In this study, a deep learning framework for multiclass ECG classification based on a 2D-to-1D data transformation is presented. The proposed methodology integrates a deep autoencoder and a long short-term memory network to enable sequential modeling of compressed 2D signal representations. This strategy effectively addresses the challenge of data format compatibility, enabling the training of models on data initially unsuitable for such architectures. ECG signals were transformed into 2D representations using techniques such as recurrence plots, visibility matrices, Gram matrices, and Markov transition fields, followed by compression into low-dimensional vectors suitable for sequence processing. The proposed method was rigorously evaluated using the MIT-BIH arrhythmia database and demonstrated superior performance compared to contemporary state-of-the-art models. The method achieved high metric values, with MTFs yielding the best results: 99.8% accuracy, 99.8% precision, 99.8% recall, 99.8% specificity, 99.8% F1-score, 99.7% Matthews correlation coefficient, and 99.3% Cohen’s kappa, confirming its effectiveness for real-time ECG classification in clinical scenarios. Furthermore, the model’s versatility was validated through its successful application to image classification tasks, highlighting its adaptability to a wide range of data types beyond ECG signals.

AI-assisted programming has gained significant traction in recent years, primarily driven by advances in large language model (LLM) technologies. This growth has resulted in the development of tools such as ChatGPT and GitHub Copilot. However, the use of AI models for code generation presents a critical challenge: The generated code is prone to security vulnerabilities, raising concerns within the domain of secure software development. This study examines the current research landscape on the security implications of LLM-generated code from a software engineering perspective. To this end, we conducted a multivocal literature review (MLR) in accordance with the guidelines outlined by Kitchenham et al. and Garousi et al. The search encompassed five academic sources: IEEE Xplore, ACM, ScienceDirect, Springer Link, and Wiley Online Library; as well as two grey literature sources: Google Scholar and Google. We supplemented the selection by means of backward and forward snowballing techniques. Through Quasi-Gold’s automated search, we identified 2796 peer-reviewed studies, from which we selected 48 primary studies. The analysis revealed seven distinct categories of security vulnerabilities in LLM-generated code, six mitigation strategies and best practices, and four tools recommended for improving security in AI-assisted coding. While research on this issue remains in its early stages, initial findings highlight the need for caution when integrating AI-powered code generation into software development, as it may introduce vulnerabilities at a higher rate than human-written code. The vulnerabilities, mitigation strategies, and tools identified in this study can serve as valuable resources for developers seeking to use LLM-based programming assistants more securely and responsibly.

Requirements engineering (RE) is critical to software development, yet ambiguous stakeholder inputs and complex dataset management often cause costly defects. This study conducts a systematic literature review (SLR) to explore how artificial intelligence (AI) techniques and natural language processing (NLP) tools enhance requirements elicitation and analysis. Through analyzing 36 primary studies, we identify 34 AI techniques, including support vector machine (SVM), random forest (RF), and BERT, alongside 25 NLP tools such as NLTK, SpaCy, and BERT4RE. These technologies enhance automation, precision, and scalability in RE. They excel in classifying requirements, extracting key abstractions, and detecting ambiguities, with techniques like RoBERTa achieving up to 97.7% F1-scores and tools like the score-based ambiguity detector and resolver (SBADR) offering 99% recall for ambiguity detection. However, key limitations include data dependency, computational demands, and challenges in handling domain-specific terminology. This synthesis extends prior elicitation-focused research, highlighting AI and NLP’s transformative potential in RE while identifying gaps in domain adaptation and computational efficiency for future exploration.

In the rapidly evolving field of quantum computing technology, postquantum cryptography (PQC) plays a pivotal role in developing secure cryptographic systems that are resistant to both quantum and classical computers. This paper focuses on the integration and implementation of PQC algorithms such as Falcon and Dilithium for digital signatures, as well as Crystal-Kyber for key encapsulation mechanisms. Through a server–client model, the study has deployed a postquantum secure communication service achieving encryption times of less than 2 ms, signing performance up to 50 ms for small files and 11 073 ms for large files, and key exchange times as low as 16 ms. These metrics provide deep insights into the resilience and security of the proposed cryptographic solutions in real-world environments.

Currently, mobile devices must meet high performance requirements and provide extended battery life, which implies low power consumption. These parameters directly depend on the processor frequency: at higher frequencies, the processor can execute more instructions per unit of time but consumes more energy, and vice versa. This paper investigates modern process scheduling methods in the Linux operating system kernel with the aim of enhancing performance and reducing power consumption in mobile devices. The process scheduling method used in Linux is modified so that it specifically focuses on the selection of the CPU core queue for a process and on the selection of the processor core frequencies. The modification is based on solving a discrete bi-criteria optimization problem. The optimization criteria consider two interrelated characteristics: performance and power consumption. The core idea of the modification is to analyze the instructions of executing tasks to identify situations where increasing the processor frequency becomes inefficient due to frequent memory accesses. The problem is solved under certain constraints: heterogeneous architectures are not considered, and hardware multithreading is not taken into account, i.e., each logical core is treated as a physical one. Based on the proposed method, software was developed and tested using ten benchmarks from the Rodinia Benchmark Suite. This suite is used to evaluate the performance of computers on various tasks, such as general-purpose computing, image processing, and signal processing. The results of the study showed that the application of the proposed approach reduces the average power consumption by 13% in mobile devices and increases the performance by 4% compared to the existing Linux kernel scheduler.

Fault-proneness is a measure that indicates the possibility of programming errors occurring within a software system. On the other hand, change-proneness refers to the potential for modifications to be made to the software. Both of these measures are crucial indicators of software maintainability, as they influence internal software metrics such as size, inheritance, and coupling, particularly when numerous changes are made to the system. In the literature, research has predicted change- and fault-proneness using internal software metrics. That research is almost a decade old. However, given the continuous evolution of software systems, it is essential to revisit and update our understanding of these relationships. Therefore, we have conducted an empirical investigation to revisit the relationship among various internal software metrics and change-proneness, and fault-proneness, aiming to provide current and relevant insights. In our study, we identified 25 internal software metrics along with the measures of change-proneness and fault-proneness within the well-known open-source systems belonging to the Apache and Eclipse ecosystems. Next, we analyzed the relationships among these metrics using statistical correlation methods. Our results revealed that most of the metrics have little to no correlation with fault-proneness. However, metrics related to inheritance, coupling, and comments showed a moderate to high correlation with change-proneness. These findings will help developers to minimize the higher correlated software metrics to enhance maintainability in terms of change- and fault-proneness. Additionally, these insights can guide researchers in developing new approaches for predicting changes and faults by incorporating the metrics that have been shown to have stronger correlations.

This paper formulates an asymptotic method for investigating vector electromagnetic fields of waveguide modes in the adiabatic approximation. The asymptotic method in the zero approximation is reduced to the symbolic investigation of a homogeneous system of ordinary differential equations; higher approximations correspond to inhomogeneous systems of ordinary differential equations with a similar matrix of coefficients. Using the asymptotic method, symbolic solutions in the zero and first approximations are obtained.

How can one check, for a given nonsingular real number matrix the entries of which have only a finite number of decimal digits, whether this matrix will remain nonsingular after some decimal digits are arbitrarily added to some (explicitly specified in advance) of its entries? It turns out that this problem is algorithmically solvable. A computer implementation of the proposed algorithmic solution is discussed.

Systems of nonalgebraic equations containing entire functions of several complex variables are considered. The number of real zeros of such systems is studied using computer algebra methods. For this purpose, a computer implementation of Newton’s recurrences and formulas for the resultant of the functions under study in Maple is proposed. The relevance of this problem is due to the fact that in applied problems, e.g., in the equations of chemical kinetics, it is necessary to determine the number of stationary states of the system.