Provenance Challenge: A Semantic Web Approach

woodruffpassionateInternet και Εφαρμογές Web

21 Οκτ 2013 (πριν από 3 χρόνια και 9 μήνες)

80 εμφανίσεις

1

Provenance Challenge:

A Semantic Web Approach

Jennifer Golbeck

MINDSWAP

University of Maryland, College Park

2

The Semantic Web


Ontologies are used to create schemas


Instances (the data) can be stored anywhere


Referenced by its URI


Semantics of Web Ontology Language (OWL)
allow for logical inferences on the data


Much support for aggregation, querying, and
browsing

3

Ontologies for Workflows


Need to represent data about


Execution of workflow


Execution of services


Information about files


Information about dependencies


http://provenance.mindswap.org/provenance.owl

4

Provenance.owl

5

Web Services for Process
Execution


Web services receive inputs


URIs of input files


The service retrieves the files from the web,
perform their action, and output the resulting
files to the web (so they have a URI)


Image/Headers/Graphics, etc


Metadata (using the ontology) describing


The provanance of the output files


The service execution

6

Metadata Aggregation


Central service is pinged or spiders to
collect metadata files from service
executions


Collection can be queried and browsed


Metadata can be generated anywhere
and stored anywhere


http://provenance.mindswap.org

7

Reasoning over the Data


The

big benefit of the Semantic Web


The semantics of properties used (e.g.
transitive, symmetric, etc) allow us to trace
back through the provenance history easily


Additional rules can be added to further
augment existing semantics


E.g. The Service Execution ancestors of a file are
all the Service Executions that produced files upon
which the file in question is dependent

8

Querying with SPARQL


SPARQL:
S
PARQL
P
rotocol
a
nd
R
DF
Q
uery
L
anguage



Find the outputs of align_warp where the inputs
are annotated with center=UChicago.


PREFIX rdf: <http://www.w3.org/1999/02/22
-
rdf
-
syntax
-
ns#>

PREFIX prov: <http://provenance.mindswap.org/provenance.owl#>

PREFIX rdfs: <http://www.w3.org/2000/01/rdf
-
schema#>

SELECT DISTINCT ?y WHERE {


?y

prov:hasInputFile ?x;



prov:serviceUsed prov:align_warp.



?x

prov:annotation "center=UChicago".

}



Query results are returned as RDF documents

9

Thanks To


Daniel Krech, creator of Redfoot, the
package used on the website, and who
helped get all the Semantic Web tool
support running


David Wang, who helped with the rules
and visualization support