This paper is devoted to the analysis of the specific focuses, directions, and peculiarities of the Strategy of Artificial Intelligence (AI) Development in Ukraine. The main paper's components are an analysis of the current state of the justification, development, and governmental approval of the National Strategy of AI in Ukraine; key elements and main priority areas of AI implementation according to IAIP-project “Strategy for AI Development in Ukraine”; proposals for AI development in short- and long-term perspectives and features of the AI implementation in Ukraine during the current wartime. Special attention is paid to such focuses in AI research and development as (a) the design of AI systems based on cognitive and conscience conceptions; (b) new solutions in intelligent robotic systems for ground, underwater and aerial applications; (c) AI perspectives in the marine industry; (d) prospective AI implementation in education; (e) linguistic competency of AI systems. Strategy, artificial intelligence, development, Ukraine, peculiarities, analysis, IAIP-project Austria, Germany, and others. According to IQ-Holon publication [5] the governments of fifty countries from different continents have created and approved AI strategies in different forms and styles as plans, ORCID: 0000-0001-7736-883X (Y. Kondratenko); 0000-0002-8446-5096 (G. Kondratenko); 0000-0002-0095-538X (A.Shevchenko); 00000002-2912-3149 (V. Slyusar); 0000-0002-6391-4382 (Y. Zhukov); 0000-0003-0772-7950 (M. Vakulenko).
Artificial intelligence (AI) plays a more and more important role in the different fields of human activity. Scientists and experts are expecting revolutionary results with
AI development and
implementation in medicine and healthcare, transportation, science, education, military and defense,
manufacturing, agriculture, space exploration, and different services [
The governments of developed countries understand the necessity of funding AI research for providing significant economic growth and for the leading position in the world’s GDP competition. Many countries created their own Strategy for AI development and determined the priority areas for AI implementation, taking into account the features of their own economic situation, national interests, the indicators and possibilities in science, the level of the education system, and others. Among the
2023 Copyright for this paper by its authors.
Modern AI products have increasing implementation for solving different complex tasks with access
for users based on fee or non-fee financial approaches. In particular, ChatGPT, GPT-4 and other AI
platforms are very popular and very important with their huge potentials and possibilities [
This paper aims to the analysis of the main focuses and features of the strategy for AI development in Ukraine. It is very important for consolidation and concentration of the research efforts for implementing AI in priority areas. The rest of this paper is organized as follows. Section 2 presents the developed “Strategy for AI development in Ukraine” with an analysis of its key components, Ukraine’s priorities in AI development, and specific features of AI implementation in the current wartime. In section 3, the authors discuss the approach to the design of AI systems based on conscience conception. New solutions in intelligent robotics for ground, underwater and aerial applications are considered in section 4. Sections 5 discuss the prospective AI implementations in the marine industry and section 6 – in education. Section 7 is devoted to the linguistic competency of AI systems. The paper ends with a conclusion in Section 8.
Let us analyze the current state of the justification, development, and governmental approval of the
“Strategy for artificial intelligence development in Ukraine” based on the IAIP’s project on AI Strategy
[
The AI field is developing and implementing very fast in Ukraine. There are more than 2,000 software development companies in Ukraine specializing in the AI industry. Ukraine has made a progressive step in the publishing open data direction, especially during the past few years. Concerning the Global Open Data Index, Ukraine places 31st position in the world.
The National Academy of Sciences, the Ministry of Digital Transformation, the Ministry of Education and Sciences, the Ministry of Strategic Industries, and many other governmental organizations in Ukraine are involved in the process of creating a National Strategy for AI Development and Implementation in Ukraine.
As a result, the Conception for AI Development and Implementation was created in Ukraine and on
2 December 2020 [
In 2020 also was started the process of creating a detailed Strategy for the development of AI in
Ukraine. The Institute of Artificial Intelligence Problems (IAIP) of the National Academy of Sciences
(NASU) and the Ministry of Education and Science of Ukraine (MESU) became the leading
organization in the IAIP-project “Creating Strategy for the Development of AI in Ukraine” [
The main steps for the development of the Ukrainian AI strategy were defined by the next tasks: • analysis and comparative review of the published national strategies of AI development in different countries from different continents; • formation of a generalized presentation of the analytical AI centers’ activities; • determination of promising directions for developing AI in Ukraine; • generalization of the basic terminology definitions, organizational principles, and main focuses of further research of Ukrainian scientists in the AI field; • identification of the priority domains for implementation of advanced AI in Ukraine; • formation of a list of necessary legislative, organizational, and investment measures for the implementation of the identified directions for the development of AI in Ukraine.
The IAIP-project was successfully executed but, unfortunately, the Russian aggression on Ukraine in February 2022 seriously influenced the global discussions and final approval of this AI Strategy as National AI Strategy at the governmental level.
Let us focus on the key components of the developed AI Strategy and the main priorities in the implementation of AI in Ukraine according to IAIP-project “Strategy for artificial intelligence development in Ukraine”.
The key content components in ten sections of the developed “Strategy for AI development in
Ukraine” [
This AI strategy was created based on the Ukrainian national characteristics and interests, the necessity to extend AI research, and the implementation of the recent AI tool developments in different fields of the Ukrainian economy. During the process of the AI Strategy creation, IAIP sent letters of inquiry to over 300 different organizations, in particular, to the majority of ministries of Ukraine, scientific institutions, state and private institutions of higher education, and commercial organizations to determine the need to implement and use AI in their work.
As a result, the next main priority areas for the implementation of AI in Ukraine were included in the Strategy for AI development in Ukraine” with detailed justifications and descriptions: • AI in the National Security and Military-Industrial Complex of Ukraine; • AI in Science and Education; • AI in Medicine and Healthcare; • AI in Manufacturing Industry and Power Sector; • AI in Telecom Industry; • AI in Transportation and Infrastructure; • AI in Agriculture; • AI in Ecology. 2.3.
The Strategy of AI Development in Ukraine (AIDU Strategy) [
To successfully implement the AIDU Strategy, the following immediate steps should be executed: Step 1. Approve and adopt the regulatory framework.
Step 2. Create the supervisory board to monitor and accomplish the tasks declared in the AIDU Strategy. Step 3. Determine the roadmap of the AIDU Strategy; Step 4. Prioritize the objectives of the AIDU Strategy; Step 5. Accomplish the most prioritized and fundamental tasks; Step 6. Provide mechanisms for quarterly and annual control over the implementation of the AIDU Strategy (reporting, optional examination, etc.).
Step 7. The final step is the reassessment of the AIDU Strategy, its analysis of compliance with the actualities of 2025, and, if necessary, its effective modification.
The AIDU Strategy should be supplemented with additional midterm (annual) deadlines, before which the aim and objectives of the relevant block must be completely accomplished. Each midterm period should be completed with an analytical report followed by an adjustment of the dynamic schedule. This component acts as a stimulus that will positively affect the intensity of the AIDU Strategy implementation.
To effectively implement the AIDU Strategy, it is necessary to take the following measures by 2025: • Create a regulatory framework that provides for the protection of economic and scientific data, as well as its storage in Ukraine. • Provide scientific and theoretical support for the execution of the AIDU Strategy. • Attract financial resources for the development of AI in Ukraine. • Provide support for fundamental and applied scientific AI research. • Increase the number of qualified AI employees and raise new technology awareness. • Improve the digital literacy of the Ukrainian people. • Build a national database system.
The main mechanism for the Strategy of Artificial Intelligence Development in Ukraine implementation is the annual action plans developed by the Committee on the Development and Implementation of Artificial Intelligence and approved by the Cabinet of Ministers of Ukraine. 2.4.
The war in Ukraine has become the first high-tech war in human history, in which both sides of the conflict began utilizing the capabilities of so-called computational artificial intelligence (AI).
The implementation of AI in Ukraine during wartime is characterized by its widespread use across various domains. First and foremost, AI plays a crucial role in tactical combat actions and military operations, particularly in enhancing the effectiveness of mass deployment of unmanned aerial vehicles (UAVs) for surveillance and reconnaissance tasks, and the evaluation of artillery fire effectiveness. According to experts, the deployment of UAVs accounts for over 70% of targets destroyed during combat operations. Developers quickly transitioned from using classical convolutional neural networks to segmenting objects based on various U-Net and PSPNet structures [11, 12]. In relatively simple object classification tasks, transfer learning and zero-shot learning methods based on neural networks previously trained on the ImageNet dataset [13, 14] performed well.
Later, in a short period, the process of adapting known neural network technology, Object Detection (OD) [15], to field datasets gained popularity. This was used not only for automatic target detection but also for simultaneous classification under varying seasonal conditions. The most ambitious projects aim to implement object identification and target class recognition, for example, combat vehicles, tanks, logistical transport, etc. This can be extended to classify a specific type of object similarly. To solve all these mass deployment tasks, various versions of YOLO family neural networks were widely used [16]. Primarily, their operation takes place not onboard the UAVs but in the command post equipment. Importantly, object detection in images is combined with video tracking algorithms for real-time incoming video streams from onboard or stationary cameras of different spectral ranges in different domains. For example, Fig. 1 shows a fragment of video tracking based on YOLO5 Small of a highspeed motor boat. The neural network effectively tracks the boat and allows counting the number of people on board. Similar results of automatic detection and tracking of a drone by a neural network are shown in Fig. 2. This OD technology is also used for detecting unexploded ordnance on the seabed using underwater drones and assessing housing and infrastructure damage.
An additional direction to enhance the capabilities of military information support was the application of intelligent chatbots in Telegram channels or based on separate mobile applications. These allow for alerts about the appearance of enemy machinery, means of air attacks, and so on. The chatbot boom has also covered areas such as psychological support for service members and legal assistance.
Natural language processing (NLP) is generally a promising direction in the field of AI, especially considering the capabilities of the language model GPT-4 and its less powerful counterparts. With Ukraine receiving Western weapons samples, effective combat operation requires translations of NATO standards and corresponding technical documentation from various European languages into Ukrainian. In this regard, smartphone translators with built-in audio and optical text recognition feature from Google, as well as translation functions implemented in ChatGPT, have become handy. Furthermore, relying on local GPT-4 analogs such as LLaMA [17], Alpaca [18], etc., automatic analysis of combat reports from units can be provided. This allows for the prompt provision of information about the current battlefield situation to commanders upon their requests, facilitating rapid response to critical threats and decision-making support.
Apart from the application of large language models (LLM), NLP algorithms, and neural networks for video tracking and image processing, an important direction is neural network processing of time series. This allows for predicting meteorological data for high-precision artillery firing, expenditure and needs in various resources, the evolution of satellite navigation correction adjustments over time, etc.
Implementing AI in war conditions has its challenges. One of these is the necessity to ensure data security and protection against cyber-attacks. Given the heightened risk of cyber threats, artificial intelligence should be viewed as a potent player in cybersecurity efforts. It is employed in algorithms designed to detect and neutralize threats, as well as to protect critical infrastructure.
The role of AI also extends to the coordination of humanitarian aid. In the logistics sector, AI ensures the efficient distribution and optimal delivery of assistance to those who need it most. In the realm of information warfare, artificial intelligence plays a significant role in detecting and countering disinformation campaigns. It enables the analysis of large volumes of data to discern patterns and trends in disinformation. Machine learning methods have become the de facto standard approach when performing social media and mass media publication analysis for Open Source Intelligence interests.
The predictive analytics capabilities of AI are also employed to forecast enemy movements and the tactics of deploying weapons and military equipment. This assists military planners in strategizing their actions and responding to potential threats.
In addition, AI is utilized in the management and servicing of critical infrastructure during warfare. It aids in monitoring and predicting potential infrastructure failures, and coordinates repair and maintenance efforts. This information is based on the latest data available as of June 2023 and is continually supplemented with new evidence of the growing role of AI in all spheres of society amidst military operations.
The novel authors’ proposal in AI Strategy for Ukraine deals with the development, design and
implementation of disruptive AI systems based on conscience conceptions. As its human cognate,
artificial consciousness (AC) is a necessary attribute of an artificial personality with AI. Artificial
consciousness manifests itself as an emergent global self-organized information phenomenon that
evaluates and controls core processes of the system, exchanges data between system components to
coordinate their behavior, provides for the social and personal perception of the environment, and
conditions internal integration and external separation of the system [
This approach suggests the synergetic treatment of AC that may be represented by a strange attractor and correspondingly simulated. Similar ideas were put forward earlier. For example, W. Calvin [19] introduces the concept of a “global workspace” that results from neuron interactions and integrates information from different brain regions. A. Bailey [20] examines James’ theory of the stream of consciousness which includes subjective feelings and emotions. The functional model of a newgeneration computer system with AI is shown in Fig. 3 with modules [21] of artificial consciousness and artificial conscience. The first attempt to create AI based on the conscience conception was made by the firm Anthropic [22]. Anthropic claims that their Claude chatbot adheres to many rules, including the principles enshrined in the Universal Declaration of Human Rights. The firm Anthropic claims that its chatbots have a “conscience”.
Creating cognitive computers and robot knowledge analysis based on cognitive computing and modular neural networks [23, 24] is also a very promising research direction in AI design.
Ukrainian scientists and policymakers pay special attention to the development and implementation of robotic systems. In intelligent robotics, research is focused on the evolution of autonomous ground vehicles (UGVs), autonomous surface (USVs), underwater vehicles (AUVs), unmanned aerial vehicles (UAVs), and the integration of these systems. The development of all these directions is propelled by the implementation of advanced AI capabilities. Machine learning is utilized for navigation, obstacle avoidance, and decision-making. There is an emerging trend towards employing multi-robot systems featuring dynamic self-organization of swarms.
In underwater robotics, equipping AUVs with advanced sensors and AI algorithms assists in pipeline inspection/protection and research of marine biology, underwater environments, and landscapes. Current developments aim to improve the autonomy of AUVs, allowing them to operate for extended periods and at considerable distances in challenging underwater conditions.
Aerial robotic systems, especially drones or UAVs, are used for a wide range of applications, from delivery services to combat operations. Future research in this field focuses on swarm robotics, where a group of drones collaboratively performs complex tasks. Alongside this, there is a need to develop AI algorithms that facilitate drone navigation in complex urban conditions, particularly in the absence of satellite navigation signals. An important direction for future research is the development of integrated robotic systems capable of operating in terrestrial, underwater, and aerial environments. This could potentially involve the creation of amphibious robots or systems where terrestrial, underwater, and aerial robots work in harmony.
As robotic systems gain greater autonomy, accompanying ethical and legal issues must be addressed. Specifically, this concerns responsibility for the consequences in case of accidents, privacy issues related to surveillance drones, and ethical implications of autonomous weapon systems.
Let us outline the future trajectories for the development of intelligent robotic systems (IRS).
Direction 4.1. Powerful LLMs such as GPT-3.5 and GPT-4 developed by OpenAI [25] use machine learning to generate human-like text and have found diverse applications in IRS. Implementation of such LLMs can enable robots to understand commands given in natural language, generate human-like responses, and even engage in conversations, making them more useful and easier to use.
Direction 4.2. By learning to understand a set of rules or guidelines laid out in natural language, a robot gains the ability to use these rules to make decisions in real-world situations. In this way, a prototype of an artificial conscience could be implemented, whose mechanism would allow for the avoidance of issues in communication with humans and other robots, making decisions that respect human rights and universal values.
Direction 4.3. Additional possibilities will be provided by the development of local GPT analogs such as LLaMA, Alpaca, etc. These models can be embedded in the onboard equipment of a robot, increasing its independence from external communication networks. This direction is closely related to the development of neural networks designed for converting audio streams into text and text-to-speech.
Although LLMs offer many potential advantages for IRS, they also pose certain challenges. For example, GPT models may sometimes generate incorrect or nonsensical responses, their training requires large volumes of data and computational resources, and their use in autonomous systems raises significant concerns regarding safety, ethics, and legal regulation. These issues will require thorough investigation to ensure the beneficial use of such AI technologies without posing excessive risks.
Direction 4.4. Importantly, (a) future solutions for IRS will allow robots to learn from past
experiences and adapt to new situations without human intervention; (b) improvements in sensor
technology can enable robots to better understand and interact with their environment [
Direction 4.5. A significant impact on the future of the IRS will be made by Augmented Reality (AR) technology [26]. It will provide a more intuitive way for humans to interact with robots, visualizing a robot's intentions, planned actions, or internal state. AR will enhance robots' autonomy by helping them better understand and navigate their environment. It is important to note that the realization of the outlined future achievements in the IRS will depend on various factors, including scientific breakthroughs, actual advances in the development of respective technologies, societal perception, as well as legal and ethical approaches.
Implementing AI technologies has a good perspective on the marine industry, which is very important for Ukraine as a marine country. For example, (a) the AI multi-software complexes are successfully used in design processes in shipbuilding and ship-repairing, (b) intelligent polymetric sensor systems are highly efficient as information components of the integrated ships' control systems [27], etc. Particular attention should be paid to increasing the efficiency of ship safety monitoring systems based on AI [28], which can (a) provide the seafarer with reliable and visualized factual information concerning ship loading and wind-wave impact to increase the soundness of his decisions for safe and efficient routing in heavy sea conditions, in particular, to provide navigational safety at stormy seas; (b) control of the autonomous (crewless) marine vehicle for fulfilling its mission with correction of the planned path, speed, and course in the current sea environment.
Early and current research on ship safety monitoring systems focused primarily on using sensors
and other hardware devices to detect hazards such as collisions, fires, capsizes, and leaks. Such systems
effectively recognize potential dangers and warn crew members and other stakeholders early. However,
these systems' existing hardware/software limitations could influence their effectiveness. For example,
these systems were often limited by the hardware devices' processing power, which can result in delays
in data analysis and decision-making. And in any case, they still are DSS - Decision Support Systems
[
Direction 5.1. The AI-based ship safety monitoring systems involve AI algorithms and advanced digital sensors to detect and analyze potential hazards in real-time operations and optimize ship routing from human, technical, commercial, and ecological safety points of view. This AI approach would involve the integration of multiple digital and smart sensors, including liquid cargoes, green fuels and technological liquids state parameters sensors, dynamic parameters of control units, weather conditions monitoring, main engine and auxiliary systems parameters monitoring, and other devices, to provide real-time data on conditions onboard the ship into cloud databases.
The proposed AI approach's practical implementation would involve installing advanced sensors and AI algorithms onboard the ship. This approach would require the development of new intelligent hardware devices and software applications that can support real-time data analysis and predictive capabilities.
Direction 5.2. Digital Twins (DT) would be widely used in ship design, manufacturing, operation, maintenance, modernization, repair, and, finally, their utilization. The output of DT returns to the ship as actions, recommendations, and even control. The outcomes of DT are used to: increase safety and reduction of operational costs; design new green and digitalized ships, equipment, etc.; training of operators and predictive maintenance; oversight and compliance monitoring, emergency response, etc.
While there are challenges to implementing this AI approach, the potential benefits make it a worthwhile investment for the shipping industry. Further research and development in this area are needed to realize the potential of AI-based ship safety monitoring systems fully.
LLMs ChatGPT and GPT-4 have significant potential for their use in the educational field. Especially the capabilities of GPT-4 increased after the introduction of access to the paid version of this language model to Internet resources and the provision of the possibility of using about 800 embedded plugins, the list of which is constantly expanding. Based on the gained practical experience of working with ChatGPT and GPT-4, it is possible to formulate a set of proposals or directions regarding potential areas of application of AI platforms built on LLMs in teaching, learning, and research processes.
Direction 6.1. GPT-4 can act as a personal tutor that provides information from different areas of knowledge and can explain concepts, helping students to better understand the learning material. AIbased platforms can create a more personalized and flexible learning environment for distance learning students [58, 59] and individuals with special needs.
Direction 6.2. GPT-4 can be used 24/7 to dialogue with students, form answers to their questions, and even host seminars and discussions, allowing students to learn on their own schedule and at their own pace. AI can play the role of a personal assistant that helps navigate learning and provides advice on choosing a course, career path, or even personal development. AI should be seen as an effective means of providing emotional support to students to help them cope with stress and maintain mental health.
Direction 6.3. AI can help educators assess assignments, provide feedback to students, and identify areas where students are struggling and need additional support. AI also can provide teachers with resources for professional development. LLMs are capable of working with different languages, allowing for the creation of multilingual learning resources and providing global access to education. AI can be a valuable tool for native and foreign language learning, providing instant feedback on grammar, pronunciation, and vocabulary. A promising trend is the integration of LLMs with generative transformers capable of synthesizing two-dimensional and three-dimensional images, and videos. This opens a wide field for creativity and improvement of teaching and learning processes.
Direction 6.4. AI can effectively manage the resources of educational institutions, for example, the optimal distribution of the teaching load among teachers, the preparation of lesson schedules, the management of library resources, or the coordination of services for students.
Direction 6.5. AI can facilitate collaborative learning by coordinating group projects, creating an environment for discussion, and providing feedback on group dynamics. LLMs are also a valuable tool to assist researchers by providing quick access to information and generating ideas for further research.
Although the proposed directions have great potential, it is also important to constantly consider the ethical implications and potential risks associated with the use of AI in education.
The policymakers in different countries pay attention to the implementation of the ChatGPT in education processes. For example, the Chancellor of the nation’s largest school system, New York City Public Schools, David C. Banks said on 18 May 2023 [29] that ChatGPT caught NYC schools off guard and now, they are determined to embrace its potential and in New York public schools, students will be taught how to use AI.
Direction 6.6. No doubt, the efficiency of training students in the AI field at the university level may be significantly increased in the framework of specialized integrated education environments [30] such as multi-university (academic) consortia and academic-industry consortia.
Linguistic competency is a recognized sign of human intelligence. It results from linguistic intelligence. The ability to express thoughts, ideas, and suggestions using human language, which constitutes the linguistic competency of an artificial personality, is an important subtask in developing AI. In turn, “accurate report”, which is a standard behavioral index indicating consciousness in humans, is best realized through human language [31]. In a similar way, these ideas are applicable to AI, which imitates human intelligence and thinking. In this sense, the developed linguistic competency of an artificial personality able to report accurately on what is going on may be regarded as a criterion indicating the rise of artificial consciousness.
Human understanding of the text or a message is based on the meanings of the used words, which are presented in explanatory dictionaries. The meaning of each word can be decomposed into elementary senses and these can be deduced from the word definition or explanation available in the dictionary. This process can be described mathematically and correspondingly formalized to automatically build semantic fields [32], resulting in the technical possibility to develop a deep intelligent instrument able to assess and compare texts and disambiguate word senses. The linguistic module of artificial personality can acquire human-like linguistic competency in this way. Chomsky argues [33a] that humans have an innate ability to acquire language, which is hard-wired into our brains. This idea has been influential in the development of NLP algorithms, which seek to replicate human language acquisition processes in machines and AI technologies.
It is important that the AI mechanisms modeling human thought and language preserve the information contained in the processed texts. If the original language uses a non-Latin alphabet, some natural language processes require its Romanization. To be able to restore the initial text, the Latinization process should be based on scholarly (strict) transliteration, which provides simplecorrespondent (one-to-one) or isomorphic correspondence between initial and Romanized graphemes. The Latinization rules using a mediator language inevitably refer to corresponding sounds in that language. That is, they are based on practical transcription rather than transliteration and, therefore, fail to preserve contained semantics. For example, the use of the English-oriented Romanization system for Ukrainian results in word form distortion and the appearance of false identities: Гальченко – Галченко (Halchenko).
Research direction, which concentrated on increasing the linguistic competency of AI systems for a correct understanding of the contents in communications between humans and intelligent robots and between different kinds of robots in multi-robotic systems, is perspective and important for future AI development and implementation.
The main peculiarities of the developed AIDU Strategy, priorities in AI implementation, and prospective research directions in the AI field are focused on and discussed in detail. The result of the “Strategy for Artificial Intelligence Development in Ukraine” implementation should be dealt with the creation of breakthrough technologies in the field of computer science and artificial intelligence as well as the creation of conscious AI-powered computers that make decisions considering ethical, moral, and legal norms. One promising way to realize the AIDU Strategy is the study of artificial consciousness based on a synergetic approach. At the next step, future research must be dealt with software and hardware development, testing and implementation of proposed new-generation intelligent systems with AI based on the conscience conception. Another important direction is the development of linguistic technologies, particularly those providing semantic text analysis that manifest the emergence of linguistic competency of an artificial personality. Besides, the authors analyzed and underlined the most important fields for AI implementation in Ukraine, as well as, developed, formalized and justified priority practical-research directions for future successful AI results and achievements, in particular, (4.1) – (4.5) in intelligent robotics, (5.1) – (5.2) in the marine industry, (6.1) – (6.6) in the education sphere. Scientific efforts must be concentrated on intensive AI research in the abovementioned directions to increase the role of Ukraine in the world as a high-technological country, strong marine country and country with high-caliber standards in education.
[11] V. Slyusar, et al., Construction of an advanced method for recognizing monitored objects by a convolutional neural network using a discrete wavelet transform, Eastern-European Journal of Enterprise Technologies, 4 112 (2021) 65-77. doi: 10.15587/1729-4061.2021.238601. [12] V. Slyusar, et al., Improvement of the object recognition model on aerophotos using deep conventional neural network, Eastern-European Journal of Enterprise Technologies 5 113 (2021) 6–21. doi: 10.15587/1729-4061.2021.243094. [13] C.H. Lampert, et al., Attribute-based classification for zero-shot visual object categorization, Pattern Analysis and Machine Intelligence, IEEE Transactions on, 36 3 (2014) 453–465. doi: 10.1109/TPAMI.2013.140. [14] Jia Deng, et al., ImageNet: A Large-Scale Hierarchical Image Database, in: 2009 IEEE Conf. on computer vision and pattern recognition, 2009, pp. 248–255. doi: 10.1109/CVPR.2009.5206848. [15] A. Bochkovskiy, et al., YOLOv4: Optimal Speed and Accuracy of Object Detection, arXiv.org, 2020, URL: https://arxiv.org/pdf/2004.10934.pdf. [16] F. Peng, et al., Traffic flow statistics algorithm based on YOLOv3, in: 2021 International Conference on Communications, Information System and Computer Engineering, CISCE’2021, 2021. doi: 10.1109/cisce52179.2021.9445932. [17] Hugo Touvron, et al., LLaMA: Open and Efficient Foundation Language Models, arXiv:2302.13971v1, 2023. doi: 10.48550/arXiv.2302.13971. [18] Zhengxuan Wu, et al., Interpretability at Scale: Identifying Causal Mechanisms in Alpaca, arXiv.org, 2023, URL: https://arxiv.org/pdf/2305.08809v1.pdf. [19] W.H. Calvin, The Cerebral Code: A Thinking Thought in the Mosaics of the Mind, The MIT Press, 1998. [20] A. Bailey, The Strange Attraction of Sciousness: William James on Consciousness, in: J. Scott Jordan (Ed.), Modeling Consciousness across the Disciplines, University Press of America, Lanham, MD, 1999, pp. 43–63. [21] A.I. Shevchenko, M.S. Klymenko, Developing a Model of Artificial Conscience, in: 15th IEEE International Scientific and Technical Conference on Computer Sciences and Information Technologies, CSIT’2020, vol. 1, 23-26 Sept. 2020, Lviv-Zbarazh, 2020, pp. 51–54. [22] Artificial intelligence now has a "conscience", 2023, URL: https://portaltele.com.ua/news/technology/u-shtuchnogo-intelektu-teper-ye-sovist.html [23] Y. Wang, et al., The Odyssey to Next-generation Computers: Cognitive Computers (kC) Inspired by the Brain and Powered by Intelligent Mathematics, Frontiers in Computer Science (2023). doi:10.3389/fcomp.2023.1152592 [24] V. Vychuzhanin, et al., Cognitive-impulse model for assessing complex technical systems survivability, CEUR Workshop Proceedings, Vol. 2711, 2020, pp. 571-585. URL: https://ceurws.org/Vol-2711/paper44.pdf [25] GPT-4 Technical Report by OpenAI, 27 March 2023. URL: https://arxiv.org/pdf/2303.08774v3.pdf. [26] Misiuk Tetyana, et al., Computer Vision Mobile System for Education Using Augmented Reality
Technology, Journal of Mobile Multimedia, 17 4 (2021) 555-576. [27] Y.D. Zhukov, et al., Intelligent Polymetric Systems Industrial Applications, CEUR Workshop
Proceedings, Vol. 2762, 2020, pp. 122 – 137. URL: https://ceur-ws.org/Vol-2762/paper8.pdf [28] Y. Kondratenko, et al., Ship Navigation in Narrowness Passes and Channels in Uncertain Conditions: Intelligent Decision Support. in: P. Shi, et al. (Eds.), Complex Systems: Spanning Control and Computational Cybernetics: Foundations, volume 414 of Studies in Systems, Decision and Control, Springer, Cham, 2022. Doi:10.1007/978-3-030-99776-2_24 [29] ChatGPT caught NYC schools off guard. Now, we’re determined to embrace its potential, 2023.
URL: https://ny.chalkbeat.org/2023/5/18/23727942/chatgpt-nyc-schools-david-banks [30] G. Kondratenko, et al., Fuzzy Decision-Making System for Model-Oriented Academia/Industry Cooperation: University Preferences, in: C. Berger-Vachon, et al. (Eds.) Complex Systems: Solutions and Challenges in Economics, Management and Engineering, volume 125 of Studies in Systems, Decision and Control, Springer, Berlin, Heidelberg, 2018, pp. 109-124. doi: 10.1007/978-3-319-69989-9_7 [31] A.K. Seth, et al., Criteria for consciousness in humans and other mammals, Consciousness and
Cognition 14 (2004) 119-139. [32] M. Vakulenko, From Semantic Metrics to Semantic Fields, in: Proceedings of the 2021 IEEE 16th International Conference on Computer Science and Information Technologies, CSIT’2021, 22–25 September 2021, Lviv, Ukraine, 2021, pp. 44–47. doi: 10.1109/CSIT52700.2021.9648675. [33] N. Chomsky, Language and Mind, Cambridge University Press, Cambridge, 2006.