SlideShare a Scribd company logo
Exchanging OWL 2 QL Knowledge Bases
Vlad Ryzhikov
joint work with E. Botoeva, D. Calvanese and M. Arenas
KRDB Research Centre, Free University of Bozen-Bolzano, Italy
ryzhikov@inf.unibz.it

Talk at University of KwaZulu-Natal, Durban, South Africa

Vlad Ryzhikov

Free University of Bozen-Bolzano

1/16
Knowledge Base Exchange
Problem
given a mapping M between the disjoint signatures Σ and Σ and a source
knowledge base (KB) K, find a target KB K that is a solution for K under
M.
M

Σ
Σ1

Σ2
target signature

source signature
A

T

D

B

T

A

C

B

C

solution

A
source KB K
Vlad Ryzhikov

A
target KB K
Free University of Bozen-Bolzano

2/16
Data Exchange vs. Knowledge Base Exchange

• In Data Exchange (DE) only mappings M (in some scenarios, also

solutions K ) use constraints – in Knowledge Base Exchange (KBE)
source KB K, M, and K use constraints.

Vlad Ryzhikov

Free University of Bozen-Bolzano

3/16
Data Exchange vs. Knowledge Base Exchange

• In Data Exchange (DE) only mappings M (in some scenarios, also

solutions K ) use constraints – in Knowledge Base Exchange (KBE)
source KB K, M, and K use constraints.
• We consider DL-LiteR as the language for the constraints; it is a

formal counterpart of OWL 2 QL standard.

Vlad Ryzhikov

Free University of Bozen-Bolzano

3/16
Data Exchange vs. Knowledge Base Exchange

• In Data Exchange (DE) only mappings M (in some scenarios, also

solutions K ) use constraints – in Knowledge Base Exchange (KBE)
source KB K, M, and K use constraints.
• We consider DL-LiteR as the language for the constraints; it is a

formal counterpart of OWL 2 QL standard.
• Some definitions of solutions in DE apply to KBE, however, KBE

allows for other natural definitions, which are easier to compute.

Vlad Ryzhikov

Free University of Bozen-Bolzano

3/16
Solutions
Solution is a target KB K that “at best” preserves the meaning of a source
KB K w.r.t. a mapping M.

Vlad Ryzhikov

Free University of Bozen-Bolzano

4/16
Solutions
Solution is a target KB K that “at best” preserves the meaning of a source
KB K w.r.t. a mapping M.

Example:
K = {A(a)}, M = {A

Vlad Ryzhikov

A },

Free University of Bozen-Bolzano

K = {A (a)} − solution?

4/16
Solutions
Solution is a target KB K that “at best” preserves the meaning of a source
KB K w.r.t. a mapping M.

Example:
K = {A(a)}, M = {A
K = {A(a), A

B}, M = {A

A },
A ,B

K = {A (a)} − solution?
B }, K = {A (a), B (a)} − solution?
K = {A (a), A

Vlad Ryzhikov

Free University of Bozen-Bolzano

B } − solution?

4/16
Solutions
Solution is a target KB K that “at best” preserves the meaning of a source
KB K w.r.t. a mapping M.

Example:
K = {A(a)}, M = {A
K = {A(a), A

B}, M = {A

A },
A ,B

K = {A (a)} − solution?
B }, K = {A (a), B (a)} − solution?
K = {A (a), A

K = {A

Vlad Ryzhikov

B}, M = {A

A ,B

B },

K = {A

Free University of Bozen-Bolzano

B } − solution?
B } − solution?

4/16
Solutions
Solution is a target KB K that “at best” preserves the meaning of a source
KB K w.r.t. a mapping M.

Example:
K = {A(a)}, M = {A
K = {A(a), A

B}, M = {A

A },
A ,B

K = {A (a)} − solution?
B }, K = {A (a), B (a)} − solution?
K = {A (a), A

K = {A

B}, M = {A

A ,B

B },

K = {A

B } − solution?
B } − solution?

Different definitions make different K above solutions!
Vlad Ryzhikov

Free University of Bozen-Bolzano

4/16
Definitions
DL-LiteR syntax: let a, b, . . . and n, m, . . . be infinite sets of constants and
nulls; Σ be a set of DL-LiteR concept names A and role names P, then
define
basic concepts B ::= A | ∃P | ∃P −
concept inclusions B1
concept disjointness B1

B2

B2

basic roles R ::= P | P −
role inclusions R1

⊥ role disjointness R1

R2
R2

⊥

concept membership B(a), B(n) role membership R(a, b), R(n, m), . . . ,

Vlad Ryzhikov

Free University of Bozen-Bolzano

5/16
Definitions
DL-LiteR syntax: let a, b, . . . and n, m, . . . be infinite sets of constants and
nulls; Σ be a set of DL-LiteR concept names A and role names P, then
define
basic concepts B ::= A | ∃P | ∃P −
concept inclusions B1
concept disjointness B1

B2

B2

basic roles R ::= P | P −
role inclusions R1

⊥ role disjointness R1

R2
R2

⊥

concept membership B(a), B(n) role membership R(a, b), R(n, m), . . . ,
DL-LiteR knowledge base is a set of concept/role inclusions/disjointness
(called TBox) and concept/role membership assertions (called ABox if
without nulls, otherwise extended ABox).

Vlad Ryzhikov

Free University of Bozen-Bolzano

5/16
Definitions
DL-LiteR syntax: let a, b, . . . and n, m, . . . be infinite sets of constants and
nulls; Σ be a set of DL-LiteR concept names A and role names P, then
define
basic concepts B ::= A | ∃P | ∃P −
concept inclusions B1
concept disjointness B1

B2

B2

basic roles R ::= P | P −
role inclusions R1

⊥ role disjointness R1

R2
R2

⊥

concept membership B(a), B(n) role membership R(a, b), R(n, m), . . . ,
DL-LiteR knowledge base is a set of concept/role inclusions/disjointness
(called TBox) and concept/role membership assertions (called ABox if
without nulls, otherwise extended ABox).
Mapping: defined over a pair of disjoint signatures Σ, Σ as the set of
concept inclusions/disjointness, where B is over Σ and B over Σ .

Vlad Ryzhikov

Free University of Bozen-Bolzano

5/16
Definitions
DL-LiteR syntax: let a, b, . . . and n, m, . . . be infinite sets of constants and
nulls; Σ be a set of DL-LiteR concept names A and role names P, then
define
basic concepts B ::= A | ∃P | ∃P −
concept inclusions B1
concept disjointness B1

B2

B2

basic roles R ::= P | P −
role inclusions R1

⊥ role disjointness R1

R2
R2

⊥

concept membership B(a), B(n) role membership R(a, b), R(n, m), . . . ,
DL-LiteR knowledge base is a set of concept/role inclusions/disjointness
(called TBox) and concept/role membership assertions (called ABox if
without nulls, otherwise extended ABox).
Mapping: defined over a pair of disjoint signatures Σ, Σ as the set of
concept inclusions/disjointness, where B is over Σ and B over Σ .
Semantics: standard, no unique name assumption.
Vlad Ryzhikov

Free University of Bozen-Bolzano

5/16
Solutions

We consider three notions:
• Universal solution
Inherited from incomplete data exchage; analogious to model conservative
extentions or Σ-model inseparability

Vlad Ryzhikov

Free University of Bozen-Bolzano

6/16
Solutions

We consider three notions:
• Universal solution
Inherited from incomplete data exchage; analogious to model conservative
extentions or Σ-model inseparability
• Universal UCQ-solution (UCQ = Union of Conjunctive Queries)
Based on what can be extracted from source and target with unions of
conjunctive queries; analogious to query conservative extentions or Σ-query
inseparability

Vlad Ryzhikov

Free University of Bozen-Bolzano

6/16
Solutions

We consider three notions:
• Universal solution
Inherited from incomplete data exchage; analogious to model conservative
extentions or Σ-model inseparability
• Universal UCQ-solution (UCQ = Union of Conjunctive Queries)
Based on what can be extracted from source and target with unions of
conjunctive queries; analogious to query conservative extentions or Σ-query
inseparability
• Representation
Like Universal UCQ-solution, but defined w.r.t. K and K containing only TBox;
uses universal quantification over possible the source and target ABoxes

Vlad Ryzhikov

Free University of Bozen-Bolzano

6/16
Universal Solutions
• Let Mod(K) be the set of models of K.
• Let I, J be a pair of DL-LiteR interpretations over signatures,

respectively Σ and Γ, s.t. Σ ⊆ Γ, then I is said to agree with J on Σ, if
aI = aJ for all constants a;
AI = AJ and P I = P J for all concept and role names A and P from Σ.

Given a Γ interpretation J , denote by agrΣ (J ) the set of all Σ
interpretations that agree with J on Σ; we also use agrΣ (J ), where J
is a set of Σ interpretations.

Vlad Ryzhikov

Free University of Bozen-Bolzano

7/16
Universal Solutions
• Let Mod(K) be the set of models of K.
• Let I, J be a pair of DL-LiteR interpretations over signatures,

respectively Σ and Γ, s.t. Σ ⊆ Γ, then I is said to agree with J on Σ, if
aI = aJ for all constants a;
AI = AJ and P I = P J for all concept and role names A and P from Σ.

Given a Γ interpretation J , denote by agrΣ (J ) the set of all Σ
interpretations that agree with J on Σ; we also use agrΣ (J ), where J
is a set of Σ interpretations.
• Let the mapping M be between the signatures Σ and Σ ; a KB K over

Σ is said to be a universal solution (US) for a KB K over Σ under M if
Mod(K ) = agrΣ (Mod(K ∪ M)).

Vlad Ryzhikov

Free University of Bozen-Bolzano

7/16
Universal Solutions contd.

Mod(K ) = agrΣ (Mod(K ∪ M))
Examples:
K = {A(a)}, M = {A

Vlad Ryzhikov

A },

Free University of Bozen-Bolzano

K = {A (a)} − US

8/16
Universal Solutions contd.

Mod(K ) = agrΣ (Mod(K ∪ M))
Examples:
K = {A(a)}, M = {A
K = {A(a), A

Vlad Ryzhikov

B}, M = {A

A },
A ,B

K = {A (a)} − US
B },

K = {A (a), B (a)} − US
K = {A (a), A
B } − not US

Free University of Bozen-Bolzano

8/16
Universal Solutions contd.

Mod(K ) = agrΣ (Mod(K ∪ M))
Examples:
K = {A(a)}, M = {A
K = {A(a), A

B}, M = {A

K = {∃R(a)}, M = {R

Vlad Ryzhikov

A },
A ,B

R , ∃R −

K = {A (a)} − US
B },

K = {A (a), B (a)} − US
K = {A (a), A
B } − not US

B }, K = {R (a, n), B (n)} − US

Free University of Bozen-Bolzano

8/16
Universal Solutions contd.

Mod(K ) = agrΣ (Mod(K ∪ M))
Examples:
K = {A(a)}, M = {A
K = {A(a), A

B}, M = {A

K = {∃R(a)}, M = {R
K = {A B ⊥,
A(a), B(b)}, M = {A

Vlad Ryzhikov

A },
A ,B

R , ∃R −

A ,B

K = {A (a)} − US
B },

K = {A (a), B (a)} − US
K = {A (a), A
B } − not US

B }, K = {R (a, n), B (n)} − US

B}

Free University of Bozen-Bolzano

− no US exists

8/16
Universal Solutions contd.

US is a fundamental and well-behaved notion in DE, however, in KBE it
has a number of issues:
• Nulls required in ABox, which are not part of OWL 2 QL standard.

Vlad Ryzhikov

Free University of Bozen-Bolzano

9/16
Universal Solutions contd.

US is a fundamental and well-behaved notion in DE, however, in KBE it
has a number of issues:
• Nulls required in ABox, which are not part of OWL 2 QL standard.
• USs cannot contain any non-trivial TBox, i.e., only (extended) ABoxes

can be materialized as the target.

Vlad Ryzhikov

Free University of Bozen-Bolzano

9/16
Universal Solutions contd.

US is a fundamental and well-behaved notion in DE, however, in KBE it
has a number of issues:
• Nulls required in ABox, which are not part of OWL 2 QL standard.
• USs cannot contain any non-trivial TBox, i.e., only (extended) ABoxes

can be materialized as the target.
• USs “very often” do not exists, when the source KB K1 contains

disjointness assertions. Reason: no unique name assumption, as it is
the case in OWL 2 QL.

Vlad Ryzhikov

Free University of Bozen-Bolzano

9/16
Universal UCQ-solutions

Universal UCQ-solution is a “softer” notion of solution, that avoids the
above mentioned issues. It uses UCQs q and certains answers cert(q, K):

Vlad Ryzhikov

Free University of Bozen-Bolzano

10/16
Universal UCQ-solutions

Universal UCQ-solution is a “softer” notion of solution, that avoids the
above mentioned issues. It uses UCQs q and certains answers cert(q, K):
• Let the mapping M be between the signatures Σ and Σ ; a KB K over

Σ is said to be a universal UCQ-solution (UUCQS) for a KB K over Σ
under M if
cert(q, K ) = cert(q, K ∪ M)
for each UCQ q over Σ .

Vlad Ryzhikov

Free University of Bozen-Bolzano

10/16
Universal UCQ-solutions

Universal UCQ-solution is a “softer” notion of solution, that avoids the
above mentioned issues. It uses UCQs q and certains answers cert(q, K):
• Let the mapping M be between the signatures Σ and Σ ; a KB K over

Σ is said to be a universal UCQ-solution (UUCQS) for a KB K over Σ
under M if
cert(q, K ) = cert(q, K ∪ M)
for each UCQ q over Σ .
• if only the inclusion ⊇ in the equation above satisfied, K is called a

UCQ-solution

Vlad Ryzhikov

Free University of Bozen-Bolzano

10/16
Universal UCQ-solutions contd.
cert(q, K ) = cert(q, K ∪ M) for each UCQ q over Σ
Examples:
K = {A(a)}, M = {A

Vlad Ryzhikov

A },

Free University of Bozen-Bolzano

K = {A (a)} − UUCQS

11/16
Universal UCQ-solutions contd.
cert(q, K ) = cert(q, K ∪ M) for each UCQ q over Σ
Examples:
K = {A(a)}, M = {A
K = {A(a), A

Vlad Ryzhikov

B}, M = {A

A },
A ,B

K = {A (a)} − UUCQS
B },

K = {A (a), B (a)} − UUCQS
K = {A (a), A
B } − UUCQS

Free University of Bozen-Bolzano

11/16
Universal UCQ-solutions contd.
cert(q, K ) = cert(q, K ∪ M) for each UCQ q over Σ
Examples:
K = {A(a)}, M = {A
K = {A(a), A

B}, M = {A

K = {∃R(a)}, M = {R

Vlad Ryzhikov

A },
A ,B

R , ∃R −

K = {A (a)} − UUCQS
B },

K = {A (a), B (a)} − UUCQS
K = {A (a), A
B } − UUCQS

B },

Free University of Bozen-Bolzano

K = {∃R (a), ∃R − B }
− UUCQS

11/16
Universal UCQ-solutions contd.
cert(q, K ) = cert(q, K ∪ M) for each UCQ q over Σ
Examples:
K = {A(a)}, M = {A
K = {A(a), A

B}, M = {A

K = {∃R(a)}, M = {R

K = {A B ⊥,
A(a), B(b)}, M = {A

Vlad Ryzhikov

A },
A ,B

R , ∃R −

A ,B

K = {A (a)} − UUCQS
B },

K = {A (a), B (a)} − UUCQS
K = {A (a), A
B } − UUCQS

B },

B}

Free University of Bozen-Bolzano

K = {∃R (a), ∃R − B }
− UUCQS
K = {A (a), B (b)}
−UUCQS

11/16
Universal UCQ-solutions contd.

• UUCQS is a notion of the solution, that is better suited for KBE.

Vlad Ryzhikov

Free University of Bozen-Bolzano

12/16
Universal UCQ-solutions contd.

• UUCQS is a notion of the solution, that is better suited for KBE.
• Still, this notion is dependent on data, i.e., ABox; computing UUCQS

requires processing big amounts of frequently changing data.

Vlad Ryzhikov

Free University of Bozen-Bolzano

12/16
Universal UCQ-solutions contd.

• UUCQS is a notion of the solution, that is better suited for KBE.
• Still, this notion is dependent on data, i.e., ABox; computing UUCQS

requires processing big amounts of frequently changing data.
• UCQ-representation is a notion of the solution, that is not dependent

on data.

Vlad Ryzhikov

Free University of Bozen-Bolzano

12/16
UCQ-representation
For the definition, we need to consider UCQ-solutions over KBs consisting
of only ABoxes. Cosider A = {A(a)} and M = {A A }, then
• A = {A (a), A (b)} - UCQ-solution;
• A = {A (b)} - not UCQ-solution.

Vlad Ryzhikov

Free University of Bozen-Bolzano

13/16
UCQ-representation
For the definition, we need to consider UCQ-solutions over KBs consisting
of only ABoxes. Cosider A = {A(a)} and M = {A A }, then
• A = {A (a), A (b)} - UCQ-solution;
• A = {A (b)} - not UCQ-solution.

Let the mapping M be between the signatures Σ and Σ ; a TBox T over Σ
is said to be UCQ-representaton (UCQR) for a TBox T over Σ under M if
cert(q, T ∪ A ∪ M) =

cert(q, T ∪ A ).
A : A is an ABox over Σ that
is a UCQ-solution for A under M

for
• each UCQ q over Σ ,
• ABox A over Σ
such that T ∪ A is consistent.

Vlad Ryzhikov

Free University of Bozen-Bolzano

13/16
UCQ-representation cont.
cert(q, T ∪ A ∪ M) =

cert(q, T ∪ A )
A : A is an ABox over Σ that
is a UCQ-solution for A under M

for each UCQ q over Σ and ABox A over Σ, such that T ∪ A is consistent.

Examples:
T = {A

Vlad Ryzhikov

A}, M = {A

A },

T = {A

Free University of Bozen-Bolzano

A } − UCQR

14/16
UCQ-representation cont.
cert(q, T ∪ A ∪ M) =

cert(q, T ∪ A )
A : A is an ABox over Σ that
is a UCQ-solution for A under M

for each UCQ q over Σ and ABox A over Σ, such that T ∪ A is consistent.

Examples:
T = {A

A },

T = {A

Vlad Ryzhikov

A}, M = {A
B}, M = {A

A ,B

T = {A
B },

A } − UCQR

T = {A
A } − not UCQR
T = {A
B } − UCQR

Free University of Bozen-Bolzano

14/16
UCQ-representation cont.
cert(q, T ∪ A ∪ M) =

cert(q, T ∪ A )
A : A is an ABox over Σ that
is a UCQ-solution for A under M

for each UCQ q over Σ and ABox A over Σ, such that T ∪ A is consistent.

Examples:
T = {A

A },

T = {A

B}, M = {A

A ,B

T = {A

Vlad Ryzhikov

A}, M = {A

B}, M = {B

T = {A

B },

B },

A } − UCQR

T = {A
A } − not UCQR
T = {A
B } − UCQR

Free University of Bozen-Bolzano

− no UCQR exists

14/16
UCQ-representation cont.
cert(q, T ∪ A ∪ M) =

cert(q, T ∪ A )
A : A is an ABox over Σ that
is a UCQ-solution for A under M

for each UCQ q over Σ and ABox A over Σ, such that T ∪ A is consistent.

Examples:
T = {A

A}, M = {A

A },

T = {A

B}, M = {A

A ,B

T = {A

B}, M = {B

B },

⊥}, M = {A

A ,B

T = {A

Vlad Ryzhikov

B

T = {A
B },

A } − UCQR

T = {A
A } − not UCQR
T = {A
B } − UCQR
− no UCQR exists

B },

T = {A

Free University of Bozen-Bolzano

B

⊥} − UCQR
T = ∅ − UCQR

14/16
Summary of Complexity Results
Membership
Universal solutions
UCQ-representations
Non-emptiness
Universal solutions
UCQ-representations

ABoxes extended ABoxes
in NP
NP-complete
NLogSpace-complete

ABoxes
extended ABoxes
in NP
PSpace-hard, in ExpTime
NLogSpace-complete

• Membership problem: given source KB K1 , target KB K2 , and the

mapping M, decide, if K2 is correct.
• Non-emptyness problem: given source KB K1 and the mapping M,

decide, if there exists a target KB K2 , such that it is correct.

Vlad Ryzhikov

Free University of Bozen-Bolzano

15/16
Summary of Complexity Results
Membership
Universal solutions
UCQ-representations
Non-emptiness
Universal solutions
UCQ-representations

ABoxes extended ABoxes
in NP
NP-complete
NLogSpace-complete

ABoxes
extended ABoxes
in NP
PSpace-hard, in ExpTime
NLogSpace-complete

• Membership problem: given source KB K1 , target KB K2 , and the

mapping M, decide, if K2 is correct.
• Non-emptyness problem: given source KB K1 and the mapping M,

decide, if there exists a target KB K2 , such that it is correct.
• Universal UCQ-solution: membership is PSpace-hard, no other results

yet - future work.

Vlad Ryzhikov

Free University of Bozen-Bolzano

15/16
Thank you
for your attention!

Vlad Ryzhikov

Free University of Bozen-Bolzano

16/16

More Related Content

What's hot

leanCoR: lean Connection-based DL Reasoner
leanCoR: lean Connection-based DL ReasonerleanCoR: lean Connection-based DL Reasoner
leanCoR: lean Connection-based DL Reasoner
Adriano Melo
 
Htb
HtbHtb
Regular Languages
Regular LanguagesRegular Languages
Regular Languages
parmeet834
 
Digital logic design lecture 07
Digital logic design   lecture 07 Digital logic design   lecture 07
Digital logic design lecture 07
FarhatUllah27
 

What's hot (6)

leanCoR: lean Connection-based DL Reasoner
leanCoR: lean Connection-based DL ReasonerleanCoR: lean Connection-based DL Reasoner
leanCoR: lean Connection-based DL Reasoner
 
Htb
HtbHtb
Htb
 
AI Lesson 16
AI Lesson 16AI Lesson 16
AI Lesson 16
 
Regular Languages
Regular LanguagesRegular Languages
Regular Languages
 
Micro teaching junior high school
Micro teaching junior high schoolMicro teaching junior high school
Micro teaching junior high school
 
Digital logic design lecture 07
Digital logic design   lecture 07 Digital logic design   lecture 07
Digital logic design lecture 07
 

Viewers also liked

Random Manhattan Indexing
Random Manhattan IndexingRandom Manhattan Indexing
Random Manhattan Indexing
net2-project
 
Extracting Information for Context-aware Meeting Preparation
Extracting Information for Context-aware Meeting PreparationExtracting Information for Context-aware Meeting Preparation
Extracting Information for Context-aware Meeting Preparation
net2-project
 
Exchanging more than Complete Data
Exchanging more than Complete DataExchanging more than Complete Data
Exchanging more than Complete Datanet2-project
 
Extending DBpedia (LOD) using WikiTables
Extending DBpedia (LOD) using WikiTablesExtending DBpedia (LOD) using WikiTables
Extending DBpedia (LOD) using WikiTablesnet2-project
 
Borders of Decidability in Verification of Data-Centric Dynamic Systems
Borders of Decidability in Verification of Data-Centric Dynamic SystemsBorders of Decidability in Verification of Data-Centric Dynamic Systems
Borders of Decidability in Verification of Data-Centric Dynamic Systemsnet2-project
 
Federation and Navigation in SPARQL 1.1
Federation and Navigation in SPARQL 1.1Federation and Navigation in SPARQL 1.1
Federation and Navigation in SPARQL 1.1net2-project
 

Viewers also liked (7)

Random Manhattan Indexing
Random Manhattan IndexingRandom Manhattan Indexing
Random Manhattan Indexing
 
Extracting Information for Context-aware Meeting Preparation
Extracting Information for Context-aware Meeting PreparationExtracting Information for Context-aware Meeting Preparation
Extracting Information for Context-aware Meeting Preparation
 
Exchanging more than Complete Data
Exchanging more than Complete DataExchanging more than Complete Data
Exchanging more than Complete Data
 
XSPARQL Tutorial
XSPARQL TutorialXSPARQL Tutorial
XSPARQL Tutorial
 
Extending DBpedia (LOD) using WikiTables
Extending DBpedia (LOD) using WikiTablesExtending DBpedia (LOD) using WikiTables
Extending DBpedia (LOD) using WikiTables
 
Borders of Decidability in Verification of Data-Centric Dynamic Systems
Borders of Decidability in Verification of Data-Centric Dynamic SystemsBorders of Decidability in Verification of Data-Centric Dynamic Systems
Borders of Decidability in Verification of Data-Centric Dynamic Systems
 
Federation and Navigation in SPARQL 1.1
Federation and Navigation in SPARQL 1.1Federation and Navigation in SPARQL 1.1
Federation and Navigation in SPARQL 1.1
 

Similar to Exchanging OWL 2 QL Knowledge Bases

Question Answering with Subgraph Embeddings
Question Answering with Subgraph EmbeddingsQuestion Answering with Subgraph Embeddings
Question Answering with Subgraph Embeddings
Karel Ha
 
L03 ai - knowledge representation using logic
L03 ai - knowledge representation using logicL03 ai - knowledge representation using logic
L03 ai - knowledge representation using logic
Manjula V
 
Local Closed World Semantics - DL 2011 Poster
Local Closed World Semantics - DL 2011 PosterLocal Closed World Semantics - DL 2011 Poster
Local Closed World Semantics - DL 2011 PosterAdila Krisnadhi
 
Csr2011 june17 15_15_kaminski
Csr2011 june17 15_15_kaminskiCsr2011 june17 15_15_kaminski
Csr2011 june17 15_15_kaminskiCSR2011
 
KR12 Semantic Index and TBox optimisation with respect to dependencies
KR12 Semantic Index and TBox optimisation with respect to dependenciesKR12 Semantic Index and TBox optimisation with respect to dependencies
KR12 Semantic Index and TBox optimisation with respect to dependencies
Mariano Rodriguez-Muro
 
Factors That Influence High Scores in Higher Secondary Examination - A Nano T...
Factors That Influence High Scores in Higher Secondary Examination - A Nano T...Factors That Influence High Scores in Higher Secondary Examination - A Nano T...
Factors That Influence High Scores in Higher Secondary Examination - A Nano T...
IJERA Editor
 
Ch5_Constraint Satisfaction Problems_Lecture 1.pdf
Ch5_Constraint Satisfaction Problems_Lecture 1.pdfCh5_Constraint Satisfaction Problems_Lecture 1.pdf
Ch5_Constraint Satisfaction Problems_Lecture 1.pdf
YanKyaw6
 
Oopsla04.ppt
Oopsla04.pptOopsla04.ppt
Csr2011 june14 15_45_musatov
Csr2011 june14 15_45_musatovCsr2011 june14 15_45_musatov
Csr2011 june14 15_45_musatov
CSR2011
 
On_Mapping_Relational_Databases_to_RDF_and_SHACL.pdf
On_Mapping_Relational_Databases_to_RDF_and_SHACL.pdfOn_Mapping_Relational_Databases_to_RDF_and_SHACL.pdf
On_Mapping_Relational_Databases_to_RDF_and_SHACL.pdf
Department of Informatics, University of Oslo
 
Dscrete structure
Dscrete  structureDscrete  structure
Dscrete structure
Abdur Rehman
 
Сергей Кольцов —НИУ ВШЭ —ICBDA 2015
Сергей Кольцов —НИУ ВШЭ —ICBDA 2015Сергей Кольцов —НИУ ВШЭ —ICBDA 2015
Сергей Кольцов —НИУ ВШЭ —ICBDA 2015
rusbase
 
Dialectica and Kolmogorov Problems
Dialectica and Kolmogorov ProblemsDialectica and Kolmogorov Problems
Dialectica and Kolmogorov Problems
Valeria de Paiva
 
conference_presentation (1)
conference_presentation (1)conference_presentation (1)
conference_presentation (1)Brandon Van Over
 
Discrete mathematics notes
Discrete mathematics notesDiscrete mathematics notes
Discrete mathematics notes
josephndengeyingoma
 
[ABDO] Logic As A Database Language
[ABDO] Logic As A Database Language[ABDO] Logic As A Database Language
[ABDO] Logic As A Database LanguageCarles Farré
 
Csr2011 june17 11_30_vyalyi
Csr2011 june17 11_30_vyalyiCsr2011 june17 11_30_vyalyi
Csr2011 june17 11_30_vyalyiCSR2011
 
Top school in delhi
Top school in delhiTop school in delhi
Top school in delhi
Edhole.com
 
Volume computation and applications
Volume computation and applications Volume computation and applications
Volume computation and applications
Vissarion Fisikopoulos
 
Publishing Math Lecture Notes as Linked Data
Publishing Math Lecture Notes as Linked DataPublishing Math Lecture Notes as Linked Data
Publishing Math Lecture Notes as Linked Data
Christoph Lange
 

Similar to Exchanging OWL 2 QL Knowledge Bases (20)

Question Answering with Subgraph Embeddings
Question Answering with Subgraph EmbeddingsQuestion Answering with Subgraph Embeddings
Question Answering with Subgraph Embeddings
 
L03 ai - knowledge representation using logic
L03 ai - knowledge representation using logicL03 ai - knowledge representation using logic
L03 ai - knowledge representation using logic
 
Local Closed World Semantics - DL 2011 Poster
Local Closed World Semantics - DL 2011 PosterLocal Closed World Semantics - DL 2011 Poster
Local Closed World Semantics - DL 2011 Poster
 
Csr2011 june17 15_15_kaminski
Csr2011 june17 15_15_kaminskiCsr2011 june17 15_15_kaminski
Csr2011 june17 15_15_kaminski
 
KR12 Semantic Index and TBox optimisation with respect to dependencies
KR12 Semantic Index and TBox optimisation with respect to dependenciesKR12 Semantic Index and TBox optimisation with respect to dependencies
KR12 Semantic Index and TBox optimisation with respect to dependencies
 
Factors That Influence High Scores in Higher Secondary Examination - A Nano T...
Factors That Influence High Scores in Higher Secondary Examination - A Nano T...Factors That Influence High Scores in Higher Secondary Examination - A Nano T...
Factors That Influence High Scores in Higher Secondary Examination - A Nano T...
 
Ch5_Constraint Satisfaction Problems_Lecture 1.pdf
Ch5_Constraint Satisfaction Problems_Lecture 1.pdfCh5_Constraint Satisfaction Problems_Lecture 1.pdf
Ch5_Constraint Satisfaction Problems_Lecture 1.pdf
 
Oopsla04.ppt
Oopsla04.pptOopsla04.ppt
Oopsla04.ppt
 
Csr2011 june14 15_45_musatov
Csr2011 june14 15_45_musatovCsr2011 june14 15_45_musatov
Csr2011 june14 15_45_musatov
 
On_Mapping_Relational_Databases_to_RDF_and_SHACL.pdf
On_Mapping_Relational_Databases_to_RDF_and_SHACL.pdfOn_Mapping_Relational_Databases_to_RDF_and_SHACL.pdf
On_Mapping_Relational_Databases_to_RDF_and_SHACL.pdf
 
Dscrete structure
Dscrete  structureDscrete  structure
Dscrete structure
 
Сергей Кольцов —НИУ ВШЭ —ICBDA 2015
Сергей Кольцов —НИУ ВШЭ —ICBDA 2015Сергей Кольцов —НИУ ВШЭ —ICBDA 2015
Сергей Кольцов —НИУ ВШЭ —ICBDA 2015
 
Dialectica and Kolmogorov Problems
Dialectica and Kolmogorov ProblemsDialectica and Kolmogorov Problems
Dialectica and Kolmogorov Problems
 
conference_presentation (1)
conference_presentation (1)conference_presentation (1)
conference_presentation (1)
 
Discrete mathematics notes
Discrete mathematics notesDiscrete mathematics notes
Discrete mathematics notes
 
[ABDO] Logic As A Database Language
[ABDO] Logic As A Database Language[ABDO] Logic As A Database Language
[ABDO] Logic As A Database Language
 
Csr2011 june17 11_30_vyalyi
Csr2011 june17 11_30_vyalyiCsr2011 june17 11_30_vyalyi
Csr2011 june17 11_30_vyalyi
 
Top school in delhi
Top school in delhiTop school in delhi
Top school in delhi
 
Volume computation and applications
Volume computation and applications Volume computation and applications
Volume computation and applications
 
Publishing Math Lecture Notes as Linked Data
Publishing Math Lecture Notes as Linked DataPublishing Math Lecture Notes as Linked Data
Publishing Math Lecture Notes as Linked Data
 

More from net2-project

Vector spaces for information extraction - Random Projection Example
Vector spaces for information extraction - Random Projection ExampleVector spaces for information extraction - Random Projection Example
Vector spaces for information extraction - Random Projection Example
net2-project
 
Mining Semi-structured Data: Understanding Web-tables – Building a Taxonomy f...
Mining Semi-structured Data: Understanding Web-tables – Building a Taxonomy f...Mining Semi-structured Data: Understanding Web-tables – Building a Taxonomy f...
Mining Semi-structured Data: Understanding Web-tables – Building a Taxonomy f...net2-project
 
Tailoring Temporal Description Logics for Reasoning over Temporal Conceptual ...
Tailoring Temporal Description Logics for Reasoning over Temporal Conceptual ...Tailoring Temporal Description Logics for Reasoning over Temporal Conceptual ...
Tailoring Temporal Description Logics for Reasoning over Temporal Conceptual ...
net2-project
 
Managing Social Communities
Managing Social CommunitiesManaging Social Communities
Managing Social Communitiesnet2-project
 
Data Exchange over RDF
Data Exchange over RDFData Exchange over RDF
Data Exchange over RDFnet2-project
 
Exchanging More than Complete Data
Exchanging More than Complete DataExchanging More than Complete Data
Exchanging More than Complete Datanet2-project
 
Exchanging More than Complete Data
Exchanging More than Complete DataExchanging More than Complete Data
Exchanging More than Complete Datanet2-project
 
Answer-set programming
Answer-set programmingAnswer-set programming
Answer-set programmingnet2-project
 
Evolving web, evolving search
Evolving web, evolving searchEvolving web, evolving search
Evolving web, evolving searchnet2-project
 
SPARQL1.1 Tutorial, given in UChile by Axel Polleres (DERI)
SPARQL1.1 Tutorial, given in UChile by Axel Polleres (DERI)SPARQL1.1 Tutorial, given in UChile by Axel Polleres (DERI)
SPARQL1.1 Tutorial, given in UChile by Axel Polleres (DERI)
net2-project
 

More from net2-project (10)

Vector spaces for information extraction - Random Projection Example
Vector spaces for information extraction - Random Projection ExampleVector spaces for information extraction - Random Projection Example
Vector spaces for information extraction - Random Projection Example
 
Mining Semi-structured Data: Understanding Web-tables – Building a Taxonomy f...
Mining Semi-structured Data: Understanding Web-tables – Building a Taxonomy f...Mining Semi-structured Data: Understanding Web-tables – Building a Taxonomy f...
Mining Semi-structured Data: Understanding Web-tables – Building a Taxonomy f...
 
Tailoring Temporal Description Logics for Reasoning over Temporal Conceptual ...
Tailoring Temporal Description Logics for Reasoning over Temporal Conceptual ...Tailoring Temporal Description Logics for Reasoning over Temporal Conceptual ...
Tailoring Temporal Description Logics for Reasoning over Temporal Conceptual ...
 
Managing Social Communities
Managing Social CommunitiesManaging Social Communities
Managing Social Communities
 
Data Exchange over RDF
Data Exchange over RDFData Exchange over RDF
Data Exchange over RDF
 
Exchanging More than Complete Data
Exchanging More than Complete DataExchanging More than Complete Data
Exchanging More than Complete Data
 
Exchanging More than Complete Data
Exchanging More than Complete DataExchanging More than Complete Data
Exchanging More than Complete Data
 
Answer-set programming
Answer-set programmingAnswer-set programming
Answer-set programming
 
Evolving web, evolving search
Evolving web, evolving searchEvolving web, evolving search
Evolving web, evolving search
 
SPARQL1.1 Tutorial, given in UChile by Axel Polleres (DERI)
SPARQL1.1 Tutorial, given in UChile by Axel Polleres (DERI)SPARQL1.1 Tutorial, given in UChile by Axel Polleres (DERI)
SPARQL1.1 Tutorial, given in UChile by Axel Polleres (DERI)
 

Recently uploaded

Palestine last event orientationfvgnh .pptx
Palestine last event orientationfvgnh .pptxPalestine last event orientationfvgnh .pptx
Palestine last event orientationfvgnh .pptx
RaedMohamed3
 
Guidance_and_Counselling.pdf B.Ed. 4th Semester
Guidance_and_Counselling.pdf B.Ed. 4th SemesterGuidance_and_Counselling.pdf B.Ed. 4th Semester
Guidance_and_Counselling.pdf B.Ed. 4th Semester
Atul Kumar Singh
 
Language Across the Curriculm LAC B.Ed.
Language Across the  Curriculm LAC B.Ed.Language Across the  Curriculm LAC B.Ed.
Language Across the Curriculm LAC B.Ed.
Atul Kumar Singh
 
Mule 4.6 & Java 17 Upgrade | MuleSoft Mysore Meetup #46
Mule 4.6 & Java 17 Upgrade | MuleSoft Mysore Meetup #46Mule 4.6 & Java 17 Upgrade | MuleSoft Mysore Meetup #46
Mule 4.6 & Java 17 Upgrade | MuleSoft Mysore Meetup #46
MysoreMuleSoftMeetup
 
Unit 2- Research Aptitude (UGC NET Paper I).pdf
Unit 2- Research Aptitude (UGC NET Paper I).pdfUnit 2- Research Aptitude (UGC NET Paper I).pdf
Unit 2- Research Aptitude (UGC NET Paper I).pdf
Thiyagu K
 
Francesca Gottschalk - How can education support child empowerment.pptx
Francesca Gottschalk - How can education support child empowerment.pptxFrancesca Gottschalk - How can education support child empowerment.pptx
Francesca Gottschalk - How can education support child empowerment.pptx
EduSkills OECD
 
How to Make a Field invisible in Odoo 17
How to Make a Field invisible in Odoo 17How to Make a Field invisible in Odoo 17
How to Make a Field invisible in Odoo 17
Celine George
 
Sha'Carri Richardson Presentation 202345
Sha'Carri Richardson Presentation 202345Sha'Carri Richardson Presentation 202345
Sha'Carri Richardson Presentation 202345
beazzy04
 
Adversarial Attention Modeling for Multi-dimensional Emotion Regression.pdf
Adversarial Attention Modeling for Multi-dimensional Emotion Regression.pdfAdversarial Attention Modeling for Multi-dimensional Emotion Regression.pdf
Adversarial Attention Modeling for Multi-dimensional Emotion Regression.pdf
Po-Chuan Chen
 
Introduction to AI for Nonprofits with Tapp Network
Introduction to AI for Nonprofits with Tapp NetworkIntroduction to AI for Nonprofits with Tapp Network
Introduction to AI for Nonprofits with Tapp Network
TechSoup
 
Supporting (UKRI) OA monographs at Salford.pptx
Supporting (UKRI) OA monographs at Salford.pptxSupporting (UKRI) OA monographs at Salford.pptx
Supporting (UKRI) OA monographs at Salford.pptx
Jisc
 
Operation Blue Star - Saka Neela Tara
Operation Blue Star   -  Saka Neela TaraOperation Blue Star   -  Saka Neela Tara
Operation Blue Star - Saka Neela Tara
Balvir Singh
 
Welcome to TechSoup New Member Orientation and Q&A (May 2024).pdf
Welcome to TechSoup   New Member Orientation and Q&A (May 2024).pdfWelcome to TechSoup   New Member Orientation and Q&A (May 2024).pdf
Welcome to TechSoup New Member Orientation and Q&A (May 2024).pdf
TechSoup
 
The basics of sentences session 5pptx.pptx
The basics of sentences session 5pptx.pptxThe basics of sentences session 5pptx.pptx
The basics of sentences session 5pptx.pptx
heathfieldcps1
 
"Protectable subject matters, Protection in biotechnology, Protection of othe...
"Protectable subject matters, Protection in biotechnology, Protection of othe..."Protectable subject matters, Protection in biotechnology, Protection of othe...
"Protectable subject matters, Protection in biotechnology, Protection of othe...
SACHIN R KONDAGURI
 
Lapbook sobre os Regimes Totalitários.pdf
Lapbook sobre os Regimes Totalitários.pdfLapbook sobre os Regimes Totalitários.pdf
Lapbook sobre os Regimes Totalitários.pdf
Jean Carlos Nunes Paixão
 
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
siemaillard
 
2024.06.01 Introducing a competency framework for languag learning materials ...
2024.06.01 Introducing a competency framework for languag learning materials ...2024.06.01 Introducing a competency framework for languag learning materials ...
2024.06.01 Introducing a competency framework for languag learning materials ...
Sandy Millin
 
TESDA TM1 REVIEWER FOR NATIONAL ASSESSMENT WRITTEN AND ORAL QUESTIONS WITH A...
TESDA TM1 REVIEWER  FOR NATIONAL ASSESSMENT WRITTEN AND ORAL QUESTIONS WITH A...TESDA TM1 REVIEWER  FOR NATIONAL ASSESSMENT WRITTEN AND ORAL QUESTIONS WITH A...
TESDA TM1 REVIEWER FOR NATIONAL ASSESSMENT WRITTEN AND ORAL QUESTIONS WITH A...
EugeneSaldivar
 
Synthetic Fiber Construction in lab .pptx
Synthetic Fiber Construction in lab .pptxSynthetic Fiber Construction in lab .pptx
Synthetic Fiber Construction in lab .pptx
Pavel ( NSTU)
 

Recently uploaded (20)

Palestine last event orientationfvgnh .pptx
Palestine last event orientationfvgnh .pptxPalestine last event orientationfvgnh .pptx
Palestine last event orientationfvgnh .pptx
 
Guidance_and_Counselling.pdf B.Ed. 4th Semester
Guidance_and_Counselling.pdf B.Ed. 4th SemesterGuidance_and_Counselling.pdf B.Ed. 4th Semester
Guidance_and_Counselling.pdf B.Ed. 4th Semester
 
Language Across the Curriculm LAC B.Ed.
Language Across the  Curriculm LAC B.Ed.Language Across the  Curriculm LAC B.Ed.
Language Across the Curriculm LAC B.Ed.
 
Mule 4.6 & Java 17 Upgrade | MuleSoft Mysore Meetup #46
Mule 4.6 & Java 17 Upgrade | MuleSoft Mysore Meetup #46Mule 4.6 & Java 17 Upgrade | MuleSoft Mysore Meetup #46
Mule 4.6 & Java 17 Upgrade | MuleSoft Mysore Meetup #46
 
Unit 2- Research Aptitude (UGC NET Paper I).pdf
Unit 2- Research Aptitude (UGC NET Paper I).pdfUnit 2- Research Aptitude (UGC NET Paper I).pdf
Unit 2- Research Aptitude (UGC NET Paper I).pdf
 
Francesca Gottschalk - How can education support child empowerment.pptx
Francesca Gottschalk - How can education support child empowerment.pptxFrancesca Gottschalk - How can education support child empowerment.pptx
Francesca Gottschalk - How can education support child empowerment.pptx
 
How to Make a Field invisible in Odoo 17
How to Make a Field invisible in Odoo 17How to Make a Field invisible in Odoo 17
How to Make a Field invisible in Odoo 17
 
Sha'Carri Richardson Presentation 202345
Sha'Carri Richardson Presentation 202345Sha'Carri Richardson Presentation 202345
Sha'Carri Richardson Presentation 202345
 
Adversarial Attention Modeling for Multi-dimensional Emotion Regression.pdf
Adversarial Attention Modeling for Multi-dimensional Emotion Regression.pdfAdversarial Attention Modeling for Multi-dimensional Emotion Regression.pdf
Adversarial Attention Modeling for Multi-dimensional Emotion Regression.pdf
 
Introduction to AI for Nonprofits with Tapp Network
Introduction to AI for Nonprofits with Tapp NetworkIntroduction to AI for Nonprofits with Tapp Network
Introduction to AI for Nonprofits with Tapp Network
 
Supporting (UKRI) OA monographs at Salford.pptx
Supporting (UKRI) OA monographs at Salford.pptxSupporting (UKRI) OA monographs at Salford.pptx
Supporting (UKRI) OA monographs at Salford.pptx
 
Operation Blue Star - Saka Neela Tara
Operation Blue Star   -  Saka Neela TaraOperation Blue Star   -  Saka Neela Tara
Operation Blue Star - Saka Neela Tara
 
Welcome to TechSoup New Member Orientation and Q&A (May 2024).pdf
Welcome to TechSoup   New Member Orientation and Q&A (May 2024).pdfWelcome to TechSoup   New Member Orientation and Q&A (May 2024).pdf
Welcome to TechSoup New Member Orientation and Q&A (May 2024).pdf
 
The basics of sentences session 5pptx.pptx
The basics of sentences session 5pptx.pptxThe basics of sentences session 5pptx.pptx
The basics of sentences session 5pptx.pptx
 
"Protectable subject matters, Protection in biotechnology, Protection of othe...
"Protectable subject matters, Protection in biotechnology, Protection of othe..."Protectable subject matters, Protection in biotechnology, Protection of othe...
"Protectable subject matters, Protection in biotechnology, Protection of othe...
 
Lapbook sobre os Regimes Totalitários.pdf
Lapbook sobre os Regimes Totalitários.pdfLapbook sobre os Regimes Totalitários.pdf
Lapbook sobre os Regimes Totalitários.pdf
 
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
 
2024.06.01 Introducing a competency framework for languag learning materials ...
2024.06.01 Introducing a competency framework for languag learning materials ...2024.06.01 Introducing a competency framework for languag learning materials ...
2024.06.01 Introducing a competency framework for languag learning materials ...
 
TESDA TM1 REVIEWER FOR NATIONAL ASSESSMENT WRITTEN AND ORAL QUESTIONS WITH A...
TESDA TM1 REVIEWER  FOR NATIONAL ASSESSMENT WRITTEN AND ORAL QUESTIONS WITH A...TESDA TM1 REVIEWER  FOR NATIONAL ASSESSMENT WRITTEN AND ORAL QUESTIONS WITH A...
TESDA TM1 REVIEWER FOR NATIONAL ASSESSMENT WRITTEN AND ORAL QUESTIONS WITH A...
 
Synthetic Fiber Construction in lab .pptx
Synthetic Fiber Construction in lab .pptxSynthetic Fiber Construction in lab .pptx
Synthetic Fiber Construction in lab .pptx
 

Exchanging OWL 2 QL Knowledge Bases

  • 1. Exchanging OWL 2 QL Knowledge Bases Vlad Ryzhikov joint work with E. Botoeva, D. Calvanese and M. Arenas KRDB Research Centre, Free University of Bozen-Bolzano, Italy ryzhikov@inf.unibz.it Talk at University of KwaZulu-Natal, Durban, South Africa Vlad Ryzhikov Free University of Bozen-Bolzano 1/16
  • 2. Knowledge Base Exchange Problem given a mapping M between the disjoint signatures Σ and Σ and a source knowledge base (KB) K, find a target KB K that is a solution for K under M. M Σ Σ1 Σ2 target signature source signature A T D B T A C B C solution A source KB K Vlad Ryzhikov A target KB K Free University of Bozen-Bolzano 2/16
  • 3. Data Exchange vs. Knowledge Base Exchange • In Data Exchange (DE) only mappings M (in some scenarios, also solutions K ) use constraints – in Knowledge Base Exchange (KBE) source KB K, M, and K use constraints. Vlad Ryzhikov Free University of Bozen-Bolzano 3/16
  • 4. Data Exchange vs. Knowledge Base Exchange • In Data Exchange (DE) only mappings M (in some scenarios, also solutions K ) use constraints – in Knowledge Base Exchange (KBE) source KB K, M, and K use constraints. • We consider DL-LiteR as the language for the constraints; it is a formal counterpart of OWL 2 QL standard. Vlad Ryzhikov Free University of Bozen-Bolzano 3/16
  • 5. Data Exchange vs. Knowledge Base Exchange • In Data Exchange (DE) only mappings M (in some scenarios, also solutions K ) use constraints – in Knowledge Base Exchange (KBE) source KB K, M, and K use constraints. • We consider DL-LiteR as the language for the constraints; it is a formal counterpart of OWL 2 QL standard. • Some definitions of solutions in DE apply to KBE, however, KBE allows for other natural definitions, which are easier to compute. Vlad Ryzhikov Free University of Bozen-Bolzano 3/16
  • 6. Solutions Solution is a target KB K that “at best” preserves the meaning of a source KB K w.r.t. a mapping M. Vlad Ryzhikov Free University of Bozen-Bolzano 4/16
  • 7. Solutions Solution is a target KB K that “at best” preserves the meaning of a source KB K w.r.t. a mapping M. Example: K = {A(a)}, M = {A Vlad Ryzhikov A }, Free University of Bozen-Bolzano K = {A (a)} − solution? 4/16
  • 8. Solutions Solution is a target KB K that “at best” preserves the meaning of a source KB K w.r.t. a mapping M. Example: K = {A(a)}, M = {A K = {A(a), A B}, M = {A A }, A ,B K = {A (a)} − solution? B }, K = {A (a), B (a)} − solution? K = {A (a), A Vlad Ryzhikov Free University of Bozen-Bolzano B } − solution? 4/16
  • 9. Solutions Solution is a target KB K that “at best” preserves the meaning of a source KB K w.r.t. a mapping M. Example: K = {A(a)}, M = {A K = {A(a), A B}, M = {A A }, A ,B K = {A (a)} − solution? B }, K = {A (a), B (a)} − solution? K = {A (a), A K = {A Vlad Ryzhikov B}, M = {A A ,B B }, K = {A Free University of Bozen-Bolzano B } − solution? B } − solution? 4/16
  • 10. Solutions Solution is a target KB K that “at best” preserves the meaning of a source KB K w.r.t. a mapping M. Example: K = {A(a)}, M = {A K = {A(a), A B}, M = {A A }, A ,B K = {A (a)} − solution? B }, K = {A (a), B (a)} − solution? K = {A (a), A K = {A B}, M = {A A ,B B }, K = {A B } − solution? B } − solution? Different definitions make different K above solutions! Vlad Ryzhikov Free University of Bozen-Bolzano 4/16
  • 11. Definitions DL-LiteR syntax: let a, b, . . . and n, m, . . . be infinite sets of constants and nulls; Σ be a set of DL-LiteR concept names A and role names P, then define basic concepts B ::= A | ∃P | ∃P − concept inclusions B1 concept disjointness B1 B2 B2 basic roles R ::= P | P − role inclusions R1 ⊥ role disjointness R1 R2 R2 ⊥ concept membership B(a), B(n) role membership R(a, b), R(n, m), . . . , Vlad Ryzhikov Free University of Bozen-Bolzano 5/16
  • 12. Definitions DL-LiteR syntax: let a, b, . . . and n, m, . . . be infinite sets of constants and nulls; Σ be a set of DL-LiteR concept names A and role names P, then define basic concepts B ::= A | ∃P | ∃P − concept inclusions B1 concept disjointness B1 B2 B2 basic roles R ::= P | P − role inclusions R1 ⊥ role disjointness R1 R2 R2 ⊥ concept membership B(a), B(n) role membership R(a, b), R(n, m), . . . , DL-LiteR knowledge base is a set of concept/role inclusions/disjointness (called TBox) and concept/role membership assertions (called ABox if without nulls, otherwise extended ABox). Vlad Ryzhikov Free University of Bozen-Bolzano 5/16
  • 13. Definitions DL-LiteR syntax: let a, b, . . . and n, m, . . . be infinite sets of constants and nulls; Σ be a set of DL-LiteR concept names A and role names P, then define basic concepts B ::= A | ∃P | ∃P − concept inclusions B1 concept disjointness B1 B2 B2 basic roles R ::= P | P − role inclusions R1 ⊥ role disjointness R1 R2 R2 ⊥ concept membership B(a), B(n) role membership R(a, b), R(n, m), . . . , DL-LiteR knowledge base is a set of concept/role inclusions/disjointness (called TBox) and concept/role membership assertions (called ABox if without nulls, otherwise extended ABox). Mapping: defined over a pair of disjoint signatures Σ, Σ as the set of concept inclusions/disjointness, where B is over Σ and B over Σ . Vlad Ryzhikov Free University of Bozen-Bolzano 5/16
  • 14. Definitions DL-LiteR syntax: let a, b, . . . and n, m, . . . be infinite sets of constants and nulls; Σ be a set of DL-LiteR concept names A and role names P, then define basic concepts B ::= A | ∃P | ∃P − concept inclusions B1 concept disjointness B1 B2 B2 basic roles R ::= P | P − role inclusions R1 ⊥ role disjointness R1 R2 R2 ⊥ concept membership B(a), B(n) role membership R(a, b), R(n, m), . . . , DL-LiteR knowledge base is a set of concept/role inclusions/disjointness (called TBox) and concept/role membership assertions (called ABox if without nulls, otherwise extended ABox). Mapping: defined over a pair of disjoint signatures Σ, Σ as the set of concept inclusions/disjointness, where B is over Σ and B over Σ . Semantics: standard, no unique name assumption. Vlad Ryzhikov Free University of Bozen-Bolzano 5/16
  • 15. Solutions We consider three notions: • Universal solution Inherited from incomplete data exchage; analogious to model conservative extentions or Σ-model inseparability Vlad Ryzhikov Free University of Bozen-Bolzano 6/16
  • 16. Solutions We consider three notions: • Universal solution Inherited from incomplete data exchage; analogious to model conservative extentions or Σ-model inseparability • Universal UCQ-solution (UCQ = Union of Conjunctive Queries) Based on what can be extracted from source and target with unions of conjunctive queries; analogious to query conservative extentions or Σ-query inseparability Vlad Ryzhikov Free University of Bozen-Bolzano 6/16
  • 17. Solutions We consider three notions: • Universal solution Inherited from incomplete data exchage; analogious to model conservative extentions or Σ-model inseparability • Universal UCQ-solution (UCQ = Union of Conjunctive Queries) Based on what can be extracted from source and target with unions of conjunctive queries; analogious to query conservative extentions or Σ-query inseparability • Representation Like Universal UCQ-solution, but defined w.r.t. K and K containing only TBox; uses universal quantification over possible the source and target ABoxes Vlad Ryzhikov Free University of Bozen-Bolzano 6/16
  • 18. Universal Solutions • Let Mod(K) be the set of models of K. • Let I, J be a pair of DL-LiteR interpretations over signatures, respectively Σ and Γ, s.t. Σ ⊆ Γ, then I is said to agree with J on Σ, if aI = aJ for all constants a; AI = AJ and P I = P J for all concept and role names A and P from Σ. Given a Γ interpretation J , denote by agrΣ (J ) the set of all Σ interpretations that agree with J on Σ; we also use agrΣ (J ), where J is a set of Σ interpretations. Vlad Ryzhikov Free University of Bozen-Bolzano 7/16
  • 19. Universal Solutions • Let Mod(K) be the set of models of K. • Let I, J be a pair of DL-LiteR interpretations over signatures, respectively Σ and Γ, s.t. Σ ⊆ Γ, then I is said to agree with J on Σ, if aI = aJ for all constants a; AI = AJ and P I = P J for all concept and role names A and P from Σ. Given a Γ interpretation J , denote by agrΣ (J ) the set of all Σ interpretations that agree with J on Σ; we also use agrΣ (J ), where J is a set of Σ interpretations. • Let the mapping M be between the signatures Σ and Σ ; a KB K over Σ is said to be a universal solution (US) for a KB K over Σ under M if Mod(K ) = agrΣ (Mod(K ∪ M)). Vlad Ryzhikov Free University of Bozen-Bolzano 7/16
  • 20. Universal Solutions contd. Mod(K ) = agrΣ (Mod(K ∪ M)) Examples: K = {A(a)}, M = {A Vlad Ryzhikov A }, Free University of Bozen-Bolzano K = {A (a)} − US 8/16
  • 21. Universal Solutions contd. Mod(K ) = agrΣ (Mod(K ∪ M)) Examples: K = {A(a)}, M = {A K = {A(a), A Vlad Ryzhikov B}, M = {A A }, A ,B K = {A (a)} − US B }, K = {A (a), B (a)} − US K = {A (a), A B } − not US Free University of Bozen-Bolzano 8/16
  • 22. Universal Solutions contd. Mod(K ) = agrΣ (Mod(K ∪ M)) Examples: K = {A(a)}, M = {A K = {A(a), A B}, M = {A K = {∃R(a)}, M = {R Vlad Ryzhikov A }, A ,B R , ∃R − K = {A (a)} − US B }, K = {A (a), B (a)} − US K = {A (a), A B } − not US B }, K = {R (a, n), B (n)} − US Free University of Bozen-Bolzano 8/16
  • 23. Universal Solutions contd. Mod(K ) = agrΣ (Mod(K ∪ M)) Examples: K = {A(a)}, M = {A K = {A(a), A B}, M = {A K = {∃R(a)}, M = {R K = {A B ⊥, A(a), B(b)}, M = {A Vlad Ryzhikov A }, A ,B R , ∃R − A ,B K = {A (a)} − US B }, K = {A (a), B (a)} − US K = {A (a), A B } − not US B }, K = {R (a, n), B (n)} − US B} Free University of Bozen-Bolzano − no US exists 8/16
  • 24. Universal Solutions contd. US is a fundamental and well-behaved notion in DE, however, in KBE it has a number of issues: • Nulls required in ABox, which are not part of OWL 2 QL standard. Vlad Ryzhikov Free University of Bozen-Bolzano 9/16
  • 25. Universal Solutions contd. US is a fundamental and well-behaved notion in DE, however, in KBE it has a number of issues: • Nulls required in ABox, which are not part of OWL 2 QL standard. • USs cannot contain any non-trivial TBox, i.e., only (extended) ABoxes can be materialized as the target. Vlad Ryzhikov Free University of Bozen-Bolzano 9/16
  • 26. Universal Solutions contd. US is a fundamental and well-behaved notion in DE, however, in KBE it has a number of issues: • Nulls required in ABox, which are not part of OWL 2 QL standard. • USs cannot contain any non-trivial TBox, i.e., only (extended) ABoxes can be materialized as the target. • USs “very often” do not exists, when the source KB K1 contains disjointness assertions. Reason: no unique name assumption, as it is the case in OWL 2 QL. Vlad Ryzhikov Free University of Bozen-Bolzano 9/16
  • 27. Universal UCQ-solutions Universal UCQ-solution is a “softer” notion of solution, that avoids the above mentioned issues. It uses UCQs q and certains answers cert(q, K): Vlad Ryzhikov Free University of Bozen-Bolzano 10/16
  • 28. Universal UCQ-solutions Universal UCQ-solution is a “softer” notion of solution, that avoids the above mentioned issues. It uses UCQs q and certains answers cert(q, K): • Let the mapping M be between the signatures Σ and Σ ; a KB K over Σ is said to be a universal UCQ-solution (UUCQS) for a KB K over Σ under M if cert(q, K ) = cert(q, K ∪ M) for each UCQ q over Σ . Vlad Ryzhikov Free University of Bozen-Bolzano 10/16
  • 29. Universal UCQ-solutions Universal UCQ-solution is a “softer” notion of solution, that avoids the above mentioned issues. It uses UCQs q and certains answers cert(q, K): • Let the mapping M be between the signatures Σ and Σ ; a KB K over Σ is said to be a universal UCQ-solution (UUCQS) for a KB K over Σ under M if cert(q, K ) = cert(q, K ∪ M) for each UCQ q over Σ . • if only the inclusion ⊇ in the equation above satisfied, K is called a UCQ-solution Vlad Ryzhikov Free University of Bozen-Bolzano 10/16
  • 30. Universal UCQ-solutions contd. cert(q, K ) = cert(q, K ∪ M) for each UCQ q over Σ Examples: K = {A(a)}, M = {A Vlad Ryzhikov A }, Free University of Bozen-Bolzano K = {A (a)} − UUCQS 11/16
  • 31. Universal UCQ-solutions contd. cert(q, K ) = cert(q, K ∪ M) for each UCQ q over Σ Examples: K = {A(a)}, M = {A K = {A(a), A Vlad Ryzhikov B}, M = {A A }, A ,B K = {A (a)} − UUCQS B }, K = {A (a), B (a)} − UUCQS K = {A (a), A B } − UUCQS Free University of Bozen-Bolzano 11/16
  • 32. Universal UCQ-solutions contd. cert(q, K ) = cert(q, K ∪ M) for each UCQ q over Σ Examples: K = {A(a)}, M = {A K = {A(a), A B}, M = {A K = {∃R(a)}, M = {R Vlad Ryzhikov A }, A ,B R , ∃R − K = {A (a)} − UUCQS B }, K = {A (a), B (a)} − UUCQS K = {A (a), A B } − UUCQS B }, Free University of Bozen-Bolzano K = {∃R (a), ∃R − B } − UUCQS 11/16
  • 33. Universal UCQ-solutions contd. cert(q, K ) = cert(q, K ∪ M) for each UCQ q over Σ Examples: K = {A(a)}, M = {A K = {A(a), A B}, M = {A K = {∃R(a)}, M = {R K = {A B ⊥, A(a), B(b)}, M = {A Vlad Ryzhikov A }, A ,B R , ∃R − A ,B K = {A (a)} − UUCQS B }, K = {A (a), B (a)} − UUCQS K = {A (a), A B } − UUCQS B }, B} Free University of Bozen-Bolzano K = {∃R (a), ∃R − B } − UUCQS K = {A (a), B (b)} −UUCQS 11/16
  • 34. Universal UCQ-solutions contd. • UUCQS is a notion of the solution, that is better suited for KBE. Vlad Ryzhikov Free University of Bozen-Bolzano 12/16
  • 35. Universal UCQ-solutions contd. • UUCQS is a notion of the solution, that is better suited for KBE. • Still, this notion is dependent on data, i.e., ABox; computing UUCQS requires processing big amounts of frequently changing data. Vlad Ryzhikov Free University of Bozen-Bolzano 12/16
  • 36. Universal UCQ-solutions contd. • UUCQS is a notion of the solution, that is better suited for KBE. • Still, this notion is dependent on data, i.e., ABox; computing UUCQS requires processing big amounts of frequently changing data. • UCQ-representation is a notion of the solution, that is not dependent on data. Vlad Ryzhikov Free University of Bozen-Bolzano 12/16
  • 37. UCQ-representation For the definition, we need to consider UCQ-solutions over KBs consisting of only ABoxes. Cosider A = {A(a)} and M = {A A }, then • A = {A (a), A (b)} - UCQ-solution; • A = {A (b)} - not UCQ-solution. Vlad Ryzhikov Free University of Bozen-Bolzano 13/16
  • 38. UCQ-representation For the definition, we need to consider UCQ-solutions over KBs consisting of only ABoxes. Cosider A = {A(a)} and M = {A A }, then • A = {A (a), A (b)} - UCQ-solution; • A = {A (b)} - not UCQ-solution. Let the mapping M be between the signatures Σ and Σ ; a TBox T over Σ is said to be UCQ-representaton (UCQR) for a TBox T over Σ under M if cert(q, T ∪ A ∪ M) = cert(q, T ∪ A ). A : A is an ABox over Σ that is a UCQ-solution for A under M for • each UCQ q over Σ , • ABox A over Σ such that T ∪ A is consistent. Vlad Ryzhikov Free University of Bozen-Bolzano 13/16
  • 39. UCQ-representation cont. cert(q, T ∪ A ∪ M) = cert(q, T ∪ A ) A : A is an ABox over Σ that is a UCQ-solution for A under M for each UCQ q over Σ and ABox A over Σ, such that T ∪ A is consistent. Examples: T = {A Vlad Ryzhikov A}, M = {A A }, T = {A Free University of Bozen-Bolzano A } − UCQR 14/16
  • 40. UCQ-representation cont. cert(q, T ∪ A ∪ M) = cert(q, T ∪ A ) A : A is an ABox over Σ that is a UCQ-solution for A under M for each UCQ q over Σ and ABox A over Σ, such that T ∪ A is consistent. Examples: T = {A A }, T = {A Vlad Ryzhikov A}, M = {A B}, M = {A A ,B T = {A B }, A } − UCQR T = {A A } − not UCQR T = {A B } − UCQR Free University of Bozen-Bolzano 14/16
  • 41. UCQ-representation cont. cert(q, T ∪ A ∪ M) = cert(q, T ∪ A ) A : A is an ABox over Σ that is a UCQ-solution for A under M for each UCQ q over Σ and ABox A over Σ, such that T ∪ A is consistent. Examples: T = {A A }, T = {A B}, M = {A A ,B T = {A Vlad Ryzhikov A}, M = {A B}, M = {B T = {A B }, B }, A } − UCQR T = {A A } − not UCQR T = {A B } − UCQR Free University of Bozen-Bolzano − no UCQR exists 14/16
  • 42. UCQ-representation cont. cert(q, T ∪ A ∪ M) = cert(q, T ∪ A ) A : A is an ABox over Σ that is a UCQ-solution for A under M for each UCQ q over Σ and ABox A over Σ, such that T ∪ A is consistent. Examples: T = {A A}, M = {A A }, T = {A B}, M = {A A ,B T = {A B}, M = {B B }, ⊥}, M = {A A ,B T = {A Vlad Ryzhikov B T = {A B }, A } − UCQR T = {A A } − not UCQR T = {A B } − UCQR − no UCQR exists B }, T = {A Free University of Bozen-Bolzano B ⊥} − UCQR T = ∅ − UCQR 14/16
  • 43. Summary of Complexity Results Membership Universal solutions UCQ-representations Non-emptiness Universal solutions UCQ-representations ABoxes extended ABoxes in NP NP-complete NLogSpace-complete ABoxes extended ABoxes in NP PSpace-hard, in ExpTime NLogSpace-complete • Membership problem: given source KB K1 , target KB K2 , and the mapping M, decide, if K2 is correct. • Non-emptyness problem: given source KB K1 and the mapping M, decide, if there exists a target KB K2 , such that it is correct. Vlad Ryzhikov Free University of Bozen-Bolzano 15/16
  • 44. Summary of Complexity Results Membership Universal solutions UCQ-representations Non-emptiness Universal solutions UCQ-representations ABoxes extended ABoxes in NP NP-complete NLogSpace-complete ABoxes extended ABoxes in NP PSpace-hard, in ExpTime NLogSpace-complete • Membership problem: given source KB K1 , target KB K2 , and the mapping M, decide, if K2 is correct. • Non-emptyness problem: given source KB K1 and the mapping M, decide, if there exists a target KB K2 , such that it is correct. • Universal UCQ-solution: membership is PSpace-hard, no other results yet - future work. Vlad Ryzhikov Free University of Bozen-Bolzano 15/16
  • 45. Thank you for your attention! Vlad Ryzhikov Free University of Bozen-Bolzano 16/16