), • -resolution ( ) Extended Horn Program:

  14. Properties of the Calculus • Composition of substitutionsalong all successful derivations returns all certain answers • Soundness • Completeness

  15. Complete Strategies for Rule Application • Loop checking • Postponing resolution with facts until the very end • equivalent to the Perfect Query Reformulation • Selecting only one atom at a time for rule application ("Live-Only strategy") • SLD-Resolution is a special case of "Live-Only“for programs without  • early failure detection

  16. Conclusions • DL-Lite has nice characteristics because it is (essentially) Horn Logic • Existing query answering algorithms correspond to resolution strategies • Ideas from computational logic lead to • alternative approaches • optimization of existing techniques

More Related