Search

Deep Convolutional Neural Networks (CNNs) have exhibited remarkable eficacy in data analysis across various domains of physical research. Within the realm of particle and high-energy physics, an imperative challenge revolves around the scrutiny of particle tracks acquired through track detectors, exemplified by liquefied argon time projection chambers (LArTPCs). These cutting-edge neural networks have showcased unparalleled performance in tackling this intricate task, demonstrating their profound potential in enhancing particle track analysis. Since the scalability requirement is imposed on neutrino detectors, it is also imposed on CNN model implementations. The scalability for LArTPC data processing is achieved due to the sparsity of the data, which have the form of thin trajectories. To process sparse data, subspecies sparse convolutional networks (SSCNs) and sparse tensor networks (STNs) have been proposed. In this paper, we present the results of semantic segmentation of the LArTPC-simulated PILArNET data using STNs and test various modifications of the classical U-Net architecture: Attention U-ResNet, U-ResNet3+, U-ResNet with an additional deep supervision block, as well as loss functions: focal loss, balanced cross-entropy to improve the accuracy of Michel electron identification. The best results were obtained for balanced cross-entropy. Further improvement is possible by combining several methods. [ABSTRACT FROM AUTHOR]

Software-Defined Networking (SDN) has emerged as a transformative paradigm for network management, offering centralized control and programmability. However, with the proliferation of Distributed Denial of Service (DDoS) attacks that pose significant threats to network infrastructures, effective mitigation strategies are needed. The subject matter of this study is to explore the importance of datasets in the mitigation of DDoS attacks in SDN environments. The paper discusses the significance of datasets for training machine learning models, evaluating detection mechanisms, and enhancing the resilience of SDN-based defense systems. Goal of the paper is to assist researchers in effectively selecting and usage of datasets for DDoS mitigation in SDN, thereby maximizing benefits and overcoming challenges involved in dataset selection. This paper outlines the challenges associated with dataset collection, labeling, and management, along with potential solutions to address these challenges. Effective detection and mitigation of DDoS attacks in SDN require robust datasets that capture the diverse and evolving nature of attack scenarios. Characterization of tasks for each section is as follows: Importance of datasets in DDoS attack mitigation in SDN, challenges in dataset utilization in DDoS mitigation in SDN, Guidelines for dataset selection, comparison of datasets used and their results and different dataset usage according to the need. Methodology involves collecting results in tabular form based on prior research to analyze the characteristics of existing datasets, techniques for dataset augmentation and enhancement, and evaluating the effectiveness of different datasets in detecting and mitigating DDoS attacks through comprehensive experimentation. Results of our findings indicate that effective detection and mitigation of DDoS attacks in SDN require robust datasets that capture the diverse and evolving nature of attack scenarios. Our findings provide valuable insights into the importance of datasets in enhancing the resilience of SDN infrastructures against DDoS attacks. In conclusion, our findings provide valuable insights into the importance of datasets in enhancing the resilience of SDN infrastructures against DDoS attacks and highlight the need for further research in this critical area. Thorough guidelines for dataset selection and impacts of different datasets used in recent studies, provide research challenges and future directions in this area.

Abstract. The increasing number of information security incidents in higher education underscores the urgent need for robust cybersecurity measures. This paper proposes a comprehensive framework designed to analyze the illegal use of internet resources in university networks in Kazakhstan. The subject of this article is the detection and mitigation of cybercriminal activities using university networks in Kazakhstan. The goal is to develop a comprehensive framework that integrates multiple educational organizations to enhance collaborative security efforts by monitoring network activity and categorizing texts using machine learning techniques. The tasks to be solved are: to formalize the procedure of integrating multiple educational organizations into a collaborative cybersecurity framework; developing a log analysis tool tailored for monitoring network activities within university networks; creating a novel dictionary of extremist terms in the Kazakh language for text categorization; to implement advanced machine learning models for network traffic classification. The methods used are: log analysis tools for real-time monitoring and anomaly detection in network activities, Natural language processing (NLP) techniques to develop a specialized dictionary of extremist terms in Kazakh, Machine learning models to classify network traffic and detect potential cyber threats, and collaborative architecture design to integrate network security efforts across multiple institutions. The following results were obtained: a comprehensive log analysis tool was developed and implemented, providing real-time monitoring of network activities in university networks; a dictionary of extremist terms in Kazakh was created, facilitating the categorization and analysis of texts related to potential security threats; advanced machine learning models were successfully applied to classify network traffic, enhancing the detection and mitigation of cyber threats; and an experimental architecture integrating multiple educational organizations was established, fostering collaborative efforts in cybersecurity. Conclusions. The scientific novelty of the results obtained is as follows: 1) a robust framework for collaborative cybersecurity in educational institutions was developed, leveraging log analysis and machine learning techniques; 2) the creation of a specialized dictionary of extremist terms in Kazakh significantly improved the accuracy of text categorization related to cybersecurity; 3) the application of advanced machine learning models to network traffic classification provided a methodological approach to effectively managing and securing network infrastructure effectively; 4) the experimental architecture demonstrated the potential for enhanced security through collaboration among educational organizations, offering strategic recommendations for improving information security in academic environments. The outcomes of this research contribute to the broader cybersecurity field by providing a structured approach to detecting and mitigating cyber threats in educational contexts. The proposed framework has potential applications extending to global security frameworks, aiming to foster a safer internet usage environment and reduce the risks associated with cyber threats and unauthorized data access.

Artificial intelligence opens up great prospects in many areas of human activity, primarily in medicine. One of the priority directions of using artificial intelligence in this field is the segmentation of medical images for the purpose of automatic diagnosis of common diseases. The application of neural network approaches to image analysis of medical images is becoming an increasingly promising direction in the field of medical diagnostics. In particular, this paper investigates the possibility of using an ensemble of neural networks for diagnosing osteoporosis. To achieve this goal, a study was conducted on the possibility of using machine learning methods to segment and determine the shape and size of certain vertebrae: Th8, Th9, Th10, Th11 of a human vertebra on X-ray images obtained in real conditions. Each network is configured and tested on different sets of medical images. Then, the two best networks were selected according to the accuracy and efficiency of the segmentation. One of the main results of the study was the selection of the two best neural networks that provide the most accurate segmentation of vertebrae. Next, the ensemble method was applied, based on the averaging of the predictions of the selected networks. This approach made it possible to improve the overall accuracy of determining the diagnostic parameters of the spine. The obtained results emphasize the effectiveness of using an ensemble of neural networks in the context of medical segmentation. Ensembles provide more stable and accurate predictions by reducing the impact of random errors of individual networks. Ensemble predictions of these networks lead to a statistically significant improvement in results compared to individual approaches. This is an important step in the direction of creating reliable systems of automated diagnostics capable of helping doctors in conducting more accurate and operative analyses.

Generating paraphrases is a fundamental problem in natural language processing. In light of the significant success of transfer learning technology, the "pre-training fine-tuning" approach has become the standard. However, popular general-purpose pre-training methods typically require large datasets and computational resources, and available pre-trained models are limited by fixed architecture and size. We propose a simple and effective approach for pre-training specifically for paraphrase generation, which significantly improves model quality and matches the quality level of general-purpose models. Both existing public data and new data generated by large language models were used. The impact of this procedure on neural networks of different architectures was investigated, and it was shown to work for all of them. [ABSTRACT FROM AUTHOR]

Augmented reality (AR) and machine learning (ML) are rapidly growing technologies with immense potential for transforming education. Web-based augmented reality (WebAR) provides a promising approach to delivering immersive learning experiences on mobile devices. Integrating machine learning models into WebAR applications can enable advanced interactive effects by responding to user actions, thus enhancing the educational content. However, there is a lack of effective methodologies to teach students WebAR development with integrated machine learning. This paper proposes a methodology with three main steps: (1) Integrating standard TensorFlow.js models like handpose into WebAR scenes for gestures and interactions; (2) Developing custom image classification models with Teachable Machine and exporting to TensorFlow.js; (3) Modifying WebAR applications to load and use exported custom models, displaying model outputs as augmented reality content. The proposed methodology is designed to incrementally introduce machine learning integration, build an understanding of model training and usage, and spark ideas for using machine learning to augment educational content. The methodology provides a starting point for further research into pedagogical frameworks, assessments, and empirical studies on teaching WebAR development with embedded intelligence.

This paper aims to assess the development pattern of resource-rich countries like Norway, Kuwait, Saudi Arabia, and the Republic of South Africa by analyzing their stock markets. As resource-rich countries have profound implications for the world's sustainable development, it is essential to investigate whether they transform natural resource wealth into sustained growth by reorientation from related to raw material industries to modern ones based on information and technologies. The analysis of resource-rich countries' stock markets allowed us to conclude that partially these countries are on the way to an economy dominated by intangible assets. Nevertheless, despite the declared intentions, some traditional attachments to raw materials are still present. At the same time, most companies in the optimal portfolios of these markets have a long-term nonlinear strategy. The impact of the pandemic COVID-19 turned out to be significant, but short-lived. The research was carried out based on R and Python packages.