This paper addresses the challenge of providing operational access to current metadata in complex, ever-changing relational data warehouses. Traditional catalogs struggle to keep up with changes in schemas, code, and processes. The paper presents a methodological approach based on a dual-loop architecture with ReAct agents and retrieval-augmented generation. The first loop, managed by an Ingestion Agent, continuously updates the semantic layer by automatically analyzing changes. The second loop uses an Assistant Agent to give analysts, developers, and support engineers an intelligent interface. This interface combines semantic search over a vector database with direct execution of diagnostic queries through an extensible set of tools. The main goal is to create a self-updating metadata ecosystem that provides operational access to contextual information for different user groups. The approach’s practical effectiveness is demonstrated through end-to-end scenarios, such as creating complex queries based on business terms or diagnosing extract-transform-load processes.

Training in supervised machine learning is based on the availability of datasets; however, medical datasets must comply with stringent privacy regulations. Generative Adversarial Networks (GANs) are a relevant alternative to solve the limitation of small medical datasets due to their ability to generate additional data with desired features. A significant drawback of these models is that they may produce unrealistic, blurred, or insufficiently diverse images. This paper proposes a data augmentation technique using GANs to create synthetic Magnetic Resonance Imaging (MRI) of four stages of Alzheimer's Disease (AD): non-demented, very mild demented, mild demented, and moderate demented. We designed a GAN based on the Pix2Pix model, which learns the features of each AD stage. Generated images are evaluated by multistage Convolutional Neural Network (CNN) models, greyscale histograms of the distribution of pixel intensities, and brain mass measurements on binarized images. The results indicate that AD synthetic MRI effectively captures disease patterns, demonstrating the potential of GANs to improve training and diagnosis of neurodegenerative diseases.

This study addresses the problem of automatic attack detection targeting Linux-based machines and web applications through the analysis of system logs, with a particular focus on reducing the computational requirements of existing solutions. The aim of the research is to develop and evaluate the effectiveness of machine learning models capable of classifying system events as benign or malicious, while also identifying the type of attack under resource-constrained conditions. The Linux-APT-Dataset-2024 was employed as the primary source of data. To mitigate the challenge of high computational complexity, model optimization techniques such as parameter quantization, knowledge distillation, and architectural simplifications were applied. Experimental results demonstrate that the proposed approaches significantly reduce computational overhead and hardware requirements while maintaining high classification accuracy. The findings highlight the potential of optimized machine learning algorithms for the development of practical early threat detection systems in Linux environments with limited resources, which is particularly relevant for deployment in IoT devices and edge computing systems.

The rise of security concerns in cloud-shared infrastructures has introduced significant challenges for maintaining privacy in data processing. Although standard encryption methods provide robust protection for data at rest and during transmission, vulnerabilities arise when data must be decrypted for processing, exposing sensitive raw information to potential privacy risks. This issue is particularly pronounced in sectors governed by stringent regulatory requirements, such as healthcare, genomics, smart government, and finance, among many others, where protecting confidential data is critical. Homomorphic Encryption (HE) cryptosystems are solutions to address privacy concerns by providing encrypted data computations. HE allows a non-trustworthy third-party resource to process encrypted information without disclosure. However, the main challenge toward deploying lattice-based HE schemes in Convolutional Neural Network (CNN) models lies in overcoming the high computational costs associated with these cryptosystems. Efficient cryptographically compatible methods become imperative for designing a privacy-preserving CNN with HE (CNN-HE). This paper proposes a method to improve the performance of CNN-HE using the Residual Number System (RNS)-based Cheon-Kim-Kim Song (CKKS) HE scheme, which enables approximate arithmetic over encrypted real numbers. The CNN-HE with CKKS-RNS enables encrypted inputs to be decomposed into several parts and propagated homomorphically and independently in parallel across the model. The RNS representation enables parallel processing in our models, significantly reducing processing time. Experimental analysis on the MNIST optical character recognition benchmark dataset demonstrates that the proposed CNN-HE-RNS models reduce classification latency concerning state-of-the-art CNN-HE solutions without compromising security and accuracy.

This paper presents a study on distributed artificial neural networks implemented using wavelet transform-based modular architectures. The research compares the performance of monolithic, vertically partitioned, and horizontally partitioned artificial neural network configurations, with particular focus on computational efficiency and recognition accuracy. Experimental results demonstrate that horizontally partitioned artificial neural networks employing Haar wavelet transforms (2 × 2 kernel) achieve comparable recognition accuracy to monolithic networks (within 1

The rank of a number in a residue number system allows for improving the efficiency of non-modular operations and determining values out of range. The paper proposes an approach that allows to find the rank from multiples of the bases of rank numbers. Proof of the correctness of the method is given. It also finds the required accuracy for the transition to the approximate computation of the rank of a number. Simulations of the considered methods in the RTL and physical synthesis environment have shown a twofold reduction in the hardware area with comparable running time for the computed ranks. An approach for determining the sign of a number in a residue number system based on the rank of a number is also proposed.

Current security cloud practices can successfully protect stored data and data in transit, but they do not keep the same protection during data processing. The data value extraction requires decryption, creating critical exposure points. As a result, privacy-preserving techniques are emerging as a crucial consideration in cloud computing. The homomorphic processing of machine learning models in the cloud represents a central challenge. The activation function is fundamental in constructing a privacy-preserving Neural Network (NN) with Homomorphic Encryption (HE). Standard activation functions require operations not supported by HE, so it is necessary to find cryptographically compatible replacement functions to operate over encrypted data. Multiple approaches address the limitation of function compatibility with polynomial approximation. These functions should exhibit a trade-off between complexity and accuracy, limiting the efficiency of conventional approximation techniques. The current literature on polynomial approximation of NN activation functions still lacks a thorough review. In this paper, we comprehensively review the standard activation functions of modern NN models and current polynomial approximation approaches. We highlight fundamental features to consider in the activation function and the approximation technique to operate over encrypted data.

Accounts are an integral part of most modern information systems and provide their owners with the ability to authenticate within the system. This paper presents an analysis of existing methods for detecting simple account passwords in automated systems. Their advantages and disadvantages are listed. A method was developed to detect simple exploitable passwords that administrators can use to supplement other existing methods to increase the overall security of automated systems against threats from accounts potentially compromised by attackers. The method was based on the analysis of commands executed in automated or manual modes with the indication of credentials in plain text. Minimum password strength requirements are provided based on the security level. A special case was considered in which all passwords analyzed in this way were found explicitly in the system logs. We developed a unified definition of the classification of passwords into simple and strong, and also developed machine learning technology for their classification. The method offers a flexible adaptation to a specific system, taking into account the level of significance of the information being processed and the password policy adopted, expressed in the possibility of retraining the machine learning model. The experimental method using machine learning algorithms, namely the ensemble of decision trees, for classifying passwords into strong and potentially compromised by attackers based on flexible password strength criteria, showed high results. The performance of the method is also compared against other machine learning algorithms, specifically XGBoost, Random Forest, and Naive Bayes. The presented approach also solves the problem of detecting events related to the use and storage of credentials in plain text. We used the dataset of approximately 770,000 passwords, allowing the machine learning model to accurately classify 98% of the passwords by their significance levels.

A substitution box (S-box) is a fundamental component in cryptographic algorithms that enhance data security by providing a complex mapping between input and output values. S-box strengthens the encryption and decryption process by introducing nonlinearity and protecting the encrypted data against various differential and linear cryptanalytic attacks. The problem of generating an S-box with optimal properties is challenging and falls under the category of NP-Hard problems. This study proposes a hybrid approach combining the Particle Swarm optimization algorithm (PSO) and the Booster algorithm to construct a highly nonlinear S-box with low computational efforts. The PSO algorithm, assisted by the Transfer function and Random Key (RK), is utilized to navigate the large permutation search space to find an S-box with acceptable cryptographic properties. The Booster algorithm works based on random applications of local operators for shuffling the elements of the S-box with each other and transforming the elements' arrangement, resulting in a modified S-box with increased nonlinearity.

The Residue Number System is widely used in cryptography, digital signal processing, image processing systems and other areas where high-performance computation is required. One of the computationally expensive operations in the Residue Number System is the parity detection of a number. This paper presents a high-speed algorithm for parity detection of numbers in Residue Number System based on Akushsky core function. The proposed approach for parity detection reduces the average time by 20.39

The Residue Number System (RNS) is a number representation system used in applications that require high-speed arithmetic operations such as signal processing, digital filters, cryptography, and error correction codes. However, one of the problems with using RNS is determining the rank of a number represented in the system, which is a computationally difficult operation. This paper presents a method for detecting and correcting errors in arithmetic operations using the properties of the normalized rank of a number. The authors develop a lemma relating two functions of the rank of a number, which is used to efficiently calculate the normalized rank. The performance of error detection and correction algorithms largely depends on the efficiency of calculating the rank of a number in RNS. The paper also describes the development of methods for calculating the rank of a number using an approximate method, which shows the relationship between the rank of a number and the inversion of a prime number and is proved using mathematical equations. This article proposes a solution to this problem and is of great importance in improving the performance of RNS-based algorithms.

The automotive sector benefits from Digital Twins (DTs), software replicas of physical assets or processes. DTs enable engineers and data scientists to obtain deeper insights into the system and solve the most difficult problems faster and more affordably. Blockchain technology is a developing and exciting technology that has the potential to offer DTs monitoring capabilities, strengthening security and enhancing DTs’ transparency, dependability, and immutability. Intelligent behavior can be integrated into blockchain-based DTs to foresee important maintenance tasks and successfully manage machine functions. Our research involves creating a theoretical framework that leverages emerging technologies such as blockchain, artificial intelligence and DTs to facilitate resolution in the predictive maintenance of industry machines with minimised governing cost.

The residue number system is widely used in cryptography, digital signal processing, image processing systems, and other areas where high-performance computing is required. One of the main tools used in the residue number system is the Akushsky core function. However, its use is limited due to the existence of so-called critical cores. This study aims to develop Akushsky core functions that effectively eliminate the occurrence of critical cores, thereby enhancing their applicability in real-world scenarios. We introduce a fundamental approach to critical core detection that reduces the average time for critical core detection by 99.48% compared to the brute force algorithm. The results of our analysis indicate not only a substantial improvement in the speed of core detection but also an enhancement in the overall performance of systems utilizing the Akushsky core function. Our findings provide important insights into optimizing residue number systems and encourage further exploration into advanced computational techniques within this domain.

Determining the sign of a number is not as simple as addition and multiplication. When using the traditional notation of a number in binary form with two's complement code, it allows you to store the sign of the number and process it. However, when representing a number in modular form, or in any other forms, for example, in the form of a homomorphic cipher, where the operation of determining the sign cannot be performed explicitly. In such cases, it is necessary to resort to various computationally complex methods. In this work, we are conducting research on the possibility of using neural networks to calculate an approximate function of the sign of a number, this will reduce the computational costs of traditional approaches to determining the sign.

Fog computing (FC) is a distributed architecture in which computing resources and services are placed on edge devices closer to data sources. This enables more efficient data processing, shorter latency times, and better performance. Fog computing was shown to be a promising solution for addressing the new computing requirements. However, there are still many challenges to overcome to utilize this new computing paradigm, in particular, reliability and security. Following this need, a systematic literature review was conducted to create a list of requirements. As a result, the following four key requirements were formulated: (1) low latency and response times; (2) scalability and resource management; (3) fault tolerance and redundancy; and (4) privacy and security. Low delay and response can be achieved through edge caching, edge real-time analyses and decision making, and mobile edge computing. Scalability and resource management can be enabled by edge federation, virtualization and containerization, and edge resource discovery and orchestration. Fault tolerance and redundancy can be enabled by backup and recovery mechanisms, data replication strategies, and disaster recovery plans, with a residual number system (RNS) being a promising solution. Data security and data privacy are manifested in strong authentication and authorization mechanisms, access control and authorization management, with fully homomorphic encryption (FHE) and the secret sharing system (SSS) being of particular interest.

A smart city has a complex hierarchical communication system with various components. It must meet the requirements of fast connection, reliability, and security without data compromise. Internet of Things technology is widely used to provide connectivity and control solutions for smart sensors and other devices using heterogeneous networking technologies. In this paper, we propose a routing solution for Wireless Sensor Networks (WSN) and Mobile Ad hoc NETworks (MANET) with increasing speed, reliability, and sufficient security. Many routing protocols have been proposed for WSNs and MANETs. We combine the Secret Sharing Schemes (SSS) and Redundant Residual Number Systems (RRNS) to provide an efficient mechanism for a Distributed dynamic heterogeneous network Transmission (DT) with new security and reliability routing protocol (DT-RRNS). We analyze the concept of data transmission based on RRNS that divides data into smaller encoded shares and transmits them in parallel, protecting them from attacks on routes by adaptive multipath secured transmission and providing self-correcting properties that improve the reliability and fault tolerance of the entire system.

This article presents a study related to increasing the performance of distributed computing systems. The essence of fog computing lies in the use of so-called edge devices. These devices are low-power, so they are extremely sensitive to the computational complexity of the methods used. This article is aimed at improving the efficiency of calculations while maintaining an appropriate level of reliability by applying the methods of the Residue Number System (RNS). We are investigating methods for determining the sign of a number in the RNS based on the core function in order to develop a new, fast method. As a result, a fast method for determining the sign of a number based on the Akushsky core function, using approximate calculations, is obtained. Thus, in the course of this article, a study of methods for ensuring reliability in distributed computing is conducted. A fast method for determining the sign of a number in the RNS based on the core function using approximate calculations is also proposed. This result is interesting from the point of view of nebulous calculations, since it allows maintaining high reliability of a distributed system of edge devices with a slight increase in the computational complexity of non-modular operations.

With the development of next-generation information technologies, especially big data and digital twins, the topic of building smart cities is increasingly dominating discussions about social change and economic performance. The purpose of this article is to analyze methods for building digital twins of a smart city. The paper describes the concepts underlying digital twins. Examples of the implementations of methods for building digital twins are investigated. Advantages of data mining and neural network modeling over other methods in the context of the considered characteristics are revealed. Based on the comparative analysis, it is shown that all methods can be complementary, as they are aimed to optimize processes, as well as predict and analyze problems.

In this paper, a study is carried out related to improving the reliability and fault tolerance of Fog Computing systems. This work is a continuation of previous studies. In the past, we have developed a method of fast operation for determining the sign of a number in the Residue Number System based on the Akushsky Core Function. We managed to increase the efficiency of calculations by using the approximate rank of a number. However, this result is not final. In this paper, we consider in detail the methods and techniques of the Akushsky Core Function. During research, it was found that the so-called weights can be equal to random variables. Based on the data obtained, we have developed a method for determining the optimal weights for the Akushsky Core Function. The result obtained allows you to obtain a performance advantage due to the preliminary identification of optimal weights for each set of moduli.

The article explores the option of using information theory’s mathematical tools to model artificial neural networks. The two primary network architectures for image recognition, classification, and clustering are the feedforward network and convolutional networks. The study investigates the use of orthogonal transformations to enhance the effectiveness of neural networks and wavelet transforms in convolutional networks. The research proposes practical applications based on the theoretical findings.