Transcript Semantic Web for Museums - Australian National University

COMP 6703 eScience Project
Semantic Web for Museums
•Student : Lei Junran
•Client/Technical Supervisor : Tom Worthington
•Academic Supervisor : Peter Strazdins
•Period : 2006 Semester 1
What is in my presentation
• Motivation
• Objectives
• Technologies
• Design Considerations
• Demonstration
• Conclusion
• Future Work
Motivation - Constraints
• Constrains of Current Museums
Collections Management Methods
– Natural features of cultural
collections — Rich associations
• eg, creator of painting A had other paintings with
the same style, which originates from another
artist, who drew painting B with the same topic…
– Collections are preserved as isolated
objects in individual museums
Museums System Example
Museums System Example
Museums System Example
Motivation - Solution
• The emerging semantic web
technology (W3C Semantic Web)
would be proposed to solve the
constraints and provide a better way
for cultural heritage preservation and
management.
Project Objectives
•
Current Objective - to develop an
effective semantic web archive
system for museums.
• Long Terms - research the promising
semantic technology for creating the
knowledge management network
among museums.
TechnologiesWhat is Semantic Web
• Tim Berners-Lee's original web vision
involved more than retrieving
Hypertext Markup Language (HTML)
pages from Web servers.
• Make the web a more collaborative
medium.
• Create a web of data that machines
can process
How to make Semantic Web
possible?
• Make the data smarter.
– application-independent, easily
discovered, to be described with
concrete relationships…
Four Levels of smart data
• Text Documents and Database Records
– Data just can be used in a single application
• XML documents using single vocabulary
– Data is now smart enough to move between
applications in this museum.
• XML documents with mixed vocabularies
– Data can be composed from multiple
museums or institutes
Four Levels of smart data
• Ontologies and rules
– data is now smart enough to be
described with concrete
relationships
– new data can be inferred from
existing data by following logical
rules
Semantic Web Elements and
technologies
• Metadata
• XML
• RDF
• Ontology
Metadata
• Meta-data: meaning of data values;
• Example:
DATA
John Smith
222 Happy Lane
META DATA
Name
Address
XML
• XML(Extensible Markup Language) is
the syntactic foundation layer of the
Semantic Web.
• Provides a simple, standard syntax
for encoding the meaning of data
values, or meta data.
• Example:
<author>
<name> John Smith </name>
<address> 222 Happy Lane </address>
</author>
XML Metadata benefits
• All data are described with a set of
predefined vocabulary and syntax.
• Enable exchange, interoperability,
information integration and
application independence.
RDF
• The resource described in RDF could
be identified by URI. The statement
about resource is combined of three
elements, or triple.
&ns;/location/
Greece
Subject
locateAt
Predicate
&ns;/location/
Europe
Object
RDF/XML Data Example
<swm:location rdf : about = "&ns; /
location / Greece">
<swm:locationAt rdf:resource = "&ns;
/ location / Europe"/>
</swm:location>
What are included in Ontology?
• Classes: Object, Activity, Location
• Relationships: object <locate at> location,
•
•
•
•
company <is a > organization
Properties: Identifier(cardinality 1:1),
Type, Creator
Constrains and Rules: If X is true, then Y
must also be true.
Functions and Process:
A formal vocabulary (defined terms) for all
above
Ontology Languages
• Ontology is represented in knowledge
representation languages
– RDFS (lightweight ontology)
• Elements: Class, label, subclassOf, Property,
Domain, range, type, subPropertyof…
– OWL (Robust ontology)
• Elements: RDFS plus someValuesFrom ∃,
allValuesFrom ∀, hasValue ∋, minCardinality ≥,
cardinality =, intersectionOf, unionOf…
Why Use Ontology
• defines the domain vocabulary.
• Improve association expression,
interoperability
• Ontology languages are backed by a
rigorous formal logic, which makes
the ontology machine-interpretable.
Semantic Levels Summary
•
Semantic Levels (Redrawn after C. Daconta, et al 2003)
Design Considerations
• Use existing ontology
– CIDOC CRM
• CIDOC:
The International Committee for
Documentation of the International Council
of Museums
• CRM: Conceptual Reference Model
• A domain ontology for cultural heritage
information
Design Considerations
• Use existing metadata standard
– Dublin Core
• A simple yet effective element set for
describing a wide range of networked
resources.
• Simplicity, Commonly understood
semantics, Extensibility
• Example Elements: Identifier,
Description, Format, Date, Creator…
CIDOC CRM
• Advantages
– Comprehensive and widely accepted
– Mappings have been established with
major metadata standards
• Disadvantages
– Includes 81 classes and 132 properties
– Vocabulary is too detailed to be used as
metadata directly
Solutions
• Use subset of CRM
• Use Dublin Core Metadata Standard
• Redesign the vocabulary of the
applied subset when DC can not
express the meaning of the subset.
• Use DC and subset vocabulary (SWM
vocabulary) as metadata
Example of CRM
Example Mixed Use of DC and
SWM Vocabulary
<swm:activity rdf : about = “ &basens;activity
/Textile Lengths 85-1002 Production">
<DC:type>production</DC:type>
<DC:identifier>Textile Lengths 85-1002
Production </DC:identifier>
<swm:beginDate>1984</swm:beginDate>
<swm:endDate>1985</swm:endDate>
<swm:locateAt rdf : resource = "&basens;
location/Ngkwarlerlaneme camp"/>
</ swm:activity>
Elements Relationships
System Architecture
Demonstration
Conclusion
• A semantic web prototype system has
been developed
• A RDF Schema has been designed
• The museums collections could be
input and transferred to RDF data
for preservation
Conclusion
• Data is now smart enough to be
described with concrete relationships
• RDF data output and Batch input
increases the interoperability with
other semantic systems and provide a
convenient transfer way to existing
data.
Review the four levels of smart data
• Ontologies and rules
– data is now smart enough to be
described with concrete
relationships
– new data can be inferred from
existing data by following logical
rules
Half way of the fourth level
• Reasons
– Use RDFS (lightweight ontology
language);
– Use subset of ontology, the
relationships is not rich enough.
– No enough constrains, rules and
associations to infer.
Future Work
• Redesign Ontology using robust
ontology language (eg. OWL)
• Add more constrains and rules for
inference
• Design system showing more benefits
of semantic web technology
• Web Services and Taxonomies in
Semantic Web.