Knowledge Engineering Meets Software Engineering

1
Knowledge Engineering Meets Software Engineering

• Hele-Mai Haav
• Institute of Cybernetics at TUT
• Software department

2
Outline
Knowledge-based systems Ontologies and their
applications (Semantic) web services
SE
KE
3
Knowledge-based Systems...

• ...for engineering computations (also
  knowledge-based SE)
• Historical field of activities at software
  department of IoC
• Systems with structural synthesis of programs
  capturing domain knowledge in form of
  computational models PRIZ, ExpertPriz, NUT, etc
  (E. Tyugu et al)
• Current system CoCoViLa supports visual
  specification of domain knowledge still using
  structural synthesis of programs (E. Tyugu, M.
  Harf, P. Grigorenko, R. Maigre, A. Ojamaa )

4
Ontologies and their Applications

• Ontology a shared conceptualization of a domain
  that is commonly agreed to by all parties, a
  specification of a conceptualization (Gruber
  1993)
• Ontology means to facilitate knowledge reuse by
  different applications, software systems and
  human resources.
• Ontologies are highly expressive knowledge models
  ? increase expressiveness and intelligence of a
  system

5
Ontologies and their ApplicationsOntology
Learning

• Ontology learning using Formal Concept Analysis
  (FCA) (H-M. Haav)
• Combining FCA and Horn logic for ontology
  extraction and representation (H-M. Haav)
• Learning user browsing behavior ontology for
  recommender systems.
• The knowledge acquired from users browsing
  behavior is used for learning profile ontology
  and formulating explicit user profiles in OWL-DL
  for recommender systems. The method exploits the
  automated reasoning capabilities provided by
  OWL-DL in order to automatically classify user
  profiles. (H-M. Haav, A. Kalja, T. Robal)
• Ontology learning from relational databases (I.
  Astrova, A. Kalja)

6
General schema of ontology learning using FCA
Domain specific texts or data
NLP based context extraction
Set of rules describing initial ontology
FCA and reduction
Concept lattice based ontology expression
Transformations
Formal Context
More rules and facts
automatic
Complete set of rules and facts representing
ontology
Inference
Expert manually
7
The process of user profile learning method
8
Ontologies and their Applications Ontology
Applications

• Smart ontology-based spatial data retrieval
• Partners IoC (H-M. Haav), companies Regio, Girf
• new project 2008-2009, partially funded by
  Enterprise Estonia via ELIKO Competence Center in
  Electronics-, Info- and Communication
  Technologies
• Semantic interoperability of large scale IS The
  Estonian public sectors case study (H-M. Haav,
  A. Kalja, P. Küngas, M.Luts)
• Automatic transformation of OWL ontologies to
  relational databases (SQL) and storing them in
  relational databases (I. Astrova, A. Kalja)

9
Semantic interoperability architecture for state
information system in Estonia
10
Modularity and layering of ontologies component
in interoperability architecture
11
(Sem)web services

• Composition of web services using structural
  synthesis of programs and visual specifications
  (the CoCoViLa system) (E. Tyugu, P. Grigorenko,
  R. Maigre)
• Web service composition using FOL theorem prover
  RqlGandalf.
• RQL (Rule-based Systems for Creation of Web
  Services) project partially funded by Enterprise
  Estonia, 2004-2005.
• Partners Institute of Computer Science of TUT,
  Cell Networks, Sampo Assets Management).
• T. Tammet, H-M. Haav, M. Kääramees, V. Kadarpik,
  K. Kindel
• Annotation of web services using OWL ontologies
  and SAWSDL for support of semantic
  interoperability of state IS (H.-M. Haav, A.
  Kalja, P. Küngas, M. Luts)

12
Web service composition with CoCoViLa

• Automatic service composition tool has been
  developed in software development environment
  CoCoViLa that supports automatic synthesis of
  programs and generates Java code from visual and
  textual model specifications
• User can define desired complex service that is
  synthesize automatically (if it is possible to
  construct the service). BPEL and WSDL
  descriptions of the complex service are then
  generated from Java code.
• Tool has been tested on federated governmental
  information system.

13
Web service composition in RQL
A goal of the system is to automatically find a
plan for service composition as an answer to the
user request. The result of program synthesis is
as a Python program corresponding to the required
composite service.
14
RQL

• Provides a new conceptual and technological
  framework for using a rule language and a rule
  engine for capturing application semantics in
  modern web-based systems.
• The approach enables to deal with two aspects of
  semantics in web-based systems business rules
  and web service composition logic.

15
Future plans

• SE 2.0
• Domain centred problem solving
• Domain semantics is the key to deal with next
  generation technologies

Dillon T. S, Chang E., Wongthongtham P.,
"Ontology-Based Software Engineering- Software
Engineering 2.0," aswec, pp. 13-23,  19th
Australian Conference on Software Engineering
(aswec 2008), 2008
16
Ontology-based SE

• Use of ontologies in different aspects of
  software engineering
• Ontology Based Multi-Site Software Development
• Ontology Mediated Information Access
• Ontology and Semantic Web Services
• Ontology based Multi Agent Systems