** **

**
**

**
**

**
**

**Acknowledgements**

I am still surprised that
fundamental mathematical objects, functions and properties in either of their
usual appearances, analytic or determined by some specific graphs (i.e. specific
sets of the form {(*a,f*(*a*))| with *a* passing through a specific set of values}/ properties represented
by tables-matrices in data bases) could be useful in modelling of the phenomena
of the real world. Analytic versus experimental: dealing with analytic forms of
these objects, with accented syntax form, or data specified form with semantic/experimental
- sensor originating, perhaps should not be as distant
as contemporary academic practice, divided into nicely organised fields shows.
In GIS, we contributed serious quantity of work to the processing of both
syntax and semantics and enjoyed in both a lot.

When I was a 2. year student on our
Math Department one of the exam problems in the course of Analysis used to be
expansion of functions in Fourier series and for quite nicely tuned input
functions students needed roughly one hour to calculate first four or five
Fourier expansion coefficients, where up
to half members used to be zero, sometimes guessing the form of the n-th member when it was easy. That procedure was purely
analytic. Now days we have numeric algorithm – fft
and for nice resolution spectra, e.g. of 32k points in perceptual or sub
perceptual time, i.e. within a 1/20^{th} of a second calculation of the
next, on 64 input channels, we have some 40M coefficients calculated in a
second multiplied 1K, which gives some 40 billion times faster calculation of
the Fourier expansions than in the example of the beginning of this paragraph.
Simple function Fourier development might result in simple Fourier series and
analytic method might be fine when we do not need real time. However, when
dealing with just acoustic phenomena, present in neuroacoustic
experimentation or in submarine echo sound analysis, we are approaching the
need to resolve –explode the experimental functions with no perceived analytic
form, within perceptual real time with parameters as in the above calculation.
Possibility to do that in real time might mean possibility to detect some
signal component of crucial importance. Recent software radio based
applications in digital signal processing improve the real time demands by some
orders of magnitude.

That gives example of Fourier -
Hegel “transition of quantity into quality” that is really exciting.

Our Group for Intelligent Systems
originates in joined inspirations with my colleague and friend Žarko
Mijajlović, with whoom I had a number of pritty mental interactions, starting
with application of quantifier ellimination for fields in polynomial
optimisations in 1976, then developping a system supporting a cumulative construction of Turing machines, end of 70’s, with
the desire to materialize beautiful achievements of Godel,
Tarski, Turing and other greats of formal thought and logic, involving syntax
and semantic problems and implementations, leading to a number of provers,
proof checkers, analysis of syntax chains of more complex organization. The
work that we initialized alone, soon involved our graduate and undergraduate
students in early 80's, E. Ljubisavljević, M. Lazić, M. Živković, T. Petrović,
Z. Ognjanović, M. Živković2 and others, and that became a kind of nice and
longer lasting fusion of energy that lead to the number of materialzed
solutions that practicaly showed what levels of a problem were feasable within
some reasonable portion of remaining future and what parts of algorithms would
evolve too slow or explode in complexity so fast that, such as they are, they are too far from any useful
application. While teaching Mathematical logic in CS and Computer applications,
a chance opened to atract greater number and groups of students into the syntax
processing projects: Quantifier elimination in algebraicaly and real closed
fields, with M. Ristić and G. Aleksić, Monadic calculus with equality that
formalized Aristotelian logic, initiated with the big group of students lead by
S. Milutinović and B. Blečić, provers in Intuitionistic logic, with G. Lazić
and M. Radić, provers in Modal logics, provers based on interpretations, with
M. Petrović and N. Lazarević, prover in Geometry, S. Kordić and P.Janičić and a
variety of syntax pattern recognition algorithms involving, besides a number of
students from our School of Mathematics, students from Biology Department S. Stamenkovic, M. Pjanić,
T. Tošić, S. Drakulic, M. Etinski and many others. Then with T. Petrović,
S. Lalević and I. Katanić
we completed the first works on syntax parallelisation. Our more recent works
included: correctness of formal specs, I. Mijajlović,
eccology for Kinematic
Turing Machines, N. Lazarević, V. Nikolin,
S. Tomić, D. Predić. D. Brzaković, optimised
algorithms for Quantifier elimination in Monadic logic with equality and an intelligent interface to the
relational data bases, involved a group of
almost 30 students where V. Aleksić, M. Lukić, M. Marić, F. Marić, R. Jovanović, M. Miličić, N. Dobrašinović, G. Marković, O. Jovanović, Nemanja Djurić, Vlada
Perović did the major work and integrated intelligent
“Cracken” prover; based on interpretation of theories and quantifier elliminatioon algorithms, provers
and syntax pattern recognition, engaged A. Perović,
N. Stefanović, M. Borovčanin,
M. Milošević, D. Ilić, M. Filipović, M. Udovičić, M. Jovanović, D. Doder, M. Marić and F. Marić.

Parallely, we have been involved
in the work on image and signal
processing, starting with NMR deconvolutions, end of
70’s through mid 80’s that lead to the integration of NMR tomography, with
S. Macura and Z. Zsolt,
fast imaging algorithms, spectral analysis and pattern recognition in
astronomy, Transputer parallelized multiple fft, T. Petrović,. S. Lalević
and I. Katanić, computerized EEG and other
biomedical signal analysis and CCD imaging in microscopy, with with S. Kordić, Z.
Obradović, G. Popović, S. Malkov, G. Lazić, I. Jovanović. That work continued
with V. Todorović, N. Andonovski, D. Perišić, G. Marković, R. Todorović, M.
Bulat, V. Ivandski, A. Uzelac, M. Oklobdzija, M. Jovanović, M. Marić, M.
Borovčanin, A. Perović and A. Jovanović, jr.
This portion of engagement of our Group is presented in this volume.

All and all, more than 200 students more or less
seriously participated in these activities.

Four groups of students received awards
from University of Belgrade for the research work of young scientists, works of
two teams were awarded by Engineering congress ETRAN for the works in
Artificial Intelligence. Great number of our coleagues continued their careers
in computer industries, a fraction continued their careers in research-science,
many all over the world. Besides software implementations, a number of papers
was generated and presentations on conferences.

I owe highest thanks to the above
mentioned coleagues and many others who contributed seriously to the
developements undertook in GIS over the mentioned longer period of time with
joy and enthusiasm and would be happy if my catalising initiatives and efforts
were useful in their later destinations.

I owe acknowledgements as well to
Dr. Dick Treffers from UCB, who introduced me in ccd imaging of quasars in his first robotic astronomic
observatory and Dr. Istvan Vince from Belgrade
Astronomic Observatory, that armed me with courage to make our first CCD
microscope for dot counting- which is how our applied developments evolved
further, here in 93. It was in the labs of Dr. Selma Kanazir,
molecular biology, Dr. Mladen Vujošević,
genetics, Dr. Rosica Veskov,
rat EEG, all from the Institute for Biology Research “Siniša
Stanković” in Belgrade, Dr. Vlada
Radivojević, human EEG, Institute for mental care,
Belgrade, prof. Nina Japundžić-Žigon,
Institute for farmacology and toxicology, School of
Medicine, Un. of

prof. Dragan
Marinković and prof. Vlada Kekić,

Portion of these works was supported by
3.5 projects of Ministery of Science and technology of Serbia, in the period
from 1994. till 1999. which helped it reach a kind of polishing that made them
look professional and useful to the intended application targets - our partners
in those projects.