The goals of this research are:
To study the expressive power of logical query languages. We have
been able, so far, to understand some aspects of the expressive power
of pure Datalog. For Datalog queries we have been able to point out
the role of persistent variables. Furthermore in our research, the
syntactic complexity of a Datalog query is associated in certain ways
with the computational complexity of the query. More specifically:
a)we found examples of 1VC computable Datalog queries which are
not expressible in linear Datalog, b)we found examples of queries
that require arbitrarily IDB-predicate width, in order to be expressed
in Datalog.
Another issue that we are interested in, is to study the parallel computational complexity of simple (e.g. one recursive rule) Datalog queries. For certain nontrivial classes of queries, we establish a complete characterization that separates NC queries from P-complete queries.
"Knowledge Bases, Representation, processing and Dissemination" funded
by the Greek ministry of Industry, General Secretariat of Research and
Technology (1993-1995)
Participants: Afrati, Kavadias, Mihail, Theodoratos, Mitakos
The term "Knowledge" is used mainly to characterise (identify) all these
systems that support a declarative instead of a procedural language. We refer
to the systems that use declarative languages and have a Data Base
Management System (DBMS) as "Knowledge Bases". The notions of
representation, processing and disseminetion of Knowledge are very closely
related to each other. Other relevant topics are metareasoning, theorem-
provers, abduction, constraint logic programming, default reasoning,
semantics. Some research areas, related to this subject are:
i) expressive power of logic programming languages and query
languages
ii)computational results related to default reasoning and circumscription
iii)semantics modification.
We intend to conduct high level original research in the general framework
of Knowledge Bases. The final results will be published in the best
magazines and presented in conferences.
Node of the Esprit Network of Excellence "Compulog Net"
COMPULOG NET is the Network of Excellence in Computational logic
supported by the ESPRIT Program of the European Community. It started
on April 1991 with 17 nodes and now counts 72 nodes. The scientific
objective of COMPULOG NET is to lay the foundations of an integrated
software environment for building knowledge rich paradigm with
enhancements from the areas of databases, artificial intelligence and logic.
The main functions of the network are to:
i)coordinate research activities and infrastructure facilities;
ii)allocate and coordinate research training fellowships;
iii)promote the exploitation of results by European industry;
iv)promote worldwide cooperation.
The research activities of the network are organized into six main areas:
i)Constrain logic programming
ii)Knowledge bases
iii)Knowledge representation and reasoning
iv)Program development
v)Programming languages
vi)Parallelism and implementation technologies.
"Constraints and Extended Support for Storage and Access"
Esprit Basic Research working group (1993-1996)
Participants: Afrati, Theodoratos, Kavalieros, Stavropoulou, Kosmopoulos.
The CONTESSA Project is a proposal for a Basic Research Project (ESPRlT Area 6). The main theme of the project is integration of Constraint Programming and Databases. More specifically, the main objective of the project is to make available database technology for application areas such as Geographic Information Systems GIS and possibly other application. The constraint database project will study the storage of constraints in databases, the use of constraint handling techniques in database query and update, and the incorporation of constraints in object data modelling and database specification. The constraint database approach will be tested mainly in GIS applications. Two additional applications, i.e. Tutoring and Scheduling, are included to broaden the basis for verifying the concepts, models and prototypes to be developed. The formal properties of constrains in Databases and database queries will be studied; their role in query optimization; the consequences of stored constraints on database indexing; constraints on between constraints as data and integrity constraints.
"Synthesis and Transformation of Logic Programs", network of the EC
"Human Capital and Mobility" program (1993- )
Participants: Afrati, Mitakos, Kitsiou, Stavropoulou.
The main goal of this project is to produce reliable and efficient software
by logic based synthesis and transformation. The general framework is that
of Computational Logic. Our focus is on synthesis based on specifications
expressed in logic and producing executable Horn clause programs meeting
these specifications. For transformation, our focus is on optimization of
Horn clause programs and on transforming them to existing Logic
Programming languages, such as Prolog and Goedel. The Human Capital
and Mobility Programme forms an ideal setting to successfully implement
this. Mobility and advanced training of researchers throughout the European
Community is necessary precondition for achieving a strong integration
within this area of research. With respect to program synthesis the main
technical advances pursued in this project are: The improvement of the
tactics and search strategies applied in the underlying proof methods and the
use of meta-logic in this contex. The comparison and integration of different
synthesis methods. The integration of data abstractions. The applicability to
advanced problems. Automatic program development for real-life
applications requires the synthesis system to contain large quantities of
knowledge. The problems of organising and structuring this knowledge need
to be addressed. In the context of program transformation, the main
technical objectives are: Further automation of transformation systems for
pure logic programs. The extension of the above transformation systems to
existing Logic programming languages such as Prolog. The study of
applicability of the developed techniques and systems to advanced
applications, such as: optimization of meta-programs and compiler
generation.
"Database Atlas for Analysing the Molecular Genetics of mammalian
Development" network of EC Biotechnology program (1993-1996)
Participants: Afrati, Mitakos, Kitsiou, Kyriazis
Our key and original intention is to produce a database of molecular-
genetic information that is graphically linked to a 3-D atlas of mouse
development (the model for all mammalian embryos that can be studied in
detail). This tool will create a focus for integrating all the genetic data of
mammalian development and its use will lead to a better understanding of
the normal development and congenital disorders of all mammals, including
farm animals and humans.
The database will be in two linked parts:
The first will be a digitised, 3-D atlas of the daily development of the
mouse embryo (7- 14 days; at 5·m resolution) with major tissues
labelled with their anatomical name, cell types, and origins. Software
will enable the anatomy to be displayed in any orientation and cut in
any plane. This "atlas" will be available on a range of media
including CD-ROM .
The second will include molecular data (e.g. from in situ
hybridization, transgene expression, and enhancer traps) mapped to
the anatomy in both graphical and non-graphical format. The
management system will include interrogation and display software.
The database will initially include data on representative genes, but
the major input of data will be done later.
Software will be developed for image reconstruction, screen display, and
database design.