GAN-technologies in intelligent environments of the subject area
The paper considers the problem of applying GAN-technology in intelligent domain environments (IDEs) for the need to structure large amounts of data and generate initial knowledge. У статті розглянуто застосування технології GAN в інтелектуальних середовищах предметної області (ІСПО) для структурува...
Gespeichert in:
| Veröffentlicht in: | Проблеми керування та інформатики |
|---|---|
| Datum: | 2025 |
| Hauptverfasser: | , |
| Format: | Artikel |
| Sprache: | Ukrainisch |
| Veröffentlicht: |
Інститут кібернетики ім. В.М. Глушкова НАН України
2025
|
| Schlagworte: | |
| Online Zugang: | https://nasplib.isofts.kiev.ua/handle/123456789/211404 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| Назва журналу: | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| Zitieren: | GAN-technologies in intelligent environments of the subject area / О. Gorda, Yu. Riabchun // Проблемы управления и информатики. — 2025. — № 3. — С. 85-95. — Бібліогр.: 8 назв. — англ. |
Institution
Digital Library of Periodicals of National Academy of Sciences of Ukraine| _version_ | 1859987296860766208 |
|---|---|
| author | Gorda, О. Riabchun, Yu. |
| author_facet | Gorda, О. Riabchun, Yu. |
| citation_txt | GAN-technologies in intelligent environments of the subject area / О. Gorda, Yu. Riabchun // Проблемы управления и информатики. — 2025. — № 3. — С. 85-95. — Бібліогр.: 8 назв. — англ. |
| collection | DSpace DC |
| container_title | Проблеми керування та інформатики |
| description | The paper considers the problem of applying GAN-technology in intelligent domain environments (IDEs) for the need to structure large amounts of data and generate initial knowledge.
У статті розглянуто застосування технології GAN в інтелектуальних середовищах предметної області (ІСПО) для структурування великих обсягів даних і генерації початкових знань.
|
| first_indexed | 2026-03-18T08:07:15Z |
| format | Article |
| fulltext |
© О. GORDA, YU. RIABCHUN, 2025
Міжнародний науково-технічний журнал
Проблеми керування та інформатики, 2025, № 3
85
UDC 004.8:004.032.26:004.93
О. Gorda, Yu. Riabchun
GAN-TECHNOLOGIES IN INTELLIGENT
ENVIRONMENTS OF THE SUBJECT AREA
Olena Gorda
Kyiv National University of Construction and Architecture,
https://orcid.org/0000-0001-7380-0533
anaelg@ukr.net
Yuliia Riabchun
Kyiv National University of Construction and Architecture,
https://orcid.org/0000-0002-8320-4038
super.etsy@ukr.net
The paper considers the problem of applying GAN-technology in intelligent do-
main environments (IDEs) for the need to structure large amounts of data and gen-
erate initial knowledge. Obtaining new knowledge, which is not explicitly formed
by experts, based on available data is an important aspect. The purpose of the
study is to determine the features of GAN-technology in IDEs, their close
integration with the system of concepts of the subject area expressed in the form of
an ontology, the development, use and development of which can be considered as
support for the intelligence and development of IDEs as a system. The correlation
of cognitive and semantic processes of IDEs is studied to determine the ontological
basis of the possibilities of ensuring the evolutionary properties and adaptability of
creating the structure and training of neural networks (NN) as NP-complex tasks.
The main difference of the study is the cognitive-semantic analysis based on the
theory of categories, mathematical logic and universal algebra, relational algebra,
namely, the transformation of the ontological dictionary, ontological constructions
in open languages of knowledge representation, which makes it possible to apply
promising methods based on metaheuristics, the mechanism of which can be rep-
resented by a problem-independent high-level algorithmic structure in the form of
a set of guiding principles or strategies. GAN-technologies in IDEs are defined as
an ontological basis, taking into account the fact that evolutionary properties are
taken into account in optimization and adaptability methods. Metaheuristics, based
on the metaphor of a natural or artificial process, as the basis for a learning scheme
for different groups of features, is an effective basis for solving specific problems
of evolutionary origin.
Keywords: information, information object, evolution, GAN-technology, intel-
lectual environment of the subject area, optimization, ontology, artificial intelli-
gence.
Introduction
The relevance of the study is due to the growing interest in the problem of data inte-
gration in various fields of activity related to the accumulation and efficient use of infor-
mation, which makes it a popular area in the modern information industry in such areas as
modelling:
• monitoring control systems — a type of automatic control system in which the
type of control influence is not known in advance;
• self-adjusting systems — having the ability to automatically change parameters
or structure during operation in order to maintain the specified indicators of quality and
efficiency of control in arbitrarily changing external conditions;
mailto:anaelg@ukr.net
mailto:super.etsy@ukr.net
86 ISSN 2786-6491
• in nonlinear models of object diagnosis based on component models for approx-
imating nonlinear functions, the mathematical basis of neural network theory is laid
down, which determines the universal approximation properties of neural networks.
At the moment, knowledge as such is becoming more and more resource-based.
The resourcefulness of knowledge lies in its definition as a set of information (data and
facts) with their value for intellectual environments, i.e., knowledge is structured infor-
mation of intelligent domain environments (IDEs) in conjunction with an assessment of
its significance for a given IDE, while there is uncertainty of the information task itself
and its formulation. The relevance of the ontological analysis of neural networks (NN)
learning based on GAN-technology in IDEs is to ensure the transfer of an extremely
large amount of knowledge and experience to users, the creation of a single distributed
information environment within IDEs.
The primary problems when analyzing the application of GAN-technology in IDEs
are the need to structure large amounts of data and to form initial knowledge. Obtaining
new knowledge, which is not explicitly formed by experts, based on available data is an
important aspect.
The purpose of the study is to determine the features of GAN-technology in IDEs,
their close integration with the system of concepts of the subject area, expressed in the
form of an ontology, the development, use and development of which can be considered
as support for the intelligence and development of IDEs as a system.
1. Literature review
The research presented in this paper was initiated and conducted in [1] and is its
continuation and development [2].
The advantages and specifics of the ontological approach are the provision of flexible
data modeling based on semantic technologies, which allows for the analysis of unstructured
information and intelligent information retrieval in heterogeneous sources. Today, there are
acute problems related to machine learning, which provides analysis and classification of da-
ta in conditions of incomplete information, knowledge management, ensuring the continu-
ous generation of new knowledge, and data analysis. For example, ontological research
methods have been conducted within the framework of knowledge engineering [3–7], in da-
ta mining [8–10], in neural network training [11, 12], in automated knowledge assess-
ment [13–15], and in artificial intelligence — knowledge clustering [16, 17]. Procedures for
creating new software concepts based on an ontological approach are discussed in [18].
Based on the results obtained, it becomes possible to define, describe and study the
formalization, to describe the problems that arise in the process of formalization and
analysis.
2. Research methods
The study uses ontological modeling to represent knowledge about the subject area
in the form of formal structures. The main thing in the construction of mathematical
models of ontology is the formalization of knowledge about the subject area while en-
suring the required level of validation and verification — the model must be checked
for compliance with knowledge about the subject area and for compliance with formal
rules and constraints. To solve this problem, we will apply an approach based on IDEs.
The IDEs is a general attitude, awareness and practical actions, measures and pro-
cedures of a generalized subject of a given generalized subject area (universal) aimed at
establishing and maintaining the intellectual system of each subject in relation to its
subject area as a subfield of the universal.
It should be especially noted that in the intellectual environment on the set of
knowledge there is a function of the importance of different knowledge in relation to the
subject of the subject area.
Міжнародний науково-технічний журнал
Проблеми керування та інформатики, 2025, № 3 87
There are relations in the IDEs information environment: accessibility, reliability,
completeness, accuracy, relevance, usefulness of information, which determine the for-
mation of conventional knowledge and conventional concepts in IDEs. At the same time,
the stages of generation and evolution of conventional knowledge are presented in the
form of classification systems, dictionaries, thesauri and other types of ontologies with dif-
ferent degrees of formalization of methods of representation in artificial intelligence (AI).
3. Main results
We further use the following information objects of IDEs:
НО (reflection and formalization, image-traces and memory, cognitive, observa-
bility in data-channels, part of space
0,Loc time interval of existence, construction and
modeling of connections between internal and external elements of images, with the
model being the generation of knowledge or connections);
0 0( , ) ?;HO Loc T +
( , );Ob Loc T
0( );K HO
( ) ;HO Ob
1 2 1 2 0{ },{ }, ;K K K K K =
1 2( ) , ( ) ;Out K Out K
1 2 0( ) ( ) ( ).K Ob K Ob K Ob =
• InPPO — information space of the subject area.
• PrInPPO — attached information space of the subject area.
• InPPO {{IE},{IO},{Information search},{Information relations},{Ontologies},
{PrInPPO},{InPPO entities},{InPPO information technologies}} — structure.
• IE — information units.
• IO — information objects.
• InDEs — software information system — homogeneous with respect to infor-
mation in terms of validity or truth.
• The structure of InDEs {{InPPO}, {Ontologies}, {InPPO entities}, {PoC crite-
ria}, {Objectives1}, {Knowledge system}}.
• The structure of the IDEs is heterogeneous with respect to information in terms
of validity or truth.
• IDEs {{InDEs}, {IDEs Ontologies}, {IDEs Subjects}, {IDEs cognitive tech-
nologies}, {IDEs information sources}, {IDEs information anachronisms}}.
Fig. 1
Thus, let us represent the structure of the IDEs ontology as:
— Structure of the IDEs ontology {{Information Processes Ontology}, {IDEs
IDEs Ontology}, {PrInPOC Ontology}, {IDEs Subjects Ontology}, {IDEs Tasks On-
tology}};
Information processes in software
— dialog exchange on
knowledge system;
— knowledge transfer;
— knowledge storage;
— knowledge loss
— optimizations;
— evalutions;
— problem statements;
— information retrieval
— formalizations on IE and IO;
— cognitive relations on IE and
IO and information tasks;
— loss of relevance on IDEs
88 ISSN 2786-6491
The IDEs(PO) subject domain is synthetic.
The subject domain model (SDM) consists of issues, problems, tasks, subjects, on-
tologies, information units, thesauri, information sources, cognitive units.
SDM ≝ {Information environment, {Ontologies}, {Thesauri}, {Objects {Prob-
lematics} {Problems} {Tasks} {Subjects }}.
To analyze GAN-technology in IDEs, let us consider the following IDEs infor-
mation objects, and by IDEs data we will understand images in IDEs as a generalization
of the concept of digital data.
— Forming the data path. Formalization as image.
— Channel image (Im).
• Channel characteristics:
• color space;
• sampling (m × n);
• im morphology;
• im topology;
• im dimensions;
• im color atlas;
• distortion and noise on Im;
• pixel loading.
Image elements (Im) include:
• light source area — area of maximum white;
• shadow area — area of darkening due to gradient from light with localizations
behind objects;
• reference lines — fixed lines in the image, such as the horizon;
• reflexes — areas in the direction from an object toward another object that have
the tonal color of the first object and are located on the second object;
• proportionality — geometric proportionality of identical objects on different im-
age plans in the direction of perspective;
• contiguity of image spots — as a division of a monolithic image into homogene-
ous spots;
• image loading — is the ratio / ,m n M where M — number of image objects;
• image homogeneity — is ,
ii
WP
KP
where iWP — loading of i-th plan, KP —
number of plans;
• for a discrete image, image heterogeneity is
i ji j
GO M
n m
, where iGO is the
number of pixels of the object area boundary;
• mage (im) as visualization is text from the point of view of perception semantics
(channel);
• interval data from channels (data set) is language alphabet;
• constructions from data (words and sentences) are language syntax;
• constructions and subject area — language semantics;
• problematics («wants») of the observing object — the basis of the subject area;
• information environment as an element of the subject area;
• ontologies and thesauri of the observing object — as elements of the subject area
(in the form of corresponding profiles);
• GAN and GA — mechanisms of development of the subject area (as knowledge
transfer and mutations);
Міжнародний науково-технічний журнал
Проблеми керування та інформатики, 2025, № 3 89
• image ontology in IDEs;
• learning in IDEs, Im — at the basis of pattern recognition and classification into
{im{}};
• object of observation;
• conditions of observation;
• characteristics of the recording equipment;
• distortions of the recorded data;
• purposes of image use;
• problem/tasks to be solved (calculation, finding an unknown parameter, optimi-
zation, design of the image);
• necessary data set for solving the problem — sufficiency;
• definition of the class of problems to be solved;
• modeling on the class of problems.
Machine learning models learn annotations of training data in order to process
new, unlabeled data in the future. In supervised learning, humans label the data and tell
the model exactly what it needs to find. However, to tell the model what to find in that
data, you must add annotations. Labeled data is data that has been augmented with la-
bels/classes that contain meaningful information.
In unsupervised learning, humans feed the model raw data without labels and the
model finds clusters in the data. Blind stages are multiple passes performed by different an-
notators. Blind stages are annotation tasks in which multiple humans (or models) assign la-
bels independently of each other; a stage is only considered passed if they all converge on
the same result. Blind stages are used to create ultra-precise training and automation data.
Annotators often miss objects, but two or more annotators are less likely to miss
the same object. Markup in blind stages is done in parallel and each participant cannot
see the progress of the others.
When all annotators have completed their version of the task, it goes through
a consensus check validating that the annotators have agreed on the same solution,
while ensuring relevance, consistency, homogeneity, completeness.
In this case, by NN training tasks we will understand:
1. Create a new .
2. Create a duplicate (not a copy) in the given subject area
3. Supplement the training with new knowledge
4. Simplify the existing one (change the task field — narrowing)
5. Create a new NN based on a set of {NN}s.
6. Roll several into one (getting superexpert)
7. Incorporating new precedents and new knowledge into the appeared new prece-
dents and new knowledge
8. Take into account the emerged new tasks in the subject area
9. Consider the emerged new in teaching the old
10. Develop the NN through its additional learning {NN}s.
11. Simplifying (optimizing) the NN to reduce the cost of operation.
12. Algebraization of the NN with respect to a fixed task and software.
13. NN algebra on software.
Neural network training is the process of teaching a neural network to solve prob-
lems based on propagation — the process of originating, moving and processing an
event.
Neural networks can come up with ideas and create content, but are not able to
think creatively.
90 ISSN 2786-6491
From the set of deep learning methods — generative-adversarial network, convolu-
tional neural network, recurrent neural networks, recursive neural networks, consider
generative-adversarial networks in IDEs.
Deep learning has a fundamental limitation: the models of deep learning are
limited in what they can represent, and most programs cannot be expressed as continu-
ous geometric morphing of data diversity. Therefore, let’s investigate from the point of
view of IDEs:
What is a good teacher in IDEs? What is a bad teacher in IDEs?
What is a good student in IDEs? What is a bad student in IDEs?
Criteria for evaluating a teacher-student pair in IDEs.
How much should the oncology of the teacher and the student match in IDEs?
Learning process — transfer of knowledge from teacher to student or not (creation
of an action mechanism, consolidation of actions) in the IDEs.
How much is the transfer cognitively in IDEs?
Determine who is the teacher (Tch) and who is the learner (Ln) in GAN IDEs using
a diagram (Fig. 2, 3):
Fig. 2
Fig. 3
Teacher–simulation model in conjunction with feedback on correction (Corr) of
responses by the learner.
: ( ) ( )Corr Ln t Ln t t→ +
( ( )) ( ( ))... ()Crit Ln t Crit Ln t t Crit + (criterion)
( ( )) ( ( ))Mach Ln t Mach Ln t t + (number of «0-elements» in the matrix of links for Ln).
The step Corr value t is determined by:
— Calibration test sample (Test), (CTS — Calibration Test Sample) and its proper-
ties for training.
The calibration tests { } { }iC Test NN are conducted in a similar manner { }.Tsk
— Questions C{ }Tsk should be correctly formulated and have correct answers in
advance.
— { }iNN relatively structured { }.Tsk
— Value { }.Test
Teacher (Tch) Information class
Generator Learner (Ln)
Tch Correct (Corr)
Test Ln Answer
t
t + t
Міжнародний науково-технічний журнал
Проблеми керування та інформатики, 2025, № 3 91
— Structured { }.Test
— Test set properties { }.Test
— The teacher (Tch) generates a student expert (Ln) on a class of tasks (class) by
teaching him (on examples and their recognition) at the expense of:
— Number of iterations ( ).n t
— Number of imitations { }.Test
— An image of the learning goal (Ob).
— The learner (Ln) should be better than the teacher (Tch), otherwise you can just
copy Tch.
— The measurable quality of learning function of NN: EF EF ({learning
sample}, {learning pairs — teacher}, {learning quality criteria Crit}, Ln). The com-
parison of the quality of NNs received is carried out according to the following pa-
rameters:
— false classification of the event;
— skipping event classification;
— correct classification of the event;
— wrong classification of the event;
— select NN in the subject area;
— selection of artificial intelligence in the subject area.
• The process of knowledge transfer in IDEs can be considered as an approxima-
tion in the information space to the informational image of the target of the information
object due to the function of the degree of inconsistency of the information objects and
the materiality of the information space relative to targets.
• Approach of heterogeneous objects in IDEs.
Consider teacher — learner P(Tch, Ln) pairs and knowledge transfer — T(Tch,
Ln), where level of knowledge — L(Tch), L(Ln), and tasks — Tsk, as well as learning in
the subject area { }.iTsk
Competencies are thus defined in relation to the objectives of the subject area —
( , { }), ( { } ).i i
j j iC Tch Tsk C Ln Tsk
Evolving subject area as the intellectual environment of the subject area:
— NN neural network — standards implemented by the classifier, as a mechanism
of response to adaptation;
— artificial intelligence AI — is cognition in the subject area as a mechanism of
adaptation and purpose.
Then, development (as a step of evolution) is determined by transfer of knowledge:
AI
1.i iNN NN +⎯⎯→
Define concepts:
— «bad» teacher Tch for this learner Ln regarding concepts: Tsk, Test, T(Tch, Ln),
Cl(Tch), Cl(Ln); Cl( ) — classification, Prpt(X) — property X, { }T — method of trans-
ferring knowledge from Tch to Ln.
— «good» learning — Pr ( ( ) Pr ( ( )) ( , );pt Cl Tch pt Cl Ln T Tch Ln
— no training { }, { }: Pr Pr ( , ).
SC Testi
k
TT T T T pr pt Tch Ln
Consider: Ont(Tch), Ont(Ln) — ontology; Ont0(Tch), Ont0(Ln) — initial oncolo-
gy;
0 0( ) ( )Ont Tch Ont Ln — oncology for these tasks; ( ) ( )k kOnt Tch Ont Ln —
92 ISSN 2786-6491
oncology for these tasks and tests as a result of the transfer of knowledge * ( , ),kT Tch Ln
which has the following properties:
* { }kT T — the result of transfer of learning;
( ) ( );Ont Tch Ont Test
0 0( ) ( );Ont Tsk Ont Ln
( ) ( );Ont Test Ont Ln
{ }T — cognitive technologies for Ln.
Gold data sets are used to compare model performance, compare different models
for greater objectivity when evaluating and comparing different algorithms and ap-
proaches. The main characteristics of gold data sets: accuracy, consistency, complete-
ness, timeliness, free from bias.
To create a golden dataset without using a generative tool is to collect data, clean
the data, create annotations and labels, check relevance, maintain the sets. Meanwhile,
there are challenges of creating golden datasets that need to be overcome to develop
golden datasets resource intensity, bias (impartiality), data privacy.
Let us define the following operations:
• operation (difference) — define a neural network Net based on neural net-
works 1Net and 2 ,Net as a result of their functional, structural and mathematical dif-
ference;
• operation → (pre-training) — train a neural network 1Net to 2 ,Net based on
their own data, such that 1 2,Net Net functionally, structurally or concurrently, wherein
1 Pr, , , ( , )Net NN Ds Dt W NN Ds
NN — neural network topology; Pr — validation of the neural network with respect to
a set of targets; Ds — training dataset; Dt — testing dataset; ( , )W NN Ds — epoch in
neural network training.
Then the transformations in neural network training can be represented as follows:
1) 1 2 2 1( ) ( ) ( ), \ ,Net Ds Net Ds Net Ds Net Net − depending on the training;
2) ( )
1 2 2 1( ) ( ), \ ,NNNet Net NN Net NN Net Net − = structurally;
3) 1 2 2 1(Pr) (Pr ) (Pr ), ,Net Net Net Net Net − functionally enabled;
4) 1 1 2 2 2 1( , ) (Pr , ) (Pr , ), ,rNet P Ds Net Ds Net Ds Net Net − extended validation;
5) 1 1 2 2 2 1( , ) ( , ) ( , ), ,Net NN Ds Net NN Ds Net NN Ds Net Net − equivalence com-
parison based on learnability;
6) 1 1 2 2 2 1(Pr, ) (Pr , ) (Pr , ), ,Net NN Net NN Net NN Net Net − comparison of clas-
ses of problems solved by neural networks;
7) 1 1 1 2 2 2(Pr, , ) (Pr , , ) (Pr , , ),Net NN Ds Net NN Ds Net NN Ds − comparison of
neural networks within the same ontology.
Conclusion
Ontological methods have become integral to data mining, enhancing the extrac-
tion of meaningful patterns by incorporating domain-specific knowledge. These meth-
ods utilize ontologies — structured representations of concepts and their interrelations
within a particular domain — to provide semantic context to data mining processes.
This integration facilitates more accurate and interpretable results, especially when deal-
ing with complex or unstructured data.
Міжнародний науково-технічний журнал
Проблеми керування та інформатики, 2025, № 3 93
The study of the correlation between the cognitive and semantic processes of IDEs
in order to determine the ontological basis for the possibilities of ensuring the evolu-
tionary properties and adaptability of creating the structure and training of NNs as NP-com-
plex tasks is defined:
— GAN-technologies in IDEs as an ontological basis, taking into account the fact
that evolutionary properties are taken into account in optimization and adaptability
methods. Metaheuristics, based on the metaphor of a natural or artificial process, as the
basis for a learning scheme for different groups of features, is an effective basis for
solving specific problems of evolutionary origin;
— visualization in the subject area (SA) — construction of conceptual space in the
intellectual environment of the image space in the general intellectual environment. In
this case, the observing object should possess concepts and perceptual images;
— GAN-technology in IDEs Im — technology of teaching a set of observing ob-
jects at the intersection of intellectual (conceptual) and not intellectual in IDEs;
— initial selection (IS) as a primary teacher;
— GAN-technologies in IDEs make it an evolving one, described on the basis of
the theory of traces and reflections;
— the subject area is generated by the problematics of the observing object (OO);
— the subject area serves to reduce the scope of the problem OO by formalizing
the problems, defining the tasks and actors of the IS;
— the problematics is determined by the need of the OO;
— the need of the OO is determined by its existentiality;
— presence of genetic modifications of GAN-technology in the IDEs.
О.В. Горда, Ю.В. Рябчун
GAN-ТЕХНОЛОГІЇ В ІНТЕЛЕКТУАЛЬНИХ
СЕРЕДОВИЩАХ ПРЕДМЕТНОЇ ОБЛАСТІ
Горда Олена Володимирівна
Київський національний університет будівництва і архітектури, м. Київ,
anaelg@ukr.net
Рябчун Юлія Володимирівна
Київський національний університет будівництва і архітектури, м. Київ,
super.etsy@ukr.net
У статті розглянуто застосування технології GAN в інтелектуальних сере-
довищах предметної області (ІСПО) для структурування великих обсягів
даних і генерації початкових знань. Важливим аспектом є отримання но-
вої інформації (яка не сформована експертами чітко та зрозуміло) відпо-
відно до наявних даних. Визначені особливості GAN-технологій в ІСПО
та їх тісна інтеграція з системою понять предметної області, виражених
у вигляді онтології, розробку, використання та розвиток якої можна роз-
глядати як підтримку ІСПО як системи. Досліджено співвідношення когні-
тивних і семантичних процесів ІСПО для з’ясування онтологічної основи
можливостей щодо забезпечення еволюційних властивостей і адаптивності
створення структури та навчання нейронних мереж (НМ) як NP-складних
задач. Головна відмінність дослідження полягає у використанні когнітив-
но-семантичного аналізу на основі теорії категорій, математичної логіки,
mailto:anaelg@ukr.net
94 ISSN 2786-6491
універсальної алгебри та реляційної алгебри, а саме — у перетворенні он-
тологічного словника та онтологічних конструкцій у відкритих мовах пред-
ставлення знань. Це дає змогу застосовувати перспективні методи, що ґрун-
туються на метаевристиках, механізм яких можна представити як проблем-
но-незалежну алгоритмічну структуру високого рівня у вигляді набору
керівних принципів або стратегій. GAN-технології в ІСПО визначаються
як онтологічна основа з огляду на те, що еволюційні властивості врахову-
ються в методах оптимізації та адаптивності. Метаевристика, що ґрунтується
на метафорі природного чи штучного процесу, як база навчання за різни-
ми групами ознак є ефективною у вирішенні конкретних завдань еволю-
ційного походження.
Ключові слова: інформація, інформаційний об’єкт, еволюція, GAN-тех-
нологія, інтелектуальне середовище предметної області, оптимізація, онто-
логія, штучний інтелект.
REFERENCES
1. Tsiutsiura M., Gorda E. Ontological analysis of cognitive information technologies of the subject
field. Proceedings of 2023 IEEE International Conference on Smart Information Systems and
Technologies (SIST). Kazakhstan : Astana, 04–06 May. 2023. P. 189–192. DOI: http://dx.doi.
org/10.1109/SIST58284.2023.10223487
2. Gorda E., Riabchun Y., Khrolenko V. Cognitive technologies for object detection and topology of
the environment’s information space. Proceedings of 2024 IEEE 4th International Conference on
Smart Information Systems and Technologies (SIST). Kazakhstan : Astana, 15–17 May. 2024.
P. 48–54. DOI: https://doi.org/10.1109/SIST61555.2024.10629620
3. Lovati S. GaN technology: challenges and future perspectives. Power Electronics News.
2 December 2021. URL: www.powerelectronicsnews.com/3-pen-ebook-dec-21-gan-technology-
challenges-and-future-perspectives/
4. Klimova M.V. Development of a method and model of knowledge verification in ontological sys-
tems. Eastern-European Journal of Enterprise Technologies. 2009. Vol. 4, N 8(40). P. 32–36.
DOI: https://doi.org/10.15587/1729-4061.2009.22184 (in Ukrainian) URL: https://journals.uran.
ua/eejet/article/view/22184
5. Huang C.-J., Trappey A.J.C., Wu C.-Y. Develop a formal ontology engineering methodolo-
gy for technical knowledge definition in R&D knowledge management. Collaborative
Product and Service Life Cycle Management for a Sustainable World . Proceedings of the
15th ISPE International Conference on Concurrent Engineering (CE 2008) / ed. by R. Cur-
ran, S.Y. Chou, A. Trappey. London : Springer, 2008. P. 495–502. DOI: https://doi.org/10.
1007/978-1-84800-972-1_46
6. Lovati S. 10 things to know about GaN. Power Electronics News. 2021. URL: https://www.
powerelectronicsnews.com/10-things-to-know-about-gan/
7. Leung N.K.Y. Towards an ontology-based knowledge management: an ontology mediation
framework to reconcile inter-organizational knowledge. University of Wollongong. 2012. URL:
https://hdl.handle.net/10779/uow.27662850.v1
8. Perez-Rey D., Maojo V. An ontology-based method to link database integration and data mining
within a biomedical distributed KDD. Artificial Intelligence in Medicine. Proceedings of the 12th
Conference on Artificial Intelligence in Medicine in Europe (AIME 2009). Italy : Verona, 18–22 Ju-
ly / ed. by C. Combi, Y. Shahar, A. Abu-Hanna. Berlin; Heidelberg : Springer, 2009. P. 355–359.
DOI: https://doi.org/10.1007/978-3-642-02976-9_48
9. Combining machine learning and ontology: a systematic literature review / S. Ghidalia, O. Lab-
bani-Narsis, A. Bertaux, Ch. Nicolle. Computing Research Repository (CoRR). 2024. 37 p. DOI:
https://doi.org/10.48550/arXiv.2401.07744
10. Gorda O.V. Аnalysis of the task of building an ontological dictionary of construction. Applied
Geometry and Engineering Graphics. 2021. Vol. 101. P. 55–95. DOI: https://doi.org/10.32347/
0131-579X.2021.101.55-95 (in Ukrainian)
11. Universal representation learning of knowledge bases by jointly embedding instances and onto-
logical concepts / J. Hao, M. Chen, W. Yu, Y. Sun, W.Wang. KDD’19: Proceedings of the 25th
ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. USA : An-
chorage, 2019. P. 1709–1719. DOI: https://doi.org/10.1145/3292500.3330838
http://www.powerelectronicsnews/
https://doi.org/10.15587/1729-4061.2009.22184
file:///D:/Downloads/10.1007/978-1-84800-972-1_46
file:///D:/Downloads/10.1007/978-1-84800-972-1_46
https://www.powerelectronicsnews.com/10-things-to-know-about-gan/
https://www.powerelectronicsnews.com/10-things-to-know-about-gan/
https://doi.org/10.1007/978-3-642-02976-9_48
https://doi.org/10.32347/0131-579X.2021.101.55-95
https://doi.org/10.32347/0131-579X.2021.101.55-95
https://doi.org/10.1145/3292500.3330838
Міжнародний науково-технічний журнал
Проблеми керування та інформатики, 2025, № 3 95
12. Ontology formation of support system for restoration of buildings, property and infrastructure ob-
jects / S. Terenchuk, A. Buhrov, R. Pasko, A. Yaschenko, I. Bosenko, B. Volokh. 2023 IEEE
4th KhPI Week on Advanced Technology (KhPI Week). Ukraine : Kharkiv, 2023. 5 p. DOI:
https://doi.org/10.1109/KhPIWeek61412.2023.10313006
13. Information system of multi-stage analysis of the building of object models on a construction
site / S. Dolhopolov, T. Honcharenko, O. Terentyev, K. Predun, A. Rosynskyi. 4th International
Conference on Sustainable Futures: Environmental, Technological, Social and Economic Mat-
ters (ICSF-2023). Ukraine : Kryvyi Rih, 2023. IOP Conference Series: Earth and Environmental
Science. 2023. Vol. 1254, N 1. ID: 012075. 10 p. DOI: http://dx.doi.org/10.1088/1755-1315/
1254/1/012075
14. Manda P., Sayedahmed S., Mohanty S.D. Automated ontology-based annotation of scientific li-
terature using deep learning. SBD’20. Proceedings of the International Workshop on Semantic
Big Data. USA : Portland, 2020. USA : New York, 2020. Art. 5. 6 p. DOI: https://doi.org/10.
1145/3391274.3393636
15. An ontology knowledge inspection methodology for quality assessment and continuous im-
provement / G.R. Roldán-Molina, D. Ruano-Ordás, V. Basto-Fernandes, J.R. Méndez. Data &
Knowledge Engineering. 2021. Vol. 133. ID: 101889. DOI: https://doi.org/10.1016/j.datak.
2021.101889
16. Vahidnia S., Abbasi A., Abbass H. A temporal ontology guided clustering methodology with a
case study on detection and tracking of artificial intelligence topics. SSRN Electronic Journal.
2022. 50 p. DOI: http://dx.doi.org/10.2139/ssrn.4200134
17. Gandon F. Distributed artificial intelligence and knowledge management: ontologies and
multi-agent systems for a corporate semantic web. 2002. 487 p. URL: http://www.theses.
fr/2002NICE5773
18. Bazaron S., Rukavichnikov A. Method specifications subject area discipline based on ontological
approach. Cybersecurity Issues. 2014. N 5(8). P. 52–58. URL: https://cyberrus.info/wp-content/
uploads/2015/02/vkb_08_10.pdf (in Russian).
Submitted 09.05.2025
https://doi.org/10.1109/KhPIWeek61412.2023.10313006
http://dx.doi.org/10.1088/1755-1315/1254/1/012075
http://dx.doi.org/10.1088/1755-1315/1254/1/012075
https://doi.org/10.1016/j.datak.2021.101889
https://doi.org/10.1016/j.datak.2021.101889
https://cyberrus.info/wp-content/uploads/2015/02/vkb_08_10.pdf
https://cyberrus.info/wp-content/uploads/2015/02/vkb_08_10.pdf
|
| id | nasplib_isofts_kiev_ua-123456789-211404 |
| institution | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| issn | 0572-2691 |
| language | Ukrainian |
| last_indexed | 2026-03-18T08:07:15Z |
| publishDate | 2025 |
| publisher | Інститут кібернетики ім. В.М. Глушкова НАН України |
| record_format | dspace |
| spelling | Gorda, О. Riabchun, Yu. 2026-01-01T19:46:21Z 2025 GAN-technologies in intelligent environments of the subject area / О. Gorda, Yu. Riabchun // Проблемы управления и информатики. — 2025. — № 3. — С. 85-95. — Бібліогр.: 8 назв. — англ. 0572-2691 https://nasplib.isofts.kiev.ua/handle/123456789/211404 004.8:004.032.26:004.93 10.34229/1028-0979-2025-3-8 The paper considers the problem of applying GAN-technology in intelligent domain environments (IDEs) for the need to structure large amounts of data and generate initial knowledge. У статті розглянуто застосування технології GAN в інтелектуальних середовищах предметної області (ІСПО) для структурування великих обсягів даних і генерації початкових знань. uk Інститут кібернетики ім. В.М. Глушкова НАН України Проблеми керування та інформатики Роботи та системи штучного інтелекту GAN-technologies in intelligent environments of the subject area GAN-технології в інтелектуальних середовищах предметної області Article published earlier |
| spellingShingle | GAN-technologies in intelligent environments of the subject area Gorda, О. Riabchun, Yu. Роботи та системи штучного інтелекту |
| title | GAN-technologies in intelligent environments of the subject area |
| title_alt | GAN-технології в інтелектуальних середовищах предметної області |
| title_full | GAN-technologies in intelligent environments of the subject area |
| title_fullStr | GAN-technologies in intelligent environments of the subject area |
| title_full_unstemmed | GAN-technologies in intelligent environments of the subject area |
| title_short | GAN-technologies in intelligent environments of the subject area |
| title_sort | gan-technologies in intelligent environments of the subject area |
| topic | Роботи та системи штучного інтелекту |
| topic_facet | Роботи та системи штучного інтелекту |
| url | https://nasplib.isofts.kiev.ua/handle/123456789/211404 |
| work_keys_str_mv | AT gordao gantechnologiesinintelligentenvironmentsofthesubjectarea AT riabchunyu gantechnologiesinintelligentenvironmentsofthesubjectarea AT gordao gantehnologíívíntelektualʹnihseredoviŝahpredmetnoíoblastí AT riabchunyu gantehnologíívíntelektualʹnihseredoviŝahpredmetnoíoblastí |