Owl gres vs quonto
This presentation is the property of its rightful owner.
Sponsored Links
1 / 59

OWL-Gres vs Quonto PowerPoint PPT Presentation


  • 61 Views
  • Uploaded on
  • Presentation posted in: General

OWL-Gres vs Quonto. Angela Alvarez Rubio. Introduction. Using ontologies as a conceptual point of view on repositories of data is increasingly. These ontologies deal with large amounts of data. Most important parameter on computational complexity of reasoning. Data size.

Download Presentation

OWL-Gres vs Quonto

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.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.


- - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - -

Presentation Transcript


Owl gres vs quonto

OWL-Gres vs Quonto

Angela Alvarez Rubio


Introduction

Introduction

  • Using ontologies as a conceptual point of view on repositories of data is increasingly

  • These ontologies deal with large amounts of data

  • Most important parameter on computational complexity of reasoning

  • Data size

  • We will want a polynomial reasoning!

  • And we want can do complex questions


Introduction1

Introduction

  • 2 stapes:

  • 1. Perfect reformulation: taking into account the TBOX T, the q query is reformulated in a new query

  • On a DL-Lite

  • Conjunctive query is a union of conjunctive wich size does not depend on A

  • We can evaluate it with LOGSPACE on the ABOX size


Introduction2

Introduction

  • 2 stapes:

  • 2. Query Evaluation: the new query is evaluated only in the ABox to produce the answer

  • The ABox, is maintained through a RDBMS (Data management systems relational) in the secondary storage to control a large data number

  • Because is the unique tecnology

  • The evaluation of the query can be delegated to an engine SQL database with optimization of querys strategies


Introduction3

Introduction

  • We presented two systems to work with large amounts of data:

  • OWL-Gres

  • Quonto


Introduction4

Introduction

  • Targets

  • Discover the DL-Lite fragment in which is based in OWL_Gres

  • Compare the OWL-Gres system with Quonto system


Quonto

Quonto

  • Is a tool that implements the DL-Lite query answering algorithm

  • Delegates to a RBDMS the storing of the ABOX

  • Is capable of answering questions about ABOXes wich containing millions of assertions

  • Their limitations will depend of the single engine DBM


Quonto dl lite a

Quonto: DL-Lite A+

  • Is the fragment DL-Lite largest known in order to obtain LOGSPACE data complexity

  • Represents the domain in terms of concepts, sets of objects, and roles and permets:

  • Value-Domains: domains that denote specific sets of values (data)

  • Concept attributes: binary relations between objects and values

  • Role attributes: ternary relations between pairs of objects and value

  • Enjoys FOL-rewritability

  • Allows for functionality assertions and role inclusion assertions, but with some restrictions:

  • No functional role or attribute can be specialized by using it in the right-hand side of a role or attribute inclusion assertions


Quonto dl lite a1

Quonto: DL-Lite A+

  • Concept inclusion assertion: B ⊑ C

The knowledge base (KB) is formed by:

  • Attribute inclusion assertion: U ⊑ V

  • T: TBOX to represent intensional knowledge

  • Value-domain inclusion assertion: E ⊑ F

  • Role inclusion assertion: Q ⊑ R

  • Attribute functionality assertion: funct U

  • Role functionality assertion: funct Q

  • Attribute Role: funct R

K=<T, A>

  • A: ABOX to represent extensional knowledge

  • Member Assertions

A(c), P(c; c0),

UC(c; d) UR(a, b, c)


Quonto query answering

Quonto:Query answering

  • Query conjunctive in a KB K:

  • Union of conjunctive queries (UCQ):

  • x: Distinguished variables

  • y: Non-distinguished variables

  • conj (x, y): atoms:

    • A(xo)

    • P(xo, yo)

    • D(xv)

    • UC(xo,xv)

    • UR(xo,y0 xv)

q(x) ←y. conj(x,y)

  • xo, yo are variables in x and y or constants in ГO

  • xv is a variable in x and y a constant in ГV

Certain answers all tuples t of elements of ГV ГO such that, when substituted to x in q(x), we have that K |= q(t)

q(x) ←Viyi. conj(x,y)


Firs target

Firs Target

  • On what DL-Lite fragment is based OWL-Gres?

  • See the characteristics of potential fragments and differentiate it

  • 2 steps:

  • Java Program

  • See if OWL-Gres accept this characteristics

  • Protege tool


First target fragments

First Target: Fragments


First target fragments1

First Target: Fragments


First target fragments2

First Target: Fragments


First target fragments3

First Target: Fragments


First target fragments4

First Target: Fragments


First target fragments5

First Target: Fragments


First target fragments6

First Target: Fragments


First target search

First Target: Search

  • We use a TBOX based in the university hierarchy:


First target search1

First Target: Search

  • Initially our TBOX is compatible with OWL-Gres:

C:\Documents and Settings\Propietario\workspace\OwlGres

21-jul-2008 12:54:42 org.coode.owl.rdfxml.parser.OWLRDFConsumer endModel

INFO: Total number of triples: 617

21-jul-2008 12:54:42 org.coode.owl.rdfxml.parser.OWLRDFConsumer endModel

INFO: Loaded http://semantics.crl.ibm.com/univ-bench-dl.owl

The TBox is compatible with DL-Lite


First target search2

First Target: Search

  • We verify for DL-Lite F:


First target search3

First Target: Search

  • We verify for DL-Lite F:

C:\Documents and Settings\Propietario\workspace\OwlGres

21-jul-2008 12:57:25 org.coode.owl.rdfxml.parser.OWLRDFConsumer endModel

INFO: Total number of triples: 618

21-jul-2008 12:57:25 org.coode.owl.rdfxml.parser.OWLRDFConsumer endModel

INFO: Loaded http://semantics.crl.ibm.com/univ-bench-dl.owl

FRAGMENT ERROR: No support for axiom OWLFunctionalObjectPropertyAxiom

On OWL Axiom: FunctionalObjectProperty(takesCourse)

The TBox is not compatible with DL-Lite

The TBos is not compatible with DL-Lite FR or DL-Lite A

DL-Lite F

DL-Lite FR

DL-Lite A

DL-Lite R


First target search4

First Target: Search

  • We verify for DL-Lite R:


First target search5

First Target: Search

  • We verify for DL-Lite R:

C:\Documents and Settings\Propietario\workspace\OwlGres

21-jul-2008 12:58:45 org.coode.owl.rdfxml.parser.OWLRDFConsumer endModel

INFO: Total number of triples: 619

21-jul-2008 12:58:45 org.coode.owl.rdfxml.parser.OWLRDFConsumer endModel

INFO: Loaded http://semantics.crl.ibm.com/univ-bench-dl.owl

The TBox is compatible with DL-Lite

OWL-Gres is based on DL-Lite R


First target search6

First Target: Search

  • But…

    • We have concept attributes…

  • IS-A for concept attribuites?

  • Range(Uc) IS-A Datatype NO

  • Person IS-A domain(Uc) assertion NO


First target conclusions

First Target: Conclusions

  • OWL-Gres is based on:

  • DL-Lite R

  • Concept attribuites

  • IS-A for concept attribuites


Second target preliminary notes

Second Target:Preliminary notes

edgeR-⊑Node

edgeR ⊑Node

edgeB-⊑Node

edgeB ⊑Node

NodeRB ⊑ edgeR

NodeRB ⊑edgeB

edgeB(a,a)

NodeRB(a)

ABOX

TBOX

q(x) ← y, z, w. edgeB(x,y)  edgeR(x,z)  edgeR(y,z)

{a}

  • Standard

2 types of semantic

q(x) ← y, z, w. edgeB(x,y)  edgeR(x,z)  edgeR(y,z)

  • Ground

{}

q(x,y,z) ← y, z, w. edgeB(x,y)  edgeR(x,z)  edgeR(y,z)

{}


Second target preliminary notes1

Second Target:Preliminary notes

q(x) ← hasSameHomeTownWith(x,y)  isMemberOf(y,z)  hasMember(z,t) isCrazyAbout(t,w)  isCrazyAbout(x,w)

Let’s consider the query 15:

isMemberOf

Y

Z

hasMember

The query is designed on purpose to establish if a reasoner is able to answer according to the standard conjunctive query semantic:

Quonto gives out 94 answers

OWLGres gives out 89 answers, like Racer, Pellet, etc..

hasSameHomeTownWith

T

isCrazyAbout

X

W

isCrazyAbout


Second target experiment conditions

Second Target:Experiment conditions

  • We have made two comparisons:

  • Without optimizations

  • Keep the reasoners near as much as possible from the optimizations point of view

  • With optimizations

  • What are they?


Second target experiment conditions1

Second Target:Experiment conditions

Optimizations


Second target experiment conditions2

Second Target:Experiment conditions

Semantic conjunctive query minimization

q(x) :- PeopleWithHobby(x), like(x,y)

PeopleWithHobby ⊑  like

Quonto

q(x) :- PeopleWithHobby(x)

Quonto

q(x) :- PeopleWithHobby(x), like(x,y)

 like⊑PeopleWithHobby

OWL-Gres

q(x) :- like(x,y)


Second target experiment conditions3

Second Target:Experiment conditions

Optimizations


Second target experiment conditions4

Second Target:Experiment conditions

Query containment

We considered:

q(x):- A(x)  q(x) :- A(x),B(x)

We can send to evaluate

q(x):- A(x)

ONLY in Quonto


Second target experiment conditions5

Second Target:Experiment conditions

Optimizations


Second target experiment conditions6

Second Target:Experiment conditions

In-expansion optimizations

q(x):-Man(x),Woman(x)

Consistent Ontology

answer {}

Man ⊑ ¬Woman

ONLY in Quonto


Second target experiment conditions7

Second Target:Experiment conditions

Optimizations


Second target experiment conditions8

Second Target:Experiment conditions

Auxiliar role optimization

  • For A ⊑R.C

Quonto and OWL-Gres

  • It’s introduced an auxiliar role

  • But has no membership assertion

  • We delete all querys with an auxiliar role


Second target experiment conditions9

Second Target:Experiment conditions

Optimizations


Second target experiment conditions10

Second Target:Experiment conditions

Selectivity optimization

ONLY in OWL-Gres

  • A concept, role or concept attribute has no membership assertions

  • We delete all the conjunctive queries with this element

  • It’s correct ?


Second target first comparison

Second Target:First Comparison


Second target first comparison1

Second Target:First Comparison


Second target first comparison2

Second Target:First Comparison


Second target first comparison3

Second Target:First Comparison


Second target quonto abox

Second Target:QuontoAbox

BaseballFanConcept:


Second target quonto abox1

Second Target:QuontoAbox

iscrazyabout Role:


Second target quonto abox2

Second Target:QuontoAbox

e-mail attribute of concept:


Second target owl gres abox

Second Target:OWL-GresAbox

TBOX_name:


Second target owl gres abox1

Second Target:OWL-GresAbox

TBOX_Concept_inclusion :


Second target owl gres abox2

Second Target:OWL-GresAbox

Individual_name :


Second target owl gres abox3

Second Target:OWL-GresAbox

Concept_assertion :


Second target owl gres abox4

Second Target:OWL-GresAbox

Object_Role_assertion:


Second target owl gres abox5

Second Target:OWL-GresAbox

Data_Role_assertion :


Second target second comparison

Second Target:Second Comparison


Second target second comparison1

Second Target:Second Comparison


Second target second comparison2

Second Target:Second Comparison


Second target second comparison3

Second Target:Second Comparison


Second target second comparison4

Second Target:Second Comparison


Conclusions

Conclusions

75 MB

55 MB


  • Login