Transcript The future of databases

The future of databases
DBDM 07/08
Applications in future data basing: High-level property
driven content retrieval through RDF and OWL
Leiden
Bas van den Berg, Patrick van Kouteren, Rosa Meijer, Mathijs Sanders
5 april 2016
1
Vermelding onderdeel organisatie
History
• Resource Description Framework (RDF) is published in
1999
• Used to define statements/properties about a resource
• Web Ontology Language (OWL) roots start in the year
2001
• OWL is an ontology language defining description of
classes, along with their related properties and
instances
• OWL can be used to process different resources to
infer knowledge
2
Long-Range Research Goal
• Relating and integrating various data by means of
standardized descriptions
• Have a standardized format for easy access querying
• Uniform way of searching through any type of content
3
Usefulness
• Making information on the web processible for automated
computer applications
• Property-driven query results
• Comparing multimedia objects by specific properties
• Problem of ambiguous object names resolved
• Better data integration due to standardization
4
Example: Detection of Audio Fingerprint
Compare with
metadata in the
RDF
Extracted from
short music fragment
Precomputed
fingerprint in RDF
5
Performance Measure
• Qualitative: user experience (after improvement, ask
the same people for their opinion)
• Performance: Hard to measure, one could depict if a
certain query on the RDFs return the expected result
as a performance of the current system
6
Current state of the art
•
RDF: Currently more applications make use of RDF
(RSS tools, Musicbrainz, Bioinformatics), but is not yet
a well known technology
•
OWL: used for inferring knowledge through logical
rules about resources (even interdisciplinary
resources)
•
POWDER: a protocol for publishing descriptions of
(e.g. metadata about) web resources using RDF, OWL
and HTTP
7
Intermediate milestones
• Increase of usage RDF and OWL
• Metadata of multimedia objects, like audio or video
signals
• Usage of OWL for knowledge inference
• The design for algorithms and usage of OWL for
interdisciplinary (different resources) knowledge
inference. One can think of finding similar images in a
video to the input image
8
Main challenges
• Introducing the new technology must be done
gradually. It should be backwards compatible at the
beginning, but it also has to offer sufficient
functionality to be useful and attractive to use
• Adapting the RDF, OWL and POWDER standards to fit
the needs of the market
9
Conclusion
• In the future one can use an efficient and propertydriven query for searching for resources through RDF
• The use of these metadata formatting technologies are
appearing more often in applications and different
fields of research
• RDF can supply the needs of the current areas which
are boosting at this very moment (YouTube, Flickr,
BLAST, ..). The abstract, but descriptive approach of
RDF is highly suitable for these type of areas
10
References
S.Decker, S.Melnik, D. Fensel, et al.: The Semantic
Web, IEEE Computing october 2000
•
N. Shadbolt, W. Hall, T. Berners-Lee: The
Semantic Web revisited, IEEE june 2006
•
J. Pan, I. Horrocks: Adding customised datatypes
into OWL, Elsevier july 2005
•
C.Sanin, E. Szczerbicki, C. Toro: An OWL Ontology
of Set of Experience Knowledge Structure, Journal of
Universal Computer Science vol. 13 february 2007
•
11