In this paper we present GOSPL, which stands for Grounding Ontologies with Social Processes and Natural Language. GOSPL is a method and tool that supports stakeholders in iteratively interpreting and modeling their common hybrid ontologies using their own terminology for semantic interoperability between autonomously developed and maintained information systems. Hybrid ontologies are ontologies in which concepts are both formally and informally described with the help of a special linguistic resource called glossary. Social interactions between the community members drive the ontology evolution process and result in more stable and agreed upon ontologies.
Christophe Debruyne, Robert Meersman: GOSPL: A Method and Tool for Fact-Oriented Hybrid Ontology Engineering. ADBIS 2012: 153-166
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
GOSPL: A Method and Tool for Fact-Oriented Hybrid Ontology Engineering
1. 19/09/12
GOSPL:
A
METHOD
AND
TOOL
FOR
HYBRID
ONTOLOGY
ENGINEERING
Christophe
Debruyne
and
Robert
Meersman
September
2012
@
ADBIS
1
2. Introduction
• Informa@on
systems
(IS)
on
the
Web
are
in
general
developed
and
maintained
autonomously
19/09/12
• For
IS
to
interoperate
seman@cally,
an
ontology
is
needed
• Agreement
among
all
the
stakeholders
• Ontologies
evolve
while
the
agreements
are
developed
• Ontologies
are
an
externaliza@on
of
the
seman@cs
outside
an
IS
• The
problem
is
not
so
much
what
ontologies
in
computer
science
are,
but
how
they
come
to
be.
In
other
words,
ontology
engineering
(OE)
is
a
cri@cal
ac@vity
2
3. Introduction
• But
ontology
engineering
methods
also
need
adequate
tool
support
19/09/12
• An
examina@on
of
related
work
showed
that
most
frameworks
did
not
• Take
into
account
a
special
role
for
informal
defini@ons
AND
• Had
tool
support
tailored
to
a
method
or
framework
AND
• Supported
the
users
in
their
elicita@on
and
agreement
processes
3
4. Contribution
• Presenta@on
of
a
method
based
on
a
framework
for
hybrid
ontology
engineering
(presented
@
ADBIS
2012)
19/09/12
• Development
of
the
GOSPL
prototype
currently
in
use
in
a
Linked
Data
project
for
Brussels
• GOSPL
stands
for
Grounding
Ontologies
with
Social
Processes
and
Natural
Language
4
5. Method:
Framework
• DOGMA
Hybrid
Ontology
Descrip@ons
<Ω,
ci,
K,
G>
• Ω
a
lexon
base,
a
finite
set
of
plausible
binary
fact
types
called
19/09/12
lexons,
e.g.,
<Vendor
Community,
Offer,
has,
is
of,
Title>
• ci
a
func@on
mapping
community-‐iden@fiers
and
terms
to
concepts
• K
a
finite
set
of
ontological
commitments
containing
• A
selec@on
of
lexons
• A
mapping
from
applica@on
symbols
to
ontology
terms
• Predicates
over
those
terms
and
roles
to
express
constraints
• G
is
a
glossary,
a
triple
with
components
• Gloss,
a
set
of
linguis@c,
human-‐interpretable
glosses
• g1,
mapping
community-‐term
pairs
to
glosses
• g2,
mapping
lexons
to
glosses
5
6. Method:
Framework
• Example
of
an
applica@on-‐commitment
19/09/12
• Ω-‐RIDL:
Verheyden
et
al.
(SWDB
2004),
6
Trog
et
al.
(RuleML
2007)
7. Method:
GOSPL
• Grounding
ontologies
with
social
processes
&
NL
• Hybrid
Ontology
Engineering
Method
19/09/12
7
8. Hybrid
Ontology
Engineering
Method
• A)
Managing
communi@es
• >1
representa@ves
of
autonomously
developed
and
maintained
19/09/12
informa@on
systems
that
need
to
interoperate
seman@cally
• B)
Managing
the
Seman@c
Interoperability
Requirements
• Set
of
key-‐terms
• Set
of
goals
• C)
Ar@cula@on
of
terms
with
glosses
• Star@ng
with
key-‐terms
• Alignment
8
9. Hybrid
Ontology
Engineering
Method
• D)
Crea@ons
of
lexons
• At
least
1
of
the
terms
needs
to
be
ar@culated
19/09/12
• E)
Crea@on
of
constraints
• An
engagement
of
the
community
members
to
comply
with
agreed
upon
constraints
in
their
applica@on-‐commitments
• Focus
on
reference
structures
“No
en@ty
without
iden@ty”
• F)
Crea@on
of
a
commitment
• Steps
A
à
E
mostly
within
one
community
• F
done
by
the
stakeholder
9
10. Hybrid
Ontology
Engineering
Method
• G)
Crea@on
of
gloss-‐equivalences
• Assert
that
two
glosses
refer
to
the
same
concept
19/09/12
• H)
Crea@on
of
synonyms
(at
level
of
lexons)
• Assert
that
two
labels
refer
to
the
same
concept
• Community
used
for
disambigua@on
• Why
this
dis@nc@on
• Independent
agreements
• Glossary-‐consistency
principle
used
a
means
for
driving
agreements
(for
every
community-‐term
pair,
if
the
glosses
used
to
ar@culate
these
terms
were
deemed
referring
to
the
same
10
concept,
then
the
labels
should
be
deemed
synonyms)
14. Tool
• Use
of
a
quasi-‐anonymous
vo@ng
system
• See
who
has
voted,
but
not
what
(cfr.
Dotmocracy)
19/09/12
• Outcomes
of
“off-‐line”
mee@ngs
(face-‐to-‐face,
teleconference,
etc…)
need
to
be
summarized
in
the
tool
• Looking
for
counter
examples
while
making
statements
• Applica@on
of
NLP
techniques
to
dis@ll
facts
from
glosses
• Exploita@on
of
the
commitments
• Link
with
mul@lingual
terminology
base
14
15. Application
• Used
in
the
context
of
a
Linked
Data
Project
in
Brussels
hmp://www.oscb.be/
19/09/12
{image, features} Picture
Image Datasets
JPSearch Input Dataset
Query Format
Flickr Knowledge collaborates
Recognition Management with others
Phone API in
Server ... Platform
JPSearch Output ...
Query Format output
Stakeholder
SPARQL
Other Datasets
Ontology
owns
Geonames
used for Agenda.be
annotation Information
DBPedia System
input
Agenda.be R2RML DB
triplifies
15
16. Discussion
• GOSPL
is
…
• Teachable
19/09/12
• Repeatable
• Traceable
• The
three
characteris@cs
of
a
method
• Usability
study
of
an
experiment
• 43
par@cipants
• Ciuciu,
O.,
Debruyne,
C.
(2012)
Assessing
the
User
Sa@sfac@on
with
an
Ontology
Engineering
Tool
based
on
Social
Processes.
In
Proceedings
of
On
the
Move
to
Meaningful
Internet
Systems
2012:
OTM
Workshops,
LNCS,
Springer
16
17. Future
work
• Mining
interac@ons
for
user
profiling
• Giving
appropriate
tools
to
different
users
19/09/12
• Early
experiment
gave
mo@va@on
• Capturing
other
means
for
discussions
• Omogenia
(Liapis
et
al.)
• Analysis
of
gloss-‐evolu@on
and
its
impact
on
the
formal
defini@ons
(to
be
reported
elsewhere)
17
18. 19/09/12
Thank
you!
Contact
chrdebru@vub.ac.be
for
more
informa@on
on
accessing
the
tool.
18