Power electronics anomaly detection and diagnosis with machine learning and deep learning methods: A survey

International Journal of Science and Research Archive, Год журнала: 2024, Номер 11(2), С. 730 - 739

Опубликована: Март 30, 2024

Power electronics pertains to the conception, regulation, and utilization of electronic power circuits proficiently administer transform electrical energy. play a crucial role in maintaining reliability, efficiency, security complex production systems. Also, increasingly important various applications such as renewable energy systems, electric vehicles, industrial automation. However, modern systems are vulnerable both cyber physical anomalies due integration information communication technologies. So far, different methods have been used detect abnormalities. This survey provides an overview state-of-the-art anomaly detection using machine learning deep methods. It highlights potential these techniques addressing growing complexity vulnerability

Язык: Английский

---

Ethics, Algorithms, and the Rules of Evidence

IGI Global eBooks, Год журнала: 2025, Номер unknown, С. 103 - 124

Опубликована: Фев. 28, 2025

Forensic intelligence, combined with the power of deep learning, has made significant leaps in revolutionizing crime investigation by allowing law enforcement agencies to process complex data, identify patterns, and predict criminal behaviors efficiency. Traditional forensic methods can be improved through machine learning techniques implementation natural language processing, which alter digital investigations. A few key ways that these two approaches benefit computer investigations include automating analysis evidence, enhancing accuracy biometrics, detecting related hacking activities traditional methods. It supports data-driven policing improves speed case settlements. Yet, concerns including algorithmic bias, data privacy, legal admissibility AI-generated evidence underscore ethical social implications technologies. This chapter will discuss transformative intelligence its applications, ethics, future.

Язык: Английский

---

Text Mining and Unsupervised Deep Learning for Intrusion Detection in Smart-Grid Communication Networks

IoT, Год журнала: 2025, Номер 6(2), С. 22 - 22

Опубликована: Март 26, 2025

The Manufacturing Message Specification (MMS) protocol is frequently used to automate processes in IEC 61850-based substations and smart-grid systems. However, it may be susceptible a variety of cyber-attacks. A protection strategy deploy intrusion detection systems monitor network traffic for anomalies. Conventional approaches detecting anomalies require large number labeled samples are therefore incompatible with high-dimensional time series data. This work proposes an anomaly method sequences based on bidirectional LSTM autoencoder. Additionally, text-mining TF-IDF vectorizer truncated SVD presented data preparation feature extraction. proposed representation approach outperformed word embeddings (Doc2Vec) by better preserving critical domain-specific keywords MMS while reducing the complexity model training. Unlike embeddings, which attempt capture semantic relationships that not exist structured protocols, focuses token frequency importance, making more suitable communications. To address limitations existing rely samples, learns properties patterns normal unsupervised manner. results demonstrate can learn potential features from maintaining high True Positive Rate.

Язык: Английский

---

Deep Learning-Based Anomaly Detection in Network Traffic for Cyber Threat Identification

Опубликована: Май 25, 2024

An essential aspect of cybersecurity is the continuously growing threat landscape, which necessitates use advanced anomaly detection techniques in network data. The traditional approach might often be inadequate when it comes to addressing intricate cyber-security issues. Therefore, possible that deep learning approaches superior terms accuracy and performance. primary objective our study provide a novel algorithm combines Convolutional Neural Networks (CNNs), Recurrent (RNNs), autoencoders, GANs create comprehensive strategy for detecting anomalies. This technique aims solve research gaps have not been previously explored. By using MTA-KDD'19 dataset, enhances precision by achieving remarkable rate 95% various types traffic abnormalities. discovery only demonstrated harmfulness learning-based but also highlighted effectiveness these measures reducing issue, particularly faced with diverse threats. development security procedures.

Язык: Английский

---

Machine Learning in Information and Communications Technology: A Survey

Information, Год журнала: 2024, Номер 16(1), С. 8 - 8

Опубликована: Дек. 27, 2024

The rapid growth of data and the increasing complexity modern networks have driven demand for intelligent solutions in information communications technology (ICT) domain. Machine learning (ML) has emerged as a powerful tool, enabling more adaptive, efficient, scalable systems this field. This article presents comprehensive survey on application ML techniques ICT, covering key areas such network optimization, resource allocation, anomaly detection, security. Specifically, we review effectiveness different models across ICT subdomains assess how integration enhances crucial performance metrics, including operational efficiency, scalability, Lastly, highlight challenges future directions that are critical continued advancement ML-driven innovations ICT.

Язык: Английский

---

Artificial intelligence and its impact on job opportunities among university students in North Lima, 2023

ICST Transactions on Scalable Information Systems, Год журнала: 2023, Номер 10(5)

Опубликована: Сен. 6, 2023

Introduction: Artificial intelligence is a technology that replaces human activities, favors business productivity and raises concerns about job losses economic social challenges. Method: The research uses quantitative approach non-experimental study design with correlational scope. It identifies two variables: artificial (AI) opportunity. evaluates students of the Adult Education Program (PFA) Universidad César Vallejo. Data collection was done through virtual survey Likert scale questions. Results: conducted descriptive analysis opportunities. A moderate positive correlation observed between both variables, suggesting significant relationship level opportunities respondents. Discussion: reveals knowledge perception important to adapt this global improve employability. Conclusion: findings support transforms society labor market. Although 86% know AI, most need more training in field, even areas projected growth AI-related employment.

Язык: Английский

---

Detection of Unfocused EEG Epochs by the Application of Machine Learning Algorithm

Sensors, Год журнала: 2024, Номер 24(15), С. 4829 - 4829

Опубликована: Июль 25, 2024

Electroencephalography (EEG) is a non-invasive method used to track human brain activity over time. The time-locked EEG an external event known as event-related potential (ERP). ERP can be biomarker of perception and other cognitive processes. success research depends on the laboratory conditions attentiveness test subjects. Specifically, inability control experimental variables has reduced in real world. This study collected data under various circumstances within auditory oddball paradigm experiment enable use active normal conditions. Then, epochs were analyzed identify unfocused epochs, affected by typical artifacts distortion. For initial comparison, ability four unsupervised machine learning algorithms (MLAs) was evaluated epochs. their accuracy compared with inspection current analysis tool (EEGLab). All MLAs typically 95-100% accurate. In summary, our finds that humans might miss subtle differences regular patterns, but could efficiently those. Thus, suggests perform better for detecting two standard methods.

Язык: Английский

---

AI-Driven Threat Detection and Response Systems for Secure National Infrastructure Networks: A Comprehensive Review

International Journal of Latest Technology in Engineering Management & Applied Science, Год журнала: 2024, Номер 13(7), С. 82 - 92

Опубликована: Авг. 5, 2024

Abstract: Due to the increased complexity and damage of cyberattacks in this digital age, security national infrastructure networks has become a vital concern. However, possible approach improve cybersecurity these crucial is incorporate artificial intelligence (AI) into threat detection response systems; rapidly evaluate large data sets, identify anomalies, automate countermeasures lessen effects cyberattacks. The impact, implementation approaches for anomaly automation AI-powered solutions safeguarding are examined paper. Understanding how AI technologies used response, reviewing operational usefulness enhancing measures evaluating deployment systems critical settings were also examined. study revealed that speed accuracy greatly by systems. capacity can potentially reduce need human analysts, while providing faster mitigation. Additionally, across sectors indicates its practicality situations it may adapt new threats. In conclusion, AI-driven an important development network cybersecurity. Therefore, improving recognize address cyber-attacks ultimately increase overall resilience infrastructures.

Язык: Английский

---

Detecting and Analyzing Network Attacks: A Time-Series Analysis Using the Kitsune Dataset

Journal of Emerging Computer Technologies, Год журнала: 2024, Номер 5(1), С. 9 - 23

Опубликована: Ноя. 2, 2024

Network security is a critical concern in today’s digital world, requiring efficient methods for the automatic detection and analysis of cyber attacks. This study uses Kitsune Attack Dataset to explore network traffic behavior IoT devices under various attack scenarios, including ARP MitM, SYN DoS, Mirai Botnet. Utilizing Python-based data tools, we preprocess analyze millions packets uncover patterns indicative malicious activities. The employs packet-level time-series visualize detect anomalies specific each type. Key findings include high packet volumes attacks such as SSDP Flood Botnet, with Botnet involving multiple IP addresses lasting over 2 hours. Notable attack-specific behaviors on port -1 targeted ports like 53195. DoS are characterized by their prolonged durations, suggesting significant disruption. Overall, highlights distinctive underscores importance understanding these characteristics enhance response mechanisms.

Язык: Английский

---

Evaluation of the effectiveness of personal electronic health assistants in monitoring patients with chronic diseases

EAI Endorsed Transactions on Pervasive Health and Technology, Год журнала: 2023, Номер 9

Опубликована: Окт. 24, 2023

Introduction: Chronic diseases pose significant challenges in healthcare, which has driven the development of electronic health solutions. The effectiveness these solutions management such as hypertension generated interest, but further in-depth, evidence-based evaluation is required.Objective: study aims to comprehensively evaluate how a customizable web platform, called "HyperVigilance", influences blood pressure control hypertensive patients, considering additional variables patient satisfaction, quality life and costs associated with treatment. In addition, aim explore possible demographic factors that could moderate results.Methodology: was conducted quasi-experimental research design included an intervention group using "HyperVigilance" platform receiving standard medical care. Statistical tests were applied age, gender socioeconomic status considered.Results: use resulted reduction pressure, increased satisfaction marked improvement life, well treatment hypertension.Conclusions: concludes effective controlling improving patients hypertension. results support growing role digital interventions chronic disease management, highlight need for long-term studies exploration different populations more complete understanding their impact.

Язык: Английский

---