holger lausen ([email protected])

21
1 Reasoning about Anonymous Resources and Meta Statements on the Semantic Web Guizhen Yang and Michael Kifer Holger Lausen ([email protected])

Upload: henry-mcgowan

Post on 30-Dec-2015

28 views

Category:

Documents


1 download

DESCRIPTION

Reasoning about Anonymous Resources and Meta Statements on the Semantic Web Guizhen Yang and Michael Kifer. Holger Lausen ([email protected]). Content. Preliminaries (RDF, F-Logic, HiLog) Translation (RDF-> „F-Logic“) Comparison RDF MT – “F-Logic” Conclusion. Preliminaries. - PowerPoint PPT Presentation

TRANSCRIPT

1

Reasoning about Anonymous Resources and Meta Statements on the Semantic Web

Guizhen Yang and Michael Kifer

Holger Lausen([email protected])

2

Content

Preliminaries (RDF, F-Logic, HiLog) Translation (RDF-> „F-Logic“) Comparison RDF MT – “F-Logic” Conclusion

3

Preliminaries

RDF, F-Logic, HiLog

4

RDF (in less then 5 min ;)

<predicate> <subject> <object> <predicate> is a property <subject> is a resource <object> is a resource or a literal

labeled graph: {creator, [http://www.w3.org/Home/Lassila], "Ora Lassila"}

5

Anonymous Resources

"http://www.w3.org/Home/Lassila has creator something and something has name Ora Lassila and email [email protected]"

{creator, [http://www.w3.org/Home/Lassila], [X]} {name, [x], „Ora Lassila“} {email, [x], „[email protected]“}

6

Reification

{type, [X], [RDF:Statement]} {predicate, [X], [creator]} {subject, [X], [http://www.w3.org/Home/Lassila]} {object, [X], "Ora Lassila"}

7

F-Logic

Explicit facts: mary[spous->john, children->>{alice, nancy}]. mary[children->>jack]. mary[married].

Inference Rule: X[children->>{C}] :- Y[spouse->X, children->>{C}].

Implicit Facts (generated by above Inference Rule): john[children->>{alice, nancy, jack}]

8

HiLOG

Variables that range over function and predicate symbols Syntax for Reification: a statement can be an

formula and an object at the same time

HiLog Example: X[attributes->>{Attribute}] :- X[Attribute =>Range].

(the set valued attribute „ attributes“ includes all defined Attributes for a particular class)

9

Translations

RDF „F-Logic“

10

Translations

http://foo.org/bulb --> [http://foo.org/bulb] Literals --> „Thomas Edison“

Example: Thomas Edison is the inventor of the bulb {[inventor], [http://foo.org/TheBulb], ”Thomas Edison”} ’http://foo.org/TheBulb’[inventor->>’Thomas Edison’].

11

Translations – anonymous resources Someone, named Thomas Edison, born in 1847, is

the inventor of the resource http://foo.org/TheBulb.

RDF {[name], [X], “Thomas Edison”} {[born], [X], “1847”} {[inventor], [http://foo.org/TheBulb], [X]}

Not In F-Logic, but in Flora-2 _#: unnumbered anonymous ID symbol

(each occurrence is a different ID) _#1: numbered anonymous ID symbol

(_#1 is identical to _#1 within the same scope)

’http://foo.org/TheBulb’[inventor->>_#1], _#1[name->>’Thomas Edison’, born->>’1847’].

12

Translations – reified statements Someone named John Doe believes that a person,

called Thomas Edison, invented the bulb (resource http://foo.org/TheBulb).

RDF {[type], [X], [RDF:Statement]} {[predicate], [X], [inventor]} {[subject], [X], [http://foo.org/TheBulb]} {[object], [X], [Y]} {[name], [Y], ”Thomas Edison”} {[name], [Z], ”John Doe”} {[believes], [Z], [X]}

Flora-2 (like) _#[name->>’John Doe’,

believes->>(’http://foo.org/TheBulb’[inventor->>_#1], _#1[name->>’Thomas Edison’]) ].

13

Formal Syntax and Semantics

(Section 4)

14

Syntax & Semantics

Believe me, there is one ;)

15

RDF MT vs F-Logic

16

Entailment Strict

Does not hold for a proper instance and the graph itself: { john[likes ->> food] } | { #[likes ->> food] } { #[likes ->> food] }

Relaxed Does not hold for a proper instance and the graph itself:

{ john[likes ->> food] } |~ { #[likes ->> food] } { #[likes ->> food]}

Relaxed entailment is defined in RDF MT

Note: A proper instance of a graph is an instance in which a blank node has been replaced by a name, or two blank nodes in the graph have been mapped into the same node in the instance.

The Paper does not make any claims which of the notions is the moreappropriate one.

17

Compositionality of the Semantics RDF MT has to differentiate between merge and union; It claims

both has its suitable application: MERGE: concatinate all triples, but rename identical blan nodes UNION: concatinate all triples, no renaming of blank nodes

Example: Document 1: {[loves], [X], [Mary]} Document 2: {[inventor], [X], [Bulb]} Union:

{[loves], [X], [Mary]} {[inventor], [X], [Bulb]} Merge:

{[loves], [X], [Mary]} {[inventor], [Y], [Bulb]} Thus this document:

{[loves], [John], [Mary]} {[inventor], [ThomasEdison], [Bulb]} Is a entailed by a merge, but not by a union of two graphs

18

Reification and blank nodes? Reification uses blank nodes, thus making twice the same

statement reified, we have two completly different statements (X,Y):

{[type], [X], [RDF:Statement]} {[predicate], [X], [inventor]} {[subject], [X], [http://foo.org/TheBulb]} {[object], [X], [http://foo.org/ThomasEdison]} {believes, [http://xyz.com/John], [X]}

With F-Logic Reification the statement itself has an ID: ’http://foo.org/TheBulb’[inventor->>’http://foo.org/ThomasEdison’][

veracity->>true, authority->>’http://www.britannica.com/’].

Statement <- Statement[veracity->> true].

•{[type], [Y], [RDF:Statement]}•{[predicate], [Y], [inventor]}•{[subject], [Y], [http://foo.org/TheBulb]}•{[object], [Y], [http://foo.org/ThomasEdison]}

19

Conclusion

20

Conclusion of the paper

F-Logic extended with anonymous resources and reification can serve as model theoretic semantics for RDF (as presented)

RDF MT has shortcomings: Non-compositional semantics Weaker then necessary treatment of reification

Identified at least 2 different treatments of RDF graph entailment

Claims F-Logic (rule / frame based) semantics have more possibilities to encode knowledge / have a bigger scope

21

Conclusion for WSM*

Anonymous IDs are useful Nice not to build on RDF MT Further explore reified statements