Knowledge, Ontology, and Business Innovation on the Web

1
Knowledge, Ontology, and Business Innovation on
the Web

• Hans Akkermans
• Amsterdam / Koedijk
• The Netherlands

2
Ontology
3
What is Ontology?

• In philosophy theory of what exists in the world
  (Aristotle, scholastics, Quine)
• In IT consensual formal description of shared
  concepts in a domain
• Aid to human communication and shared
  understanding, by specifying meaning
• Machine-processable (e.g., agents use
  ontologies in communication)
• Ontology key technology in semantic information
  processing
• Applications knowledge management, e-business,
  multi-agent systems, industrial engineering,web
  (service)

4
What Is An Ontology?

• Definition an ontology a formal specification
  of a shared conceptualization of a certain domain
  
• Goal embed this semantic knowledge into systems
  themselves so that they better serve us
• must be (1) human-understandable, and
• must be (2) computer-processable

• Note in philosophy, ontology theory of what
  can exist in the world in general
• W.V.O. Quine What we suppose to exist, in order
  for our theories to be true what exists is what
  can be quantified over to be is to be the
  value of a variable
• ICT what a community assumes to exist in their
  (part of the) world in order to carry out their
  tasks

5
Ontological (Semantic)Approaches to
Interoperability

• Knowledge-rich organizations, people and systems
  must be able to interoperate
• People must communicate, agree on, share
  information and knowledge across boundaries
• Smart organizations internal/external alignment
  of business models, processes, outputs,
  intelligence
• Systems and intelligent devices exchange data,
  devices become Internetworked (ambient
  intelligence), interface with people as knowledge
  workers
• Ontology Web
• (semi)-automated semantic approach to
  interoperability focused on shared meaning of
  exchanged information

6
Web
7
Semantic Web Smart Next Generation of World Wide
Web
The Semantic Web will globalise knowledge
representation, just as the WWW globalised
hypertext (W3C-director Tim Berners-Lee)
8
WWW Stack of Languages

• XML
• Surface syntax, tags for structuring documents,
  no semantics
• XML Schema Describes structure of XML documents
• RDF Resource Description Framework
• Datamodel for relations between things
  object-attribute-value triplets and graphs
• RDF Schema RDF Vocabulary Definition Language
• OWL Web Ontology Language
• W3C standard OWL extends RDF Schema to
  fully-fledged knowledge representation language
• data-typing, cardinality, logical expressions,
  quantifiers

9
Exploiting World-Wide Information Resources

• Web world-wide knowledge sharing
• BUT information resources are heterogeneous,
  distributed, poorly structured, enormous in size
• Need to combat information overload
  (infarction) by new content- and
  meaning-oriented methods
• W3C vision next generation WWW must enable
  built-in semantic information processing
• navigation based on content, not location
• Semantic Web as web of meanings, not just
  hyperlinks
• Semantic Web Ontology (semi)-automated
  approach focused on shared meaning of exchanged
  information
• Semantic approach to interoperability
• New intelligent Internet infrastructure

10
Business Innovation
11
Semantic Search and Browsing

• Knowledge transfer via website?
• Issue hyperlinks or keyword search do not tell
  much, and are a waste of time
• Better ask questions and just get the answer

Ontology helps separate different meanings of
communication
2
Hyperlink structure does not really help!
1
(Semantic sitemaps from OTK/ Aduna Spectacle
tool)
12
Online Interpretation of Multimedia Resources

• Cultural heritage
• News multimedia
• Guus Schreiber et al.

13
From Web Search to Understand and Answer
Questions

• What is this apes colour?
• What is this ape doing?
• Tricky problems for computer
• Concept Trouble to understand meaning and
  interpret content
• Context Trouble with setting or situatedness
• Everyday real world and common sense is a problem
  for computer understanding

14
Q What does this photograph mean?

• General
• People walking around at night
• Specific
• Fall of Berlin Wall in 1989
• Abstract
• End of Cold War / Iron Curtain

15
Q Find other paintings of the same style
MATISSE, Henri Le bonheur de vivre (The Joy of
Life), 1905-1906 Oil on canvas Barnes Foundation,
Merion, PA
DERAIN, Andre The Turning Road, L'Estaque,
1906 Oil on canvas Museum of Fine Arts, Houston,
TX
16
Social Networks Visualization and Analysis

• Peter Mika et al.
• J. Web Semantics, 2004, 2005, 2006
• FOAF ontology
• Friend Of A Friend
• Flink system
• www.flink.semanticweb.org

17
Business Ontology Research From Technology to
Value

• Model/ontology-based theory formation of
  networked value constellations
• Achieve seamless business-IT technology alignment
• Practical applications for business innovation
  with advanced IS/IT design
• Research feedback loop between theory and
  industry practice (action research)

18
Whats in an E-business Model? The e3-value
ontology

• Key ontology concepts
• Actor
• Value Object
• Value Interface
• Value Offering
• Value Transfer
• Market Segment
• Value Activity
• Composed actor

19
Business Model Network of Value Exchanges
between Actors
20
Tulips from Amsterdam
21
Applications of e3 Ontologies in Business and
Industry

• Internet radio and music rights clearance
• Smart Power Networks
• And many others

See videoclips at www.e3value.com
22
FENIX EU-IP in Energy Industry

• Innovative architectures and services in smart
  power networks (FENIX EU-IP)
• Commercial aggregation of many small power
  production and consumption units (DER)
• Virtual Power Plant concept

23
Smart Power Networks

• Real-time imbalance in demand-supply match of
  power grid is very costly
• (and critical for security of supply)
• e3value studies significant business case for
  Distributed Balancing services
• Distributed Control by eMarket technologies
• Field tests Automatic imbalance reduction gt 40
• Commercialization underway

24
Offering eService Bundles for Dementia Care at
Home

• Application based on e3service ontology

25
Online Design of Events (San Sebastián)

• In addition to e3service, domain ontologies for
  events (from tourist industry organization)
• Blackboard-style opportunistic reasoning for web
  service composition

26
Knowledge 1Practice
27
Ontology Application Areas

• Major ontology application areas
• electronic business
• knowledge management
• multi-agent applications (communication)
• virtual communities of interest with high-value
  knowledge areas e.g.,
• industrial engineering (automotive, telecoms, )
• bio- and life sciences informatics
• art/museums, multi-media resources
• linguistic, medical, legal, and other
  professional domains
• Web information system and service integration
  Semantic Web as next generation of WWW

28
Web, Ontology, and KM

• Mika and Akkermans, Knowledge Engineering Review,
  2004

29
Create Common Understanding with all Stakeholders

• Thats what ontologies are about

Source Financial Times, e-procurement, Oct. 2000
30
Ontological Communication Problems (human
computer)
What does this mean?
Fma
And what does a computer think about
that? VIR?

• Its all about communication

(Financial Times, e-procurement, Oct. 2000)
31
Knowledge 2Research and Technology
32
Ontological Engineering

• Different types / generality levels of
  ontologies
• Foundational (upper) ontology (space-time,
  part-of, ...)
• Domain and task ontologies
• Ontology specs express micro-theories of a
  domain
• Lightweight top-level structure in formal
  diagram form (e.g. UML)
• Heavyweight specified as finite axiom sets in
  (some subset of) first-order logic
• Industry practice often starts with metadata
  concept taxonomies (e.g. XML e-commerce
  standards)
• but ontology encompasses much more

Ontology is advance, on top of continuity with
established conceptual modelling and knowledge
representation
33
Issues in Ontology Research

• Ontologies often start as standardized
  vocabularies and concept hierarchies
• But in practice (1) different types of
  hierarchical links (2) multiple class
  hierarchies needed
• Ontologies are useful to express multiple
  viewpoints on the same thing (But how to map?)
• Ontology modularization (horizontal, vertical)
  and Ontology composition (import, specialization,
  mapping)
• Ontology dynamics and evolution, and its
  management
• Ontology representation issues class vs.
  instance flexibility graphical vs. formal specs
  expressivity vs. computational efficiency

34
XML-based product standards in e-commerce
mapping issues
35
Problems with Hierarchies and Taxonomies (and
Folksonomies)

• Whats in a link?
• Hierarchical links often have different semantics
• E.g. Kind-of and Part-of
• Dimensions in making distinctions heterarchy
  rather than hierarchy
• (Multiple) classification along different
  dimensions within single hierarchy creates
  confusion and makes applications unnecessarily
  complex
• In reasoning Hierarchy enforces a single fixed
  sequence of dimensions
• fixed ordering not always possible or desirable

36
Ontology and Views on the World
e3value ontology

• PhysSys ontology library -
• Modularization principles
• Viewpoints (horizontal)
• Abstraction levels (vertical)

37
Web and KM Future Trends
38
Knowledge 3 Epistemology and Philosophy
39
Knowledge Representation and Reasoning

• Stone-age AI knowledge representation works (!)
• Frames (OO classes taxonomy, OAV triplets, RDF)
• Taxonomies and Rules (a.o. simple forms of
  reasoning) are natural for humans
• Context-dependence, but rather stable knowledge
  patterns do exist
• Formal logic vs. practical argument/reasoning
• Differences less and more

(Toulmin, 1958)
40
Ontology and Conceptualization
41
Ontology as Scientific Method Conceptualization
Theory Formation

• Theory Coherent set (system) of concepts plus
  assertions on how they relate
• Conceptualization is crucial foundation upon
  which theory is built (cf. first principles in
  science)
• Scientific method is relative (and subordinate)
  to
• Teleology research goals (what you want and
  hence, context)
• Technology (broadly what you can/not achieve,
  so also theory)
• Theories have implications, or rather,
  implication networks
• Evaluation of theory is not hypothesis testing
  but much more network goodness of fit
• Ontology is (new!) scientific method for theory
  formation
• Formal conceptualization
• Test by computational implementation

42
Context

• Conceptualization and reasoning methods depend on
  context
• Real-world reference, lifeworld
• Task and Domain specifics

• But Context transcendence is also empirically
  present
• Stable knowledge patterns of representation and
  reasoning exist
• Many ontologies appear to be useful to
  communities
• FOAF, WordNet, Upper-O, etc.

43
Can Computers Understand?

• Android epistemology
• Knowing subjects, Nature/ways of knowing only
  human(s)?
• Is cognition social? (cf. ontology as paradigm)
• Knowledge engineering provides a theory of
  knowledge
• Dealing with context and context transcendence
• Actionable knowledge
• Rational reasoning, good/acceptable argument ?
  formal logical inference/deduction
• Ontology as static representation ? Dynamics?
  Reflection and Learning (Argyris/Schon) ?
  Emergence/self-organization
• Communicative action
• Syntax ? Semantics ? Pragmatics
• Creation of society, person, culture via
  interaction (Habermas)

44
Knowledge 4Science and Society
45
Sciences 19th Century Societal Context

• Industrial Revolution
• Watts patent 1784 steam engine as a generic
  enabling technology for industry
• Many applications took decades to take shape
• Sciences (as we know them now) co-evolved with
  (rather than caused) Industrial Revolution
• For example general concept and laws of energy
  emerged only after 1840
• Academic ideals are from this period, but no good
  reason to assume that they still fit 21st
  century societal context
• being connected to societal needs but still
  catering for abstraction, reflection, Bildung
• help shaping (designing) new structures rather
  than just studying existing ones

46
Sciences 21st Century Societal Context

• New phenomenon information and knowledge as
  Ding an Sich
• Human factor is central but enabled, pushed and
  pulled, by ICT technology and infrastructure
• new social fabric, in part driven by technology
• Industrial revolution mechanized manual labour
• Likewise, knowledge economy and information
  society revolutionize intellectual labour
  (info/knowledge work)
• Computer steam engine of knowledge economy?
• (but if so, we still live in the early 19th
  century!)
• This is the rationale behind Information Science
  as an emerging new area and discipline

47
Science and Society (Take 21)

• University always had, has, must have a social
  function
• Ongoing contextualization of science (Nowotny
  et al., 2001) society also speaks back to
  science
• Todays technological and socio-economic
  developments - summarized as Information Society
  and knowledge economy - constitute a fundamental
  and lasting societal innovation
• This justifies, and makes useful, new academic
  research that transgresses ancient disciplinary
  boundaries
• On the way, it is necessary to rethink and
  reshape the universitys societal function in its
  21st century context
• This will ultimately also impact what is science
  as such as perceived by academia itself

48
Future(s) of the Information Age

• Images and metaphors influence developments
• Kuhn paradigms and their change
• Marx man creates society, but in given external
  conditions
• Antiquity Homo homini lupus
• Neo-conservatism society and individual as
  free market competition
• Industrial society mechanical clock as (not just
  a) metaphor for social organization
• Cf. Taylorism
• Information society computational metaphors move
  into social reality, and vice versa
• Cf. computer/network information processing
  paradigms agents, conceptual models
  (ontologies), markets network IS the knowledge
• But which metaphors and images will become
  reality? And how can we use our societal design
  freedom to our own human benefit?

49
Thank You For Your Attention!
50
Further information on www.e3value.com