|
|
Informatics
In 1962, a laboratory was organized in the Institute; and one of the tasks
of that new laboratory was to develop external devices of computers, in
particular, devices for information exchange between a man and computer.
Main attention was paid to means of intercourse with computer with the aid
of devices of recognition and synthesis of speech signals. Some time later
this laboratory headed by doctor of technical sciences N. G. Zagorujko
grew into one of leading scientific schools in the field of speech technologies.
On its basis, a unique scientific-organizational formation emerged: All-Union
summer school-workshop on the problem “Automatic recognition of hearing
Patterns” (ARHP), which joined efforts of specialists working over problems
of speech communication in various scientific fields: philologists, physiologists,
psychologists, engineers, and mathematicians. In the period from 1965 till
1972, 17 schools-workshops ARHP were held, and representatives of more than
150 organizations took part in them. The results of the Laboratory in the
field of speech investigations were marked in 1992 by the First Diploma
on the competition of works in the field of speech recognition and synthesis
that was held by Hewlett-Packard among collectives from the USSR and countries
of East Europe. Investigations of problems of speech recognition naturally
grew into investigations of general problems of Pattern recognition. Then
it became clear that the problem of recognition is just one of the problems
of a wider and new field of applied mathematics, field of Data Mining.
The basis of this field is mathematical methods of automatic discovering
of regularities concealed in large volumes of information. Works in this
direction developed in the frames of the Department of Informatics and
are represented by a wide range of investigations and developments beginning
from methodological foundations and mathematical problems, via algorithms
and programs, to applied software used in geology, medicine, economics,
sociology, and many other fields.
Applied software package OTEKS
for analysis of experimental and statistical data was handed over for
usage to more than 100 organizations. Versions of this package were awarded
two silver medals of the Exhibition of National Economic Achievements
of the USSR.
Scientific results obtained in the collective served as a basis for many
doctor and candidate theses. Several researchers who worked or now work
in the Department became doctors of sciences: G. Ya. Voloshin,
G. S. Lbov, K. F. Samokhvalov, V. I. Kotjukov,
D. I. Sviridenko, A. V. Kel’manov, Yu. A. Ustjugov.
Candidates of sciences V. D. Gusev, V. A. Skorobogatov,
V. N. Elkina, L. I. Makarov, E. P. Zolotukhin,
and others work fruitfully and head scientific collectives. Many learners
of our scientific school work in New Independent States and foreign countries:
Ukraine, Uzbekistan, Tajikistan, USA, Israel, Germany, Australia, and
New Zealand.
Discovering of regularities (machine learning) is the central problem
in the field of artificial intelligence. Developments of systems of artificial
intelligence that are carried out in the Department, in particular, development
of expert systems (ES), are connected with that problem. Instrumental
system EKSNA was created, which allows to construct applied expert systems
of new generation: expert systems of partner type (PS). PS can make profound
analysis of data, extract knowledge from data, measure distances between
knowledges, and on this basis automatically find contradictions between
knowledges, structurize the knowledge base, simulate the human way of
reasoning “by analogy”, etc.
Theoretical investigations are accompanied by solution of various applied
problems both in those areas where information is presented in the form
of tables “object-property-time” and in those ones where the source data
have the form of real objects or processes. Here one has to know how to
measure informational signs, gather tutorial material, and only then one
can apply methods of data mining: classification, recognition, prediction,
etc. Among such areas the following ones should be noted:
- analysis and recognition of speech signals. In the
recent time, the works in this direction are carried on under guidance
of doctor of physical and mathematical sciences A. V. Kel’manov
and are directed mainly at development of recognition methods that are
stable with respect to strong acoustical noise and signal distortions
in communication channels;
- analysis of symbolic sequences. This work is headed
by candidate of technical sciences V. D. Gusev. Effective
methods of discovering of regularities in symbolic sequences of large
length were developed; these methods are used in analysis of genetic
and musical texts;
- analysis of structured objects. Under guidance of
candidates of technical sciences V. A. Skorobogatov and L. I. Makarov,
methods for analysis the structure of molecules of organic compounds
were developed, which allows to predict biological properties of compounds
by peculiarities of their structure.
In recent time, works were begun in the Department over the problems
of data and knowledge analysis arising in investigation of problems of
sustainable development.
Now 15 researchers, including 3 doctors and 9 candidates of sciences,
work in two laboratories of the Department.
Each year 8-10 students, magisters, and postgraduates of NSU (Novosibirsk
State University) work on in the Department. Over 200 graduates of the
Chair of Theoretical Cybernetics were prepared during the period of existence
of the Department.
|
 |
