Pragmatic Web 4.0
Kommission „Die Natur der Information“
Göttinger Akademie, 8.11.2013
Prof. Dr. Adrian Paschke
Corporate Semantic Web (AG-CSW)
Institut für Informatik, Freie Universität Berlin
paschke@inf.fu-berlin
http://www.inf.fu-berlin/groups/ag-csw/
and
Department of Information Systems
Poznan University of Economics
paschke@inf.fu-berlin
Agenda
What is Semantics?
Declarative Knowledge
Representation in IT
The Semantic Web – An
Introduction
Semantic Web and it’s Relations
What comes next?
Data vs. Information
Data
A “given,” or fact; text, a number, a statement, or a
picture, …
The raw materials in the production of information
Information
Data that has been put into a meaningful and useful
context.
Example Data vs. Information
data
95
information My score on the final exam is
95%
knowledge I have passed the exam with
excellent mark bdb
data
representation,
e.g. relational
DB
data + context
+ information
representation
data /
information +
meaning
interpretation
Search Results from Publication
Database
Title
Lorenz P,
Transcriptional repression
mediated by the KRAB domain of the human
Author
C2H2 zinc finger protein Kox1/ZNF10 does not
require histone deacetylation.
Biol Chem. 2001 Apr;382(4):637-44.
Fredericks WJ. An engineered PAX3-KRAB
transcriptional repressor inhibits the malignant
Journal
Year
phenotype of alveolar rhabdomyosarcoma
cells harboring the endogenous PAX3-FKHR
oncogene.
However, for a machine things look different!
Mol Cell Biol. 2000 Jul;20(14):5019-31.
Results from Publication Database
Lorenz P, Transcriptional repression
mediated by the KRAB domain of the
human C2H2 zinc finger protein
Kox1/ZNF10 does not require histone
deacetylation.
Biol Chem. 2001 Apr;382(4):637-44.
Fredericks WJ. An engineered PAX3KRAB transcriptional repressor inhibits
the malignant phenotype of alveolar
rhabdomyosarcoma cells harboring the
endogenous PAX3-FKHR oncogene.
Mol Cell Biol. 2000
Solution:
Tags (XML)?
Jul;20(14):5019-31.
Results from Publication Database
<author>Lorenz P</author><title>Transcriptional repression
mediated by the KRAB domain of the human C2H2 zinc finger
protein Kox1/ZNF10 does not require histone deacetylation.
</title>
<journal>Biol Chem </journal><year>2001<year>
<author>Lorenz P</author><title>Transcriptional repression
mediated by the KRAB domain of the human C2H2 zinc finger
protein Kox1/ZNF10 does not require histone deacetylation.
</title>
<journal>Biol Chem </journal><year>2001<year>
However, for a machine things look different!
...
Results from Publication Database
<author>Lorenz
P</author><title>Transcriptional
repression mediated by the KRAB
domain of the human C2H2 zinc finger
protein Kox1/ZNF10 does not require
histone deacetylation. </title>
<journal>Biol Chem
</journal><year>2001<year>
<author>Lorenz
P</author><title>Transcriptional
repression mediated by the KRAB
domain of the human C2H2 zinc finger
protein Kox1/ZNF10 does not require
histone deacetylation. </title>
<journal>Biol Chem
</journal><year>2001<year>
Solution: Use Semantic
Knowledge
Representation
Example: Traffic Light
Syntax – Semantics - Pragmatics
Syntax
green (bottom); yellow; red
Semantics
green = go; …; red = stop
Pragmatics
If red and no traffic
then allowed to go
Example: Question-Answer Interaction
Syntax – Semantics - Pragmatics
Syntax
“What time is it?” (English)
Semantics
Question about current time (Meaning)
Pragmatics
An answer to the question is obligatory
(even if time is unknown) (Understanding
and Commitment)
Example - XML Syntax vs. Semantics
Adrian Paschke is a lecturer of Logic Programming
<course name=“Logic Programming">
<lecturer>Adrian Paschke</lecturer>
</course>
<lecturer name=“Adrian Paschke">
<teaches>Logic Programming</teaches>
</lecturer>
Opposite nesting (syntax), same meaning (semantics)!
Syntax – Semantics - Pragmatics
Syntax
about form
Semantics
about meaning
Pragmatics
about use.
Semantic Technologies for
Declarative Knowledge Representation
1. Rules
Describe derived conclusions if premium(Customer)
and reactions from given
then discount(10%)
information (inference)
2. Ontologies
equal
with
Customer
Ontologies represent the conceptual
knowledge of a domain (concept
semantics)
is a
Partner
Client
What is an Ontology? (in IT)
An Ontology is a
formal specification
Executable, Discussable
of a shared
Group of persons
conceptualization
About concepts; abstract class
of a domain of interest
e.g. an application, a specific
area, the “world model”
[Gruber 1993] - T.R. Gruber, Toward Principles for the Design of Ontologies Used for
Knowledge Sharing, Formal Analysis in Conceptual Analysis and Knowledge
Representation, Kluwer, 1993.
What is a Rule? (in IT)
1. Rules
•
•
Derivation rules (deduction rules): establish / derive new information
Reaction rules that establish when certain actions or effects should
take place :
• Condition-Action rules (production rules)
• Event-Condition-Action (ECA) rules + variants (e.g. ECAP).
• Messaging Reaction Rules (event message reaction rules)
2. Constraints
•
•
•
Structural constraints (e.g. deontic assignments).
Integrity constraints and state constraints
Process and flow constraints
[Paschke, A., Boley, H.]: Rule Markup Languages and Semantic Web Rule Languages, in Handbook of Research on Emerging Rule-Based
Languages and Technologies: Open Solutions and Approaches, IGI Publishing, ISBN:1-60566-402-2, 2009
Example: Ontology and Rules
Ontology
Object
is_a-1
Person
is_a-1
is_a-1
knows
has
Topic
described_in
Prior Art
Document
related_to
related_to
is_a-1
Patent
Application Priority
Skill
Patentee
Technique described_in
Teaching
writes
is_a-1
Patent
date
becomes
granted
RULES:
Topic
Patentee
writes
described_in
Patent
Application
Document
is_about
Topic
Topic
Patentee
Patentee
is_about
knows
has
Document
Topic
Skill
Ontologies and their relatives
informal
formal semantics
expressiveness
Based on AAAI’99 Ontologies Panel – McGuiness, Welty,
Ushold, Gruninger, Lehmann
Many Ontology Languages
No special ontolgy languages,
Entity Relationship Modell
but might be used to describe
ontologies
UML with OCL
Frames, F-Logic
Predicate Logic
Common Logic
formal logic
Description Logic
specialized
SHOE, XOL, OML, SKOS, OBO, SBVR, …
Web Ontology Languages
RDFS, DAML+OIL -> OWL
ODM
Ontology Transformation
…
Logic and Knowledge Representation
in IT
Declarative Knowledge Representation
express what is valid, the responsibility to interpret
this and to decide on how to do it is delegated to an
automated interpreter / reasoner
Formal logic-based languages for the
representation of knowledge with a clear
semantics
Main Requirements of a Logic-based
Ontology / Rule Language in IT
a well-defined syntax
a formal semantics
sufficient expressive power
efficient reasoning support
convenience/adequacy of
expression syntax
Logic-based Knowledge Representation
First Order Logic
Expressive syntax: constants, functions, predicates, equality,
quantifiers, variables
Objects and relations are semantic primitives represented as
predicate formula
But: reasoning not efficient and undecidable
Solution: Restriction to Subsets of FOL
Horn Logic (Logic Programming / Rules)
Descripition Logics (Ontologies)
But: convenience of expression: only formal syntax +
semantics, but not a Web representation format
=> Semantic Web Syntax and Semantic Web Data Model needed
Semantic Web – An Introduction
"The Semantic Web is an
extension of the current web in
which information is given welldefined meaning, better
enabling computers and people
to work in cooperation."
Tim Berners-Lee, James Hendler,
Ora Lassila, The Semantic Web
„Make the Web understandable
for machines“
W3C Stack 2007
Main Building Blocks of the
Semantic Web
1.
2.
3.
4.
Explicit Metadata on the WWW
Ontologies
Rule Logic and Inference
Semantic Tools ,Semantic Web Services,
Software Agents
The (current) W3C Semantic Web Stack
Ontologies
RDF Query
Language
Rules
Semantic Web
Information
Model
Standard
Internet
Technologies
W3C Semantic Web Stack since 2007
Overview on the Semantic Web
Technologies
URI/IRI: Web Resource Identifiers
Note: Representational State Transfer (REST) – Resources are
abstraction from an information/knowledge (e.g. „weather in Göttingen“)
XML: Syntactic basis
RDF: Resource Description Framework
RDF as Web data model for facts and metadata
RDF schema (RDFS) as simple ontology language
(mainly taxonomies)
SPARQL as a RDF query language
Linked Data – data publishing method
Overview on the Semantic Web
Technologies (2)
Ontology
RDF Schema (RDFS) and Web Ontology Language (OWL)
Rules / Logic
Rule Interchange Format (RIF, RuleML)
Proof
Generation of proofs-, interchange of proofs, validation
Trust
Digital signatures
recommendations, ratings
Semantic Web Applications & Interfaces
e.g. Semantic Search, Semantic Agents, …
Metadata Problem Domains
Syntax:
Which representation and interchange format for
metadata? (Microformats, RDF, RDFa, Microdata)
Semantics:
Which metadata are allowed for Web resources
(expressiveness, metadata vocabulary, schema)
Association problem:
How to connect metadata with resources? (who
defines the metadata, are metadata separated
from the content (RDF vs. RDFa), etc.)
RDF Triple Stores
A specialized database for RDF triples
Supports a query language
SPARQL is the W3C recommendation
Triple stores might be in memory or provide a
persistent backend
Presistence provided by an underlying relational DBMS
(e.g., mySQL) or a custom DB for efficiency.
Example: SPARQL SELECT
SELECT:
SELECT Variables
FROM Dataset
WHERE Pattern
Examples:
PREFIX foaf: <http://xmlns.com/foaf/0.1/>
SELECT ?name
WHERE ( ?x foaf:name ?name )
PREFIX foaf: <http://xmlns.com/foaf/0.1/>
SELECT *
WHERE ( ?x foaf:name ?name )
The (current) W3C Semantic Web
Architecture
W3C Semantic Web Stack since 2007
OWL vs. RDFS
More Expressiveness
Definition of relations between classes
Definition of constraints and cardinalities
Constraints on properties: exists, forall, cardinality
Definition of equivalences between classes (e.g.
different ontologies)
Properties of properties
Boolean combinations of classes and constraints
…
Example: OWL Ontology
peopleAtUni
range
id
Student
involves
domain
domain
Course
OWL
Staff Member
range
domain
subClassOf
equivalentClass
subPropertyOf
isTaughtBy
phone
domain
unionOf
T-Box
Model
range
Literal
1
Faculty
Academic Staff Member
subClassOf
range
subClassOf
subClassOf
disjointWith
Professor
Assistant
Professor
Associate
Professor
rdf:type
rdf:type
RDF
isTaughtBy
Semantic Web
Adrian Paschke
A-Box
Model
Reasoning with OWL
Semantics of OWL is defined by Description Logics (DL)
Satisfiability: whether the assertions in an TBox and ABox has
a model (i.e. non-contradicting)
Subsumption: whether one description is more general than
another one
Equivalence: whether two classes denote same set
Consistence: whether its set of assertions is consistent
Instantiation: check if an individual is an instance of class C
Retrieval: retrieve a set of individuals that instantiate C
The (current) W3C Semantic Web
Architecture
W3C Semantic Web Stack since 2007
Example: Rule Markup Language
Standards (RuleML)
RuleML 1.0 (Deliberation, Reaction, Defeasible, Modal, …)
Semantic Web Rule Language (SWRL)
Uses RuleML Version 0.89
Semantic Web Services Language (SWSL)
Uses RuleML Version 0.89
W3C Rule Interchange Format (RIF)
Uses RuleML Version 0.91 with frames and slots
OASIS LegalRuleML
Uses RuleML Version 1.0
OMG Production Rules Representation (PRR)
Input from RuleML
OMG Application Programming Interfaces four KBs (API4KB)
Input from Reaction RuleML 1.0
Unifying Logic
• Not standardized in W3C Semantic Web Stack yet
• Which semantics? (e.g., Description Logics, F-Logic, Horn Logic, …)
• Which assumptions? (e.g., Closed World, Open World, Unique Name)
• …
W3C Semantic Web Stack since 2007
Example Decision Scenario
You need to wait if the
traffic light is not green.
I know that the traffic light
is green, so I’m allowed to
cross the street
I’m not sure if the traffic
light is green, so I’m
allowed to cross the street
????
Open World vs. Closed Word
Assumption
Open World Assumption (typical for ontologies)
explicitly prove the truth of negation
Closed World Assumption (typical for rules / logic programs)
if we do not know (from our closed knowledge base) we assume
falsity
This difference has practical implications
Traffic light example:
Under open world assumption we need to explicitly
prove that the light is not red => cross street
Under closed world assumption it is enough if we
prove that there is no information if the light is red
=> cross street
Who is responsible if an accident happens?
Unique-Names Assumption
isTaughtBy
domain
Course
range
1
Academic Staff Member
A course is taught by at most one staff member.
The course „Ontologies in IT“ is taught by
„Prof. Paschke“ and „Prof. Wecel“
OWL does not adopt the unique-names assumption
of database systems/logic programs (rules)
If two instances have a different name or keys/IDs does not
imply that they are different individuals
An OWL reasoner does not flag an error
Instead it infers that the two resources are equal
A rule reasoner / deductive database does flag an error
Proof and Trust
• Proof Markup Languages, Justifications and Argumentations
• Claims can be verified, if there are evidences from other (trusted) Internet
sources
• Semantic Reputation Models
Example Scenario – eCommerce
E-Shop
Review
Relying Party
Reseller Bob
Delivery
Service
Buyer
Monitored
Delivery
Performance
Business
Owner/Seller/Factory
used for service
management
used for buying
decisions
Semantic
Reputation Object
Semantic Web Reputation and Trust Management
http://www.corporate-semantic-web.de/rule-responder.html
Other Buyers
Use Cases / Applications / Tools
Semantic-enriched Search
Content management
Knowledge management
Business intelligence
Collaborative user interfaces
Sensor-based services
Linking virtual communities
Grid infrastructure
Multimedia data management
Semantic Web Services
Etc. see e.g.SWEO’s use case collection
http://www.w3.org/2001/sw/sweo/public/UseCases/
Example:
Semantic Desktop Systems
Combine desktop systems with Semantic Web
Technologies
Extract, manage, visualize and use semantic and
contextual associations respectively metadata for
Personal Information Management (PIM)
e.g. Gnowsis, Nepomuk, Beagle++, Social Semantic Desktop, Haystack
Example: Job Portal
Semantic Recommendation
d (Java, C++) = d (Java, Object Oriented) + d (C++, Object Oriented)
= (0.25-0.0.0625) + (0.25-0.0625)
= 0.375
sim(Java, C++) = 1 – 0.375 = 0.625 (Semantic Similarity is 0,625)
Example:
Query „Job offers for Java Programmer“ + expanded with Personal Skill Profile (Java +
C++ Knowledge)
=> also recommend job offers for C++ programmer
(see Semantic Matchmaking Framework: http://www.corporate-semantic-web.de/technologies.html
Example: OMG Ontology Definition Metamodel (ODM)
Ontology Definition Metamodel
MOF
MOF XMI
Of UML
MOF XMI
Of ODM
UML
ODM
User
UML Model
UML XMI
Of User Model
User
Ontology
Ontology XMI
Of User Model
ISO
Topic Maps
M2
M1
User
Instances
UML 2
(+OCL)
M3
M0
ISO
CL
W3C
RDFS
W3C
OWL
ODM brings together the communities (SE+KR) by providing:
Broad interoperation within Model Driven Architecture
MDA tool access to ontology based reasoning capability
UML notation for ontologies and ontological interpretation of UML
Example: Rule Markup Language
Standards (RuleML)
RuleML 1.0 (Deliberation, Reaction, Defeasible, Modal, …)
Semantic Web Rule Language (SWRL)
Uses RuleML Version 0.89
Semantic Web Services Language (SWSL)
Uses RuleML Version 0.89
W3C Rule Interchange Format (RIF)
Uses RuleML Version 0.91 with frames and slots
OASIS LegalRuleML
Uses RuleML Version 1.0
OMG Production Rules Representation (PRR)
Input from RuleML
OMG Application Programming Interfaces four KBs (API4KB)
Input from Reaction RuleML 1.0
Social Semantic Web
The concept of the Social Semantic Web
subsumes developments in which social
interactions on the Web lead to the creation
of explicit and semantically rich knowledge
representations. (Wikipedia)
Corporate Semantic Web
Corporate Semantic Web (CSW) address
the applications of Semantic Web
technologies and Knowledge Management
methodologies in corporate environments
(semantic enterprises).
(www.corporate-semantic-web.de)
Corporate Semantic Web
Public Semantic Web
Corporate Semantic Web
Business Context
Corporate
Semantic
Engineering
Corporate
Semantic
Search
Corporate
Semantic
Collaboration
Corporate Business Information Systems
Pragmatic Web
The Pragmatic Web consists of the tools,
practices and theories describing why and how
people use information. In contrast to the
Syntactic Web and Semantic Web the Pragmatic
Web is not only about form or meaning of
information, but about interaction which brings
about e.g. understanding or commitments.
(www.pragmaticweb.info)
Pragmatic Agent Web
The Pragmatic Agent Web utilize the Semantic Web with
multiple interacting intelligent agents which collaborate on
the Web and put independent meta data, ontologies and
local data into a pragmatic context such as communicative
situations, organizational norms, purposes or individual
goals and values.
Duration & Connectedness
Intelligence
Knowledge
Pragmatic
Semantic
Information
Syntax
Data
(Machine) Understanding
Pragmatic Agent Web (2)
Utilize the heterogenous Semantic Web resources, meta data and
meaning representations with intelligent agents and web-based services
with the ability to understand the others intended meaning (pragmatic
competence)
Formal Logic Representation vs. (Controlled) Natural Language Representation
Collaborate in a communicative conversation-based process where
content and context is interchanged in terms of messages (relation of
signs) between senders and receivers (interpreters/users).
Loosley-coupled vs. de-coupled interactions
Fixed negotiation and coordination protocols vs. free conversations
Pragmatic layer/wrapper around semantic/content e.g. by KQML / ACL
like speech-act primitives (e.g. assert(content), retract(content), query(kb))
Model, negotiate and control shared and invividual meanings
requires learning and knowledge adaption / updates
Challenges for the Semantic Web
Connectedness
Intelligence / Wisdom
understanding
principles
Pragmatics
Knowledge
Understanding
patterns
Sematics
Information / Content
Understanding relations
Data
Ontologies
(Logic)
Rules
(Logic)
Syntax
???
(Human Logic +
Machine Logic)
Understanding
Ubiquitous Pragmatic Web 4.0
Pragmatic Agent
Ecosystems
Machine
Understanding
Situation Aware Real-time Semantic
Complex Event Processing
Ubiquitous Pragmatic Web 4.0
Pragmatic Web
Connects Intelligent Agents and Smart Things
Massive
Multi-player Web Gaming
Ubiquitous autonomic
Smart Services and
Things
Smart Web TV
Social Semantic Web 3.0,
Web of Services & Things,
Corporate Semantic Web Connects
Semantic Web
Smart Content
People, Services and Things
Semantic Web 2.0
Connects Knowledge
Syntactic Web
World Wide Web 1.0
Smart Content
Passive
Active
Desktop Computing
Syntactic
Web
Semantic
Web
Consumer
Smart
Agents
XML
RDF
Monolithic
Systems Era
HTML
Desktop
Content
Producer
Connects Information
Pragmatic
Web
Ubiquitous Next Generation Agents and Social Connections