ontological analysis for conceptual modelinghome.deib.polimi.it/schiaffo/tfi/lecture...
TRANSCRIPT
Ontological Analysis for Conceptual Modeling:
Foundational Ontologies
Roberta Ferrario Laboratorio di Ontologia Applicata (LOA)
Istituto di Scienze e Tecnologie della Cognizione (ISTC-CNR) Trento, Italy
www.loa-cnr.it
2
Summary
• The semantic dimension : from form to content • The role of ontological analysis: what are ontologies and what
are they for • Formal ontology and the bases for ontological analysis • Ontology-driven information systems • An application and an approach: services
The focus of ontological analysis: from form to CONTENT
§ The key problems content-based information access (semantic matching) content-based information integration (semantic integration)
To approach them, content must be studied, understood, analyzed as such,
independently of the way it is represented. Computer technologies are not really good for that... …and users of computer systems are often confused by technology
Ontologies: a magic solution?!
Do we know what to REpresent?
First ontological analysis, THEN knowledge representation…
Unfortunately, this is not the current practice…
6
So… Nothing new?!
• Investing on analysis always pays off (most of all in a re-use perspective)
• To use instruments of ontological analysis and of semantic technologies offers further advantages:
• The analysis is more robust, well-founded, re-usable • Semantic technologies offer advantages (as automatic reasoning)
both at design time and at run time • Analysis (and ontology) can “automatically” guide nearly everything
in the system
A common alphabet is not enough…
“XML is only the first step to ensuring that computers can communicate freely. XML is an alphabet for computers and as everyone who travels in Europe knows, knowing the alphabet doesn’t mean you can speak Italian or French”
Business Week, March 18, 2002
Standard glossaries can help, but...
• Defining standard vocabularies is difficult and time-consuming • Once defined, standards don’t adapt well • Heterogeneous domains need a broad-coverage vocabulary • People don’t implement standards correctly anyway • Vocabulary definitions are often ambiguous or circular
Ontology, lexicon, semantics
Distinctions among contents: Ontology (capital “o”) Reference to content: Lexicon, via Semantics Every organization, every computer system
Makes (implicit) ontologic assumptions Adopt a certain lexicon, to which an intended semantics is
ascribed.
Ontology and ontologies
Ontology: the philosophical discipline Study of what there (possibly) is Study of the nature and structure of reality
Domain of entities Categories and relations Characterizing properties
An ontology: a theoretical or computational artifact “An explicit and formal specification of a conceptualization” (Gruber) A specific artefact expressing the intended meaning of a vocabulary in terms
of the nature and structure of the entities it refers to
11
The importance of philosophy for ontologies
• In order to build a model that is implementable in an information system
one needs to understand what are the objects that populate the domain the system applies to and this is an important exercise of philosophy
• The disciplines that deal directly with those objects usually don’t ask
questions on the foundations of the domain, taking them for granted
• Example: organization theory or sociology don’t try to answer questions like those inherent to the materiality of organizations or to the nature of individuals possessing intentions
12
The importance of ontologies for philosophy
• On the other hand, facing the concrete modeling activity allows
philosophers to see old problems under a new light, thus showing new perspectives that wouldn’t otherwise have emerged
• When one builds a model (especially a formal one) on the basis of a certain theory, such theory in some sense undergoes a validation
• A model abstracts from what is not relevant, thus highlighting what is relevant; this allows to grasp more easily imprecisions and theoretical inconsistencies
What is a conceptualization?
Formal structure of (a piece of) reality as perceived and organized by an agent, independently of:
the vocabulary used the actual occurrence of a specific situation
Different situations involving same objects, described by different vocabularies,
may share the same conceptualization.
apple
mela same conceptualization
LI
LE
What is a conceptualization? A cognitive approach
Humans isolate relevant invariances from physical reality (quality distributions) on the basis of:
Perception (as resulting from evolution) Cognition and cultural experience Language
A set of atomic stimuli (input pattern) is associated to each situation Synchronic level: spatial invariants
Unity properties are ascribed to input patterns: topological and morphological wholes (percepts) emerge
Diachronic level: temporal invariants
Objects: equivalence relationships among input patterns belonging to different situations
Events: unity properties are ascribed to percepts patterns belonging to different situations
Ontology
Ontologies and intended meaning
Language L
Conceptualization C (relevant invariants across
situations: D, ℜ)
Intended models IK(L)
State of affairs State of
affairs Situations
Ontological commitment K (selects D’⊂D and ℜ’⊂ℜ)
Interpretation I
Ontology models IK(L)
Models MD’(L)
Bad Ontology
~Good
Ontology quality: precision and coverage
Low precision, max coverage
Less good
Low precision, limited coverage
WORSE
High precision, max coverage
Good
Max precision, limited coverage
BAD
Levels of ontological precision
Ontological precision
Axiomatic theory"
Glossary"
Thesaurus
Taxonomy
DB/OO scheme Catalog"
Precision vs. accuracy
In general, a single intended model may not discriminate among relevant alternative situations because of Lack of entities
Lack of primitives
Capturing all intended models is not sufficient for a “perfect” ontology
Precision: non-intended models are excluded Accuracy: non-intended situations are excluded
When is a precise (and well-founded) ontology useful?
1. When subtle distinctions are important
2. When recognizing disagreement is important
3. When general abstractions are important
4. When careful explanation and justification of ontological commitment is
important
5. When mutual understanding is more important than interoperability
Ontologies vs. classifications
Classifications focus on: access, based on pre-determined criteria
(encoded by syntactic keys)
Ontologies focus on:
Meaning of terms Nature and structure of a domain
23
Applied ontology • Differently from other computer science approaches, applied
ontology has as a goal to produce conceptual models that are independent form applications and from ways of presenting data
• Ontologies as axiomatic theories that have the purpose of describing explicitly the fundamental entities of a certain field of enquiry underlying their relevant properties and the relations among them
• The methodology consists in preliminarily using the analytical tools taken from different disciplines (cognitive sciences, linguistics…) among which philosophy plays a dominant role
24
Different kinds of ontologies
• Foundational ontologies and domain ontologies
• Foundational (top level) ontologies can be reused and specialized in domain ontologies or in task ontologies
• Domain and task ontologies have the purpose of characterizing a precise domain of analysis, singling out the theoretical primitives that can better deal with the problems they are built to solve
• Ideally, domain and task ontologies should rely on a foundational ontology that shows and drives modeling choices
• Only at the end of such specialization process one can build application ontologies that are really functional
25
The modular approach
• As it is well known, within each discipline very heterogeneous
positions – often irreducible one to the other - may co-exist; for such reason assuming a modular approach is particularly helpful
• Instead of relying on a unique monolithic ontological system, a
library of ontologies linked by formal relations can be built, in which each module, i.e. each ontology, expresses a particular position
A single, imperialistic ontology?
An ontology is first of all for understanding each other ...among people, first of all! not necessarily for thinking in the same way
A single ontology for multiple applications is not necessary Different applications using different ontologies can co-exist and co-
operate (not necessarily inter-operate) ...if linked (and compared) together by means of general enough basic
categories and relations (primitives).
If basic assumptions are not made explicit, any imposed, common ontology risks to be seriously misused or misunderstood opaque with respect to other ontologies
Which primitives? The role of ontological analysis
Theory of Essence and Identity Theory of Parts (Mereology) Theory of Wholes Theory of Dependence Theory of Composition and Constitution Theory of Properties and Qualities
The basis for a common ontology vocabulary
Idea of Chris Welty, IBM Watson Research Centre, while visiting our lab in 2000
The problem of primitives
Representation primitives vs. ontological primitives (against arbitrary interpretations)
Let's aim at general primitives, similarly to what happens in
mathematics: set, relation, transitive, symmetric…
29 07/05/2007
The ontological level
Level Primitives Interpretation Main feature
Logical Predicates, functions
Arbitrary Formalization
Epistemological Structuring relations
Arbitrary Structure
Ontological Ontological relations
Constrained (meaning postulates)
Meaning
Conceptual Conceptual relations
Subjective Conceptualization
Linguistic Linguistic terms
Subjective Language dependence
30
Towards a methodology
• Linguistic analysis: • To simplify composite terms
• hasMunicipalityJob, hasMunicipalityDateofBirth... • hasStateVerificationDueContribution...
• Analysis of dependences: • relational/non relational terms
• To associate data with the entities that have generated them (and the attributes to the entities they depend from)
• To make explicit temporal entities: events, stories, situations... • …and their participants • Ex.: work relation (with its participants)
• To analyze the effects of possible changes • Address/geographical place...
31
Towards a methodology (II) • First steps of a formalization
• Choice of the domain of discourse • Choice of the relevant concepts and relations for the given domain • Choice of the primitive concepts and relations (son vs. relative o ancestor)
1. Alignment with a reference ontology (DOLCE?)
1. Objects 2. Events 3. Qualities 4. Facts 5. Descriptions 6. ...
• Systematic ontological analysis in terms of formal properties and relations...
32
Ontology-driven information systems
• Each IS has its own (implicit or explicit) ontology • The ODIS perspective: explicit ontologies play a central role, driving
all aspects and components of an IS • Two (main) dimensions to claim the role of an explicit ontology :
• Temporal dimension: development time vs. run time • Structural dimension : impact on the various components of an IS:
• Database component • Application program • User interface
An application: services!
33
Services are everywhere..."Services do not operate in isolation, they interact with other
services, thus they are part of a service system!Service systems are themselves part of wider contexts, i.e. socio-
technical systems, that are composed by services, various technological artefacts, various other resources and social agents"
Different disciplines are interested in the study of services, but they lack a common understading of what services are"
"Need of a Services Science!"Chesbrough & Spohrer, A Research Manifesto for Services Science. Communications of the ACM, 2006"
""
Some terminological issues!
What is a service?"An action"A generic type of action"The capability to perform some action"A computational procedure capable to perform some action"An agent in charge of performing an action"An agent ready to perform an action"The result of an action"The (subjective) value of an action"..."
What is a service provider?"The authority which guarantees the presence of a service"The actual agent who executes the service actions (possibly on behalf
of somebody else) "…""
35
From the black box to the glass box view!Current web services models (black box view): transfer function from an
input state to an output state Yet, business applications need to specify
how the service is performed at the business level, referring to internal details whose nature varies a lot from service to service
when the various processes involved in a service occur Business applications need to monitor and evaluate services quality with
respect to their actual impact on the whole service system, which includes external events, objects, people, organizations... (context-aware services)
Service Level Agreements need to refer both to internal and contextual details
Well-known gap between business services and IT Need to look inside the box and out of the box...
Looking outside the box !
"Taking people and society into account "
Putting people in the loop: from users to actors !Putting social interactions in the loop "Putting social institutions in the loop """
Taking external environment into account "External natural environment "Internal aspects of artifacts !Interfaces (sensors) "
36
Looking inside the box!
"Cognitive transparency, to support: "
"At design time "
"participatory, concurrent, agile design "
""At run time "
"governance "control "
37
Services as events: the basic idea!
""A service is an agent’s availability to guarantee some action useful for
somebody, upon prior agreement, in correspondence of certain events """
• How can you tell that a service is present, here and now?"• What do you pay for, when you invest in a service?"
A service is essentially a promise [O’Sullivan 2006]
The main actors!
Service Provider is the person or authority who commits to have the service executed"
Service Producer is the person or organization that actually performs the actions constituting the delivery of the service "Provider and producer may coincide, but this is not always the case. "
"Service Customer is the one that requests the service and then
negotiates for its customized delivery" Service Consumer is the direct beneficiary of the service "Customer
and consumer may coincide, but this is not necessarily so"Consumer and producer may also coincide, in very particular situations""
41
Examples of elementary axioms in DOLCE
Parthood: “x is part of y”P(x, y) ! (AB(x) " PD(x)) # (AB(y) " PD(y))
Temporary Parthood: “x is part of y during t”P(x, y, t) ! (ED(x) # ED(y) # T(t))
Constitution: “x constitutes y during t”K(x, y, t) ! ((ED(x) " PD(x)) # (ED(y) " PD(y)) # T(t))
Participation: “x participates in y during t”PC(x, y, t) ! (ED(x) # PD(y) # T(t))
Quality: “x is a quality of y”qt(x, y) ! (Q(x) # (Q(y) " ED(y) " PD(y)))
Quale: “x is the quale of y (during t)”ql(x, y) ! (TR(x) # TQ(y))ql(x, y, t) ! ((PR(x) " AR(x)) # (PQ(y) " AQ(y)) # T(t))