1. Collaborative Ontology
Engineering and Management
May 20-24, 2013
The Sheraton San Diego Hotel &
Marina
San Diego, California, USA
The 2013 International Conference on
Collaboration Technologies and Systems
(CTS 2013)
Michele Missikoff
Polytechnic University of Marche and LEKS-CNR, Italy
2. Content
• What is an ontology? Why we need them?
• The Semantic Web and Social Semantic
Networks
• On the nature of (computational) knowledge
• Conceptual modeling: principles
• From perception to representation
• Ontology Engineering
• The social dimension of Ontology Building
And ...
• Some practical exercises
2CTS 2013, San Diego
3. The Speaker
Michele Missikoff
• Scientific Advisor at Univ Polytechnic of
Marche (Ancona), for the European BIVEE1
Project
• Coordinator of Lab for Enterprise Knowledge
and Systems, Italian National Research Council
• European Task Force Leader of FInES - Future
Internet Enterprise Systems Research
Roadmap
• Professor of Enterprise Information Systems at
International University of Rome
3CTS 2013, San Diego
(1Business Innovation in Virtual Enteprise Environments)
4. 4
What is an ontology? What is the
Semantic Web? Why we need
them?
Ontology Introduction
5. Ontology: Origins and History
Ontology Introduction 5
• In Philosophy,
fundamental branch of
metaphysics
– Studies “being” or
“existence” and their basic
categories
– Aims to find out what
entities and types of
entities exist
– Identifies and characterises
their properties
(Credits: I. Horrocks)
6. Ontology Introduction 6
What is a Computational Ontology?
From Philosophy to practical use of an Ontology
– It is about what exists, and is relevant for our
purposes, in our domain of interest;
– Needs the consensus of a group which is
representative of the community of interest
– Aims at reaching a shared view of the domain of
interest
– Allows for reduction or elimination of
terminological and conceptual confusion
An ontology is an evolving repository of relevant
concepts, continuously incorporating new meanings
from the interaction with the environment
7. Ontology Introduction 7
An Ontology is …
“… a theory about the nature of beings” (Philosophical view)
“… a formal, explicit specification of a shared
conceptualisation.” (AI view – T.R. Gruber)*
– ‘Formal' refers to the fact that the ontology should be
machine understandable.
– 'Explicit' means that the type of concepts used and the
constraints on their use are explicitly and fully defined.
– 'Shared' reflects the notion that ontology captures
consensual knowledge, that is, it is not restricted to some
individual, but accepted by a group / community.
– A 'conceptualisation' refers to an abstract model of some
phenomena in the world, it identifies the relevant concepts
related to that phenomena.
• In formal terms: Ont = (Conc, Rel, Axioms, Inst)
8. Ontology Introduction 8
An Ontology is … (con’d)
"An ontology defines the common terms and
concepts (meaning) used to describe and
represent an area of knowledge. An ontology can
range in expressivity from a Taxonomy
(knowledge with minimal hierarchy or a
parent/child structure), to a Thesaurus (words
and synonyms), to a Conceptual Model (with
more complex knowledge), to a Logical Theory
(with very rich, complex, consistent and
meaningful knowledge)." [www.omg.org]
9. Ontology Introduction 9
Conceptual models and ontologies
They have common roots, but ...
Conceptual Model
• Traditionally conceived for inter-human communication
• Typically in diagrammatic form (e.g., UML, BPMN)
• Semi-formal representation
– Formal syntax, but intuitive semantics
• Used with a precise goal (e.g., IS engineering)
Computational Ontologies
• Conceived to be ‘fully’ processed by a computer
• Linear (textual) form (supports equivalent diagrammatic
forms)
• Typically represented with a formal language (e.g., RDF(S),
OWL, CG, F-Logic, ...)
• Used to represent an application domain, not a specific system
10. Ontology Introduction 10
Ontologies as Social Artefacts
• An Ontology is a socio-cultural phenomenon, but we
want to describe the concepts in a formal and
unambiguous way, processable by a computer
An ontology contains:
– a set of concepts (e.g., entities, attributes, processes) seen as
relevant in a given domain
– the definitions and inter-relationships among these concepts
– set of Axioms (e.g., constraints) and, in case, instances
• To be used by computers, ontologies must
– have precise definitions, with a formal semantics (Tarski)
– evolve according to an evolving reality and adapt to current
needs and usage of both human users and computers
– be supported by an Ontology Management System
12. Ontology Introduction 12
First Motivations
When starting a cooperation (to work together
or in interacting in social settings), people
and organizations may have different:
– viewpoints
– assumptions
– needs
about the same domain, due to different
contexts, goals, backgrounds and cultures
13. 13
motivations (cont’d)
Furthermore, the frequent use of different:
– jargon
– terminology
sometimes diverging or overlapping, generate confusion.
Even worse,
– concepts
may be mismatched or ill-defined (e.g., delivery_date).
Goal
allow people, organizations, computer applications, smart
objects to effectively cooperate, despite the mentioned
differences
• All computers today can communicate, but it does not imply
that they cooperate (due to different services & data
organization)
• People and organizations do communicate and cooperate, but
with low automatic support (and several misunderstandings)
14. Ontology Introduction 14
Cooperation Problems
The lack of a shared understanding leads to a poor
communication that impacts on:
– effectiveness of people’s cooperation
– flaws in enterprise operations
– even social fragmentation (… tension)
When Information Systems Engineering is involved, further
problems arise on:
– the identification of the requirements for the system
specification
– potential reuse and sharing of system components
– interoperability among systems
Then … ONTOLOGIES
15. Ontology Introduction 15
Benefits of Reference Ontologies
• Business Opportunity analysis
• Partnering
• Interoperability
• Semantic Knowledge Management
• Business / IT Alignment
• Social / Shared vision
By means of
• A collaboration practice for a shared context
understanding
• Ontology management – Building an EO
• Semantic Annotation
• Interoperability among legacy systems
• Sem Search: exact / approximate
• Similarity reasoning
17. Ontology Introduction 17
Progression of Domain
specification
Lexicon - Set of terms (also multi-word) representing
relevant entities and relationships in the domain
Glossary - Alphabetically ordered terms, with their
descriptions, in natural language. First
categorizations according to an Ontology
Framework (e.g., OPAL)
Taxonomy - hierarchy of terms according to a
refinement relation (e.g., ISA)
Thesaurus - First introduction of elationships, such as:
synonyms, antonyms; BT, NT, RT
Semantic Net - Full fledged deployment of Concepts
and Relations: Gen/Spec, part of/HasPart, Sim,
InstOf, … + domRel
20. 20
The Knowedge Society
• European Council: Lisbon Strategy for growth
and jobs
“Europe needs will achieve the largest and most
competitive knowledge-based economy in the planet”
• Investing in knowledge and innovation is
intended to spur the EU's transition to a
knowledge-based and creative economy.
• The "fifth freedom" – the free movement of
knowledge – should thus be established
• Knowledge is a value if embodied in models and
practices of the Society and Production
systems (…New Economy).
(europa.eu/legislation_summaries/employment_and_social_policy/eu2020/growt
h_and_jobs/c11806_en.htm)
Ontology Introduction
21. World is Changing...
... and we need new:
• Systems of values
• Development models
• Social relationships to guarantee
sustainability at:
– Social, economic, environmental levels
Ontology Introduction 21
22. 22
The Advent of the Semantic Web
The collaborative, shared dimension of
Knowledge: The Semantic Web
“The Semantic Web is an extension of the
current Web in which information is given
well-defined meaning, better enabling
computers and people [and Smart Objects] to
work in cooperation.”
(Tim Berners-Lee, James Hendler and Ora Lassila, The
Semantic Web, Scientific American, May 2001)
Ontology Introduction
25. 25
Two kind of resources
Documental Resources (DR): Human-oriented
information and knowledge
Factual K, such as: the Rome Sheraton Hotel has
250 rooms, the prices are…
Intensional K: An Hotel is composed by: a
reception, some rooms, etc…
Procedural K: To make a reservation, prepare first
the credit card, then enter the hotel Web site, …
Semantic Resources (SR): Knowledge to be
‘understood’ and processed by a computer.
x H, y: hotel(x) has(x,y) reception(y) …
Ontology Introduction
26. 26
Human-readable vs Computer-
readable
According to Tim Berners-Lee:
“Today’s web pages are conceived to be human-
readable (in terms of content), we need to find
solutions to make them computer-readable.”
A technical intuition:
• HTML is the language of the Traditional Web,
to represent human-oriented hypermedia docs
• RDF is the language of the Semantic Web, to
represent computer-oriented knowledge
Ontology Introduction
27. What’s computer ‘readability’?
Ontology Introduction 27
What We Say to Dogs
"Stay out of the garbage!
Understand, Ginger? Stay out
of the garbage!"
What Dogs understand
"... blah blah blah blah GINGER
blah blah blah blah ..."
28. 28
Functions of the Traditional Web
• Keyword-based
Information Retrieval
• Hypertext Navigation
• Manual Classification
• Specialised search robots
(Softbots, crawlers, ..)
!
?
Retrieval quality
(precision & recall)
inversely proportional
to data quantity
Ontology Introduction
29. 29
Functions of the the Semantic Web
• Semantic Information
Retrieval
• Machine Reasoning
• Machine-machine
advanced cooperation
• Shared
Conceptualisations
(shared ontologies)
with std knowledge
representation
Retrieval quality directly
proportional to knowledge
quantity
(and reasoning capabilities)
Ontology Introduction
32. 32
The dimensions of knowledge
• Level of explicitness (Nonaka, theory of Ba): Tacit,
Implicit, Explicit
• Addressee (Human, Machine, both)
• Level of declarative (vs procedurale) approach
• Level of formalization (from NL text to
algebra/logics)
• Level of abstraction (from factual to conceptual to
metaCon)
• Synchronic vs Diachronic (Structural vs
Behavioural)
33. 33
Representing Knowledge? For whom, for what
It depends on the:
• Who is the Addressee
– for people (easy to read and manipulate)
– for machines (easy to process automatically)
– to exchange K between people and machines
• What Activity it supports (for people and/or computers)
– preliminary domain investigation and analysis
– decision support and recommender systems
– Data mining
– detail analysis, design and (sw) implementation
– Business transactions
– Knoweledge storage and retrieval
– Semantic query processing (with reasoning)
– Semantic interoperability
– Intelligent user interfaces
34. 34
Level of declarativeness
According to the OMG-MDA vision:
• Descriptive (Computational Independent
Model)
– Ex. Class Diagram, abstract Business Process
model (EPC, UML, …)
• Prescriptive (Platform IM)
– Workflow Management System (Savvion,
TeamWare, OpenFlow, …), no transaction exec
• Operational (P Specific M)
– Process/action exec specification (e.g., BPEL,
BPMN)
– Enterprise Information System, ERP, SCM, …
35. 35
The three formalisation levels
-Informal: typically textual documents (free or
loosely structured text)
-Semiformal: diagrams, tables, forms (rigorous
structure/syntax, intuitive semantics: UML,
EPC, Purchase order, invoice, etc.)
-Formal: rigorous specification languages
(rigorous syntax and semantics: RDF, OWL,
KIF, Z++, PSL/Pi Calculus, Ontolingua, etc.)
36. 36
The Knowledge Tiers
- Factual knowledge: ground information,
representing individuals (DB technology)
- Conceptual knowledge: representing abstract
entities and operations (Enterprise models and
IS design blueprints)
- Methodological knowledge: representing
languages and guidelines for KB construction
(knowledge engineering languages methods,
metamodels, modeling ideas)
40. Social Ontology Building
and Evolution (SOBE)
SOBE supports the building of shared
ontologies through:
• Automatic knowledge extraction
– Analysis of textual documents by using NLP
techniques
• Social participation
– Voting and discussing (forum) for validating
and enriching extracted knowledge
Ontology Introduction 40
41. SOBE Methodology
• Step-wise approach through five incremental
steps (Milestones)
– Lexicon (M1): plain list of terms
– Glossary (M2): terms + natural language definition
– Concept Categorization (M3): in accordance with
the OPAL (e.g., Object, Process, Actor)
– Taxonomy (M4): definition of ISA hierarchy
– Ontology enrichment (M5): additional
relationships (e.g., predication, relatedness)
Ontology Introduction 41
43. Collaborative Dimension
Driven by Web 2.0 and social communities
philosophy
• Voting: accept/discard results of the
automatic extraction (lexicon and glossary)
• Proposing: new terms and definitions to be
validated by participants
• Discussing: for reaching an agreement on
glossary definitions (dedicated forums)
Ontology Introduction 43
44. 44
Conclusions
• Semantic Web applications will involve humans (H),
smart objects and devices (O), mainly improving:
– O2O communication and cooperation, when devices interact
to support human activities and goals achievments
– H2O and H2H (tech-enhanced), with digital technology that
will progressively disappear, allowing ‘natural’ interactions
• Semantic Web needs Ontologies to interpret meanings of
(digital) resources
• Ontologies effectiveness depends on representation
languages, reasoning, and collaborative consensus
reaching
Ontology Introduction