SYNTHESIS is not only the name of the group but also the name of the framework project embracing the main research activities of the group. SYNTHESIS project includes projects being developed by the group in frame of collaborative research and grants obtained.
Research in the area of heterogeneous interoperable information resource environments (HIRE) is focusing on the fundamental issues of semantic interoperability and integration reasoning. Such issues include: reaching of the coherence of the ontological contexts of the heterogeneous information resources (databases, ontologies, services, processes) and of the application; searching for resources and their components that could serve as the concretizations of the application specification; creation of the composition of resources (or their components) that could serve as a consistent, coherent concretization of the application; justifying (proving) that constructed specification is truly a concretization of the application.
A notion of semantically interoperable complete framework (SIRF) is investigated. The notion should provide necessary modeling, methodological and architectural capabilities for analysis, decision making, provable reasoning and implementation based on the pre-existing resource integration and reuse.
The SYNTHESIS group is developing approaches for complete and two-dimensionally uniform specifications of the applications and of the available information resources, for application context definition and coherency reasoning, for support of provable refinement of application specifications by the information resource compositions, for canonical information model development and metainformation support, for HIRE dynamics (multiactivities, workflows) specification, composition and reuse.
Specifications should be complete to provide for reasoning that a resource is applicable to a given problem (perhaps, after some structural, contextual, functional, concurrently behavioral, extensional, etc. reconciliation). Specifications should be uniform horizontally to provide canonical definitions of heterogeneous information resources and vertically to be used for different objectives: for domain analysis and application specifications, for design and for resource to application mappings definitions.
Semi-formal well-defined canonical models (hybrid semistructured/object-oriented and declarative) are being investigated. These models are mapped into purely mathematical formal models supporting provable stepwise refinement of the development process.
Compositional, cooperative infrastructures supporting information systems developers and component providers cooperation in course of component-based information systems construction are emphasized. Information systems formed by interoperation from pre-existing components, integrated distributed multidatabases, subject mediators, workflow compositions are different examples of forms of resource compositions considered.
Metainformation refers to the characterization of information resources for purposes of locating, evaluating, and accessing appropriate sets of resources. The metainformation considered should provide for contextual, canonical and formal specification of applications (mediators) and information resources to be used on different phases of semantic interoperation reasoning. Contextual (ontological) specifications are treated as well organized collections of concept definitions. Concepts are treated as types (in canonical model) to be interrelated by their refinement.
The group is developing approaches and software tools prototypes to support semantic interoperation and integration of distributed information resources including:
- the canonical semi-formal model intended for two-dimensionally uniform specifications of applications (mediators) and heterogeneous information resources;
- a mapping of the canonical model into the formal model-theoretic notation to provide for provable justification of the semantically interoperable design;
- a technology for refining mapping of heterogeneous definitions of various information resources into the canonical specifications (information resource model commutative mapping method based on the well-defined notion of refinement has been developed): the method preserves information and operations of types of specific information models while mapping them into the canonical types;
- a design method specifically emphasizing reuse of the pre-existing components; the output of the existing methods are interpreted and extended (if required) in frame of the canonical SYNTHESIS model;
- a scenario and procedure for the ontological integration of different application contexts based on associations between vectors of verbal terms used for the concepts definitions as well as their refinement relationships;
- modeling and support of metainformation for the broad spectrum of heterogeneous information modeling (structured, semistructured and object data, services and concurrent behaviors, ontologies) including interoperation level specifying federated metainformation intended for unified access to multiple resources with relevant information;
- developing of infrastructures for mediating heterogeneous distributed information resources in the interests of problem solving over multiple distributed heterogeneous information resources in e-science.
Currently the SYNTHESIS group is focused mostly on methods and tools for heterogeneous information resources integration applying subject mediation methodology. Application-driven approach is emphasized (a description of an application subject domain (in terms of concepts, data structures, functions, processes) is created, into which resources relevant to the application should be mapped) in contrast to resource-driven approach (when an integrated schema of multiple resources is created independently of specific applications). Application-driven approach assumes creation of subject mediator that supports an interaction between an application and resources on the basis of the application domain definition (description of the mediator). Various methods and tools have been developed recently. Briefly main results were summarized in the paper and slides available at our site (Mediation Framework for Enterprise Information System Infrastructures).
The results of this work are planned to be applicable in e-science, specifically for Virtual Observatory developed for astronomers.
Recent collaborative projects in which the SYNTHESIS group participated are listed below:
- Object model integration, project with GTE Labs (Michael Brodie), 1992 – 1993
- Semantically Interoperable Information Systems, INTAS Project, GMD IPSI (Coordinator: E.Neuhold; W.Klas, P.Fankhauser, K. Aberer), PRISM laboratory, France (George Gardarin), Kiev State University (N. Nikitchenko), 1995 – 1997
- Compositional software development using pre-existing heterogeneous information resource. Research project with the Siemens Software Arch. and Reuse Department, 1994 - 1997
- Modelling and management of semi-structured data for dynamic world-wide-web applications, INTAS Project, Aristotle University of Thessaloniki (Coordinator: Y.Manolopoulos), Rome University III (P. Atzeni, G. Mecca, R. Torlone), St. Petersburg University (B. Novikov), Yerevan University (M. Manukyan), 2000 – 2001
- DELOS Network of Excellence, 2000 – 2003; DELOS WG on Ontology Harmonization (Martin Doerr); DELOS WG on Metadata Registries (Thomas Baker)
- UNESCO project: Digital Libraries in Education, groups from USA and Europe coordinated by L.A.Kalinichenko, 2002
- Various research projects granted by the Russian Foundations for Basic Research (RFBR) and the Russian Academy of Sciences (RAS)