Modern Networks Research Articles

What should an ordinary person understand about the work of generative artificial intelligence? Proceedings of the competition "TRIZformashka-2024"

Three basic aspects of the operation of generative neural network models (generative artificial intelligence) are discussed in the article: the concept of a "token", the probabilistic nature of the generated response, and the concept of a "large model", the size of which ensures the pseudo-intelligent behavior of neural network chatbots. The issues of implementing generative models, areas and methods of their application are not discussed in principle.The materials of the "TRIZformashka-2024" competition, which was dedicated to neural network models, are provided. The fact of pseudo-intelligence of generative models is demonstrated. It turns out that a model trained on a single phrase "mama myla ramu" ("mama washed the frame") and using a context of a single letter for generation can sometimes behave as if it knows the rules of declension in the Russian language and is able to change a word by case!The concept of a "token" is considered in relation to the generation of texts, pictures and passwords. On the basis of "tokens" a practically useful method of generating passwords is built, difficult to guess, but easy to reproduce (difficult to forget).The concept of a "large model" is presented clearly and intelligibly due to its "visualization" by comparing it with physical quantities. (If one parameter of a neural network weighed one gram, then 200 freight trains would be needed to transport it. If one parameter had a length one millimeter, then the neural network would revolve around the Earth at the equator 25 times. If one second was required to learn one parameter, then it would be necessary to start training a modern neural network in the times of the Cro-Magnons.)The materials will be useful for studying generative artificial intelligence at any age.

Large Language Models Meet Next-Generation Networking Technologies: A Review

The evolution of network technologies has significantly transformed global communication, information sharing, and connectivity. Traditional networks, relying on static configurations and manual interventions, face substantial challenges such as complex management, inefficiency, and susceptibility to human error. The rise of artificial intelligence (AI) has begun to address these issues by automating tasks like network configuration, traffic optimization, and security enhancements. Despite their potential, integrating AI models in network engineering encounters practical obstacles including complex configurations, heterogeneous infrastructure, unstructured data, and dynamic environments. Generative AI, particularly large language models (LLMs), represents a promising advancement in AI, with capabilities extending to natural language processing tasks like translation, summarization, and sentiment analysis. This paper aims to provide a comprehensive review exploring the transformative role of LLMs in modern network engineering. In particular, it addresses gaps in the existing literature by focusing on LLM applications in network design and planning, implementation, analytics, and management. It also discusses current research efforts, challenges, and future opportunities, aiming to provide a comprehensive guide for networking professionals and researchers. The main goal is to facilitate the adoption and advancement of AI and LLMs in networking, promoting more efficient, resilient, and intelligent network systems.

High data rate spin-wave transmitter

Spin-wave devices have recently become a strong competitor in computing and information processing owing to their excellent energy efficiency. Researchers have explored magnons, the quanta of spin-waves, as an information carrier and significant progress has occurred in both excitation and computation. However, most transmission designs remain immature in terms of data rate and information complexity as they only utilize simple spin-wave pulses and suffer from signal distortion. In this work, using micromagnetic simulations, we demonstrate a spin-wave transmitter that operates reliably at a data rate of 4 Gbps over significant (multi-micron) distances with error rates as low as 10−14. Spin-wave amplitude is used to encode information. Carrier frequency and data rate are carefully chosen to restrict dispersion spreading, which is the main reason for signal distortion. We show that this device can be integrated into either pure-magnonic circuits or modern electronic networks. Our study reveals the potential for achieving an even higher data rate of 10 Gbps and also offers a comprehensive and logical methodology for performance tuning.

A network security protection scheme for tax system based on elliptic curve cryptography

With the rapid development of modern information technology and network technology, the construction of tax informatization plays an important role in improving the level of tax management and work efficiency. However, in the centralized data processing mode, the establishment of large data center and the efficient, stable, and secure connection of server bandwidth have become key challenges. Moreover, the openness and interconnectivity of the tax system network pose various network security threats, seriously threatening the integrity and security of data. For the above issues, this article proposes a tax system network security protection scheme based on elliptic curve cryptography. This scheme combines hash function and elliptic curve cryptography to achieve user anonymity and secure generation of session keys, effectively ensuring the secure transmission of confidential information. The security and effectiveness of the scheme are verified through informal and formal security analysis based on security proof, BAN logic and AVISPA tool. And a comprehensive evaluation of the scheme is conducted, and the results show that the scheme significantly reduced the costs while providing security features.

Comparison of Cloud Computing, SaaS, SOA and Microservices

A network application is an application that runs on a single host and facilitates communication with another application running on a different computer or host. Numerous network applications have been introduced, including cloud computing, SaaS, SOA, and microservices. However, these technologies have often been used interchangeably, leading to confusion among developers and researchers. Therefore, the objective of this study is to review and compare the available network applications, specifically cloud computing, Software as a Service (SaaS), Service Oriented Architecture (SOA), and microservices. This study conducts a comparative analysis by examining several important criteria such as definition, type, language, interoperability, characteristics, and examples. By providing clear and precise definitions for each technology, the study aims to clarify their distinctions and appropriate usage contexts. The findings successfully establish the correct terminology for each technology and offer detailed discussions on the specific platforms and situations where each is most suitable. This comprehensive review serves as a valuable resource for developers and researchers seeking to navigate the complexities of modern network applications.

Securing the Internet of Things with Ascon-Sign

With a Cryptographically-Relevant Quantum Computer (CRQC) estimated to be viable within the next 15 years, the development of post-quantum security is imperative. Previously secure networks may soon fall victim to these CRQCs as they will likely attack the weakest link in a network. In modern networks, these weak-links are often present in the form of Internet of Things (IoT) devices, as the resource constrains imposed by these wireless nodes leads to lowered security. We offer the first Ascon-Sign implementation for resource constrained FPGAs, which allows a wireless sensor network to verify nodes. Our design runs twice as fast as similarly-area constrained devices, and shows a 33% reduction in power per operation. We demonstrate the capability of our design by integrating it with a wireless sensor network for weather detecting. We also propose an amendment to the Ascon-Sign specification that allows for shortened processing time and lower memory requirements.

Efficient Privacy-Preserving Machine Learning with Lightweight Trusted Hardware

In this paper, we propose a new secure machine learning inference platform assisted by a small dedicated security processor, which will be easier to protect and deploy compared to today's TEEs integrated into high-performance processors. Our platform provides three main advantages over the state-of-the-art: (i) We achieve significant performance improvements compared to state-of-the-art distributed Privacy-Preserving Machine Learning (PPML) protocols, with only a small security processor that is comparable to a discrete security chip such as the Trusted Platform Module (TPM) or on-chip security subsystems in SoCs similar to the Apple enclave processor. In the semi-honest setting with WAN/GPU, our scheme is 4X-63X faster than Falcon (PoPETs'21) and AriaNN (PoPETs'22) and 3.8X-12X more communication efficient. We achieve even higher performance improvements in the malicious setting. (ii) Our platform guarantees security with abort against malicious adversaries under honest majority assumption. (iii) Our technique is not limited by the size of secure memory in a TEE and can support high-capacity modern neural networks like ResNet18 and Transformer. While previous work investigated the use of high-performance TEEs in PPML, this work represents the first to show that even tiny secure hardware with very limited performance can be leveraged to significantly speed-up distributed PPML protocols if the protocol can be carefully designed for lightweight trusted hardware.

Towards fine-grained load balancing with dynamical flowlet timeout in datacenter networks

In modern datacenter networks (DCNs), load balancing mechanisms are widely deployed to enhance link utilization and alleviate congestion. Recently, a large number of load balancing algorithms have been proposed to spread traffic among the multiple parallel paths. The existing solutions make rerouting decisions for all flows once they experience congestion on a path. They are unable to distinguish between the flows that really need to be rerouted and the flows that potentially have negative effects due to rerouting, resulting in frequently ineffective rerouting. Fine-grained rerouting will also cause severe packet reordering, especially in asymmetric topology scenarios. To address the above issues, we present a fine-grained traffic-differentiated load balancing (TDLB) mechanism, which aims to distinguish flows that are necessarily to be rerouted and reroute traffic in fine-grained without packet reodering. Specifically, TDLB distinguishes the traffic that must be rerouted through the host pair information in the packet header, and selects an optimal path for rerouting. To prevent severe packet reodering caused by excessive path delay differences, TDLB dynamically adjusts the flowlet timeout to segment the traffic and select the optimal path for rerouting. The NS-2 simulation results show that TDLB effectively reduces tail latency and average flow completion time (FCT) for short flows by up to 49% and 46%, respectively, compared to the state-of-the-art load balancing schemes.

Потенциальные угрозы, связанные с применением искусственного интеллекта в преступных действиях

The article is devoted to the issues of assessing the threats associated with the use of artificial intelligence in the implementation of criminal actions, taking into account the rapid development of technical capabilities and areas of application of modern neural networks, since the degree of readiness for counteraction directly affects the effectiveness of such counteraction and the efficiency of its application.

Lifelong Generative Adversarial Autoencoder.

Lifelong learning describes an ability that enables humans to continually acquire and learn new information without forgetting. This capability, common to humans and animals, has lately been identified as an essential function for an artificial intelligence system aiming to learn continuously from a stream of data during a certain period of time. However, modern neural networks suffer from degenerated performance when learning multiple domains sequentially and fail to recognize past learned tasks after being retrained. This corresponds to catastrophic forgetting and is ultimately induced by replacing the parameters associated with previously learned tasks with new values. One approach in lifelong learning is the generative replay mechanism (GRM) that trains a powerful generator as the generative replay network, implemented by a variational autoencoder (VAE) or a generative adversarial network (GAN). In this article, we study the forgetting behavior of GRM-based learning systems by developing a new theoretical framework in which the forgetting process is expressed as an increase in the model's risk during the training. Although many recent attempts have provided high-quality generative replay samples by using GANs, they are limited to mainly downstream tasks due to the lack of inference. Inspired by the theoretical analysis while aiming to address the drawbacks of existing approaches, we propose the lifelong generative adversarial autoencoder (LGAA). LGAA consists of a generative replay network and three inference models, each addressing the inference of a different type of latent variable. The experimental results show that LGAA learns novel visual concepts without forgetting and can be applied to a wide range of downstream tasks.

Enhancing Wireless Communication Efficiency Using Dynamic Nonlinear Distortion Adaptive Optimization Analysis

Abstract Modern communication networks require improved signal quality and spectrum efficiency. Modern wireless communication relies heavily on the multiple input multiple output (MIMO) technology to improve spectrum efficiency and achieve faster data rates. However, the performance of MIMO systems is significantly hampered by non-linear distortions in power amplifiers (PA), especially under dynamic operating conditions. Traditional methods to reduce these distortions often lack the adaptability required for effective optimization. The proposed dynamic nonlinear distortion adaptive optimization analysis (DNDAOA) therefore introduces an adaptive approach that dynamically adjusts the pre-distortion signals using real-time feedback mechanisms. DNDAOA effectively reduces non-linear distortions and thus improves the performance of MIMO systems. Simulation results show that DNDAOA significantly increases spectrum efficiency, reduces error rates, and improves signal quality. This method is widely used in areas such as wireless telecommunications, internet of things (IoT), autonomous vehicles, and smart infrastructure, driving advances in connectivity, reliability, and data integrity.

Revisiting Douai Ms 785: Musical Cultures of Exiled English Convents through the Lens of a Miscellany

Abstract Scholarship on early modern English Catholic music after the reformations tends to focus on the activities of male musicians and male institutions. Despite increased study of English convent culture by scholars of religious, social, and literary history, there remains little specialist examination of music at post-Reformation English convents in exile, and their role in wider musical networks in early modern Europe is markedly under-acknowledged. This article aims to highlight how complex miscellanies with links to English monastic institutions in exile can offer insight into the convents’ otherwise elusive musical world. Using a hitherto unanalysed miscellany – Douai Ms 785 – this article will show how codicological study of manuscripts, combined with study of concordances and unica, can illuminate the role of English convents in early modern musical networks. In doing so, it will demonstrate the need to understand miscellanies like Douai Ms 785 as witness to interacting, overlapping musical and religious ecosystems in early modern Europe.

PROBLEMS OF MANAGING TECHNOLOGICAL REGIMES IN GAS DISTRIBUTION NETWORKS AND MODERN APPROACHES TO THEIR SOLUTION

This article systematically discusses the problems of managing technological regimes in gas distribution networks and modern approaches to their resolution. The scientific-theoretical and practical aspects of managing gas distribution networks (GDN) are analyzed. Based on the obtained results, a scientifically grounded comparative explanation of the existing contradictions in modern and traditional gas distribution networks is provided, and the role of key factors in the emergence of these contradictions is investigated. It is shown that technological backwardness in traditional GDNs increases energy consumption and operational costs, thereby reducing the economic efficiency of the system and increasing the risk of accidents. The research suggests decision-making methods for solving management problems in traditional GDNs by maintaining the technical-structural level and implementing telemetry systems. It is indicated that the widespread use of telemetry and telemechanics tools in the gas distribution industry enhances the safe delivery of gas to end-users and significantly reduces service costs. Another focus of the research is on the issues related to the processing of spatial information for the management of the technical components of gas distribution systems. The article also analyzes scientific research dedicated to the application of modern IT methods and tools in the management of GDNs, investigating management and forecasting models using neural networks and intelligent methods. Keywords: Gas distribution networks, modern GDN networks, traditional GDN networks, GIS systems, digital technologies, technogenic risks, neural networks, forecasting methods

A Hybrid Approach for CT Image Noise Reduction Combining Method Noise-CNN and Shearlet Transform

The presence of gaussian noise commonly weakens the diagnostic precision of low-dose CT imaging. A novel CT image denoising technique that integrates the non-subsampled shearlet transform (NSST) with Bayesian thresholding, and incorporates a modern method noise Deep Convolutional neural network (DCNN) based post-processing operation on denoised images to strengthen low-dose CT imaging quality. The hybrid method commences with NSST and Bayesian thresholding to mitigate the initial noise while preserving crucial image features, such as corners and edges. The novel aspect of the proposed approach is its successive application of a DnCNN on initial denoised image, which learns and removes residual noise patterns from denoised images, thereby enhancing fine detail preservation. This dual-phase strategy addresses both noise suppression and image-detail preservation. The proposed technique is evaluated through the use of metrics, such as PSNR, SNR, SSIM, ED, and UIQI. The results demonstrate that the hybrid approach outperforms standard denoising techniques in preserving image quality and fine details.

Data-Driven Relationship for Reduction Factor of Ballasted Railway Bridge Deflections Due to Load Distribution Within Track

Increasing the operating speed of the trains on modern networks necessitates performing dynamic analyses to assess the performance of bridges under passage of trains. The detailed investigation of their responses requires constructing complex computational models capable to take the train–track–bridge interaction effects into account. Such models have successfully been developed; however, employing those elaborated models for practical engineering applications, or to perform studies that require a large number of analyses may become infeasible. Among such situations are conducting probabilistic investigations, screening of entire networks, or sensitivity analyses. These concerns have been addressed by employing simplified models mostly relying on moving load modeling strategy which disregards the train–track–bridge interaction effects. Those neglected contributions can be compensated by implementing additional correction factors. The distribution of loads within track is one of those disregarded effects where a reduction factor is recommended by design guidelines to take its contribution into account. It has been shown that the existing relationship for these reduction factors delivers an acceptable performance for vertical accelerations, while showing a less favorable performance for displacements. Then, a data-driven strategy is adopted in this study to propose easy-to-apply relationships for reduction factors of deflections, due to load distribution within the track. In this context, three different distributive lengths of triangular load footprints have been considered, namely 2.0, 2.5 and 3.0[Formula: see text]m. The procedure employed has trained and tested for more than 1[Formula: see text]200 train configurations, comprising conventional, articulated and regular vehicles, and including several tens of thousand data points for each distributive length. The performance observed in the new models revealed a considerable improvement with respect to the existing relationship.

Learnability of state spaces of physical systems is undecidable

Despite an increasing role of machine learning in science, there is a lack of results on limits of empirical exploration aided by machine learning. In this paper, we construct one such limit by proving undecidability of learnability of state spaces of physical systems. We characterize state spaces as binary hypothesis classes of the computable Probably Approximately Correct learning framework. This leads to identifying the first limit for learnability of state spaces in the agnostic setting. Further, using the fact that finiteness of the combinatorial dimension of hypothesis classes is undecidable, we derive undecidability for learnability of state spaces as well. Throughout the paper, we try to connect our formal results with modern neural networks. This allows us to bring the limits close to the current practice and make a first step in connecting scientific exploration aided by machine learning with results from learning theory.

Will Network Infrastructure Optimization Make Manufacturing Enterprises "Out of Real to Virtual"? --- A Quasi-Natural Experiment from the "Broadband China" Policy

Manufacturing is the most important component of the real economy, and the application of modern network information technology can effectively improve the operational efficiency of manufacturing enterprises. In the context of "Made in China 2025", it is particularly important to prevent manufacturing enterprises from deviating from their main business and becoming excessively financialized. Maintaining the essential business activities of manufacturing enterprises is of great importance to China's economic development. Based on the operating data of all A-share listed manufacturing companies from 2010 to 2020 and combining the "Broadband China" demonstration city policy, this paper uses a multi-point double-differenced model to explore the relationship between network infrastructure optimization and excessive financialization of manufacturing enterprises. The research shows that: 1) Network infrastructure optimization will make manufacturing enterprises increase their financialization level, and this effect is mainly manifested through optimizing accounts receivable management. 2) From the perspective of enterprise size, compared with small and medium-sized manufacturing enterprises, large manufacturing enterprises are more likely to improve their financialization level through network infrastructure optimization. 3) From the perspective of the degree of diversified operations, manufacturing enterprises with high diversified operations are more likely to take advantage of network infrastructure to improve their financialization level. This study provides reference and guidance for later researchers on the impact of "new infrastructure" represented by network infrastructure construction on the business model of manufacturing enterprises.

ShuffleNeMt: modern lightweight convolutional neural network architecture

ShuffleNeMt: modern lightweight convolutional neural network architecture

Computer Networks Cybersecurity Monitoring Based on Deep Learning Model

ABSTRACTEffective cybersecurity monitoring is essential for safeguarding computer networks against evolving threats. However, the increasing scale, complexity, and data volume of modern networks pose significant challenges to traditional monitoring methods. To address these challenges, this article proposes a deep learning‐based approach for computer network cybersecurity monitoring. Leveraging the MnasNet‐LSTM model, network traffic data is classified into distinct categories, including normal traffic and cyberattacks. The model is trained using the CICIDS2017 dataset, yielding promising results with a classification accuracy of approximately 97.05% and a minimal error rate of 8.46%.

A REVIEW OF MACHINE LEARNING AND FEATURE SELECTION TECHNIQUES FOR CYBERSECURITY ATTACK DETECTION WITH A FOCUS ON DDOS ATTACKS

This study provides a systematic review of machine learning (ML) techniques applied in intrusion detection systems (IDS), with a particular focus on Random Forest (RF), Support Vector Machine (SVM), and Decision Tree (DT). Following the PRISMA guidelines, a comprehensive search of relevant databases identified 205 articles, from which 68 were selected for detailed analysis. The findings highlight that RF consistently outperforms other models, achieving accuracy rates as high as 99.72% in detecting Distributed Denial of Service (DDoS) attacks due to its ensemble learning approach. SVM, while effective in specific scenarios with binary classification tasks, struggles with scalability and high-dimensional datasets, though feature selection significantly improves its performance. DT models, known for their simplicity and interpretability, are prone to overfitting, but this issue is mitigated when combined with feature selection techniques. The study further emphasizes the importance of feature selection in enhancing IDS accuracy and efficiency across various models. Additionally, ensemble and hybrid methods, which combine multiple ML techniques, offer promising improvements in detection accuracy and real-time performance. These findings underscore the potential of machine learning, particularly through the use of ensemble and hybrid approaches, to significantly improve cybersecurity measures in modern networks.