AI-Driven Cybersecurity: An Overview, Security Intelligence Modeling and Research Directions

SN Computer Science, Journal Year: 2021, Volume and Issue: 2(3)

Published: March 26, 2021

Language: Английский

---

Machine Learning: Algorithms, Real-World Applications and Research Directions

Published: March 8, 2021

In the current age of Fourth Industrial Revolution ($4IR$ or Industry $4.0$), digital world has a wealth data, such as Internet Things (IoT) cybersecurity mobile business social media health etc. To intelligently analyze these data and develop corresponding real-world applications, knowledge artificial intelligence (AI), particularly, machine learning (ML) is key. Various types algorithms supervised, unsupervised, semi-supervised, reinforcement exist in area. Besides, deep learning, which part broader family methods, can on large scale. this paper, we present comprehensive view that be applied to enhance capabilities an application. Thus, study's key contribution explaining principles different techniques their applicability various applications areas, cybersecurity, smart cities, healthcare, business, agriculture, many more. We also highlight challenges potential research directions based our study. Overall, paper aims serve reference point for not only application developers but decision-makers researchers particularly from technical view.

Language: Английский

---

A modified Adam algorithm for deep neural network optimization

Neural Computing and Applications, Journal Year: 2023, Volume and Issue: 35(23), P. 17095 - 17112

Published: April 25, 2023

Abstract Deep Neural Networks (DNNs) are widely regarded as the most effective learning tool for dealing with large datasets, and they have been successfully used in thousands of applications a variety fields. Based on these trained to learn relationships between various variables. The adaptive moment estimation (Adam) algorithm, highly efficient optimization is algorithm fields training DNN models. However, it needs improve its generalization performance, especially when large-scale datasets. Therefore, this paper, we propose HN Adam, modified version Adam Algorithm, accuracy convergence speed. HN_Adam by automatically adjusting step size parameter updates over epochs. This automatic adjustment based norm value update formula according gradient values obtained during Furthermore, hybrid mechanism was created combining standard AMSGrad algorithm. As result changes, like stochastic descent (SGD) has good performance achieves fast other algorithms. To test proposed evaluated train deep convolutional neural network (CNN) model that classifies images using two different datasets: MNIST CIFAR-10. results compared basic SGD addition five recent In comparisons, outperforms algorithms terms AdaBelief competitive testing speed (represented consumed time), HN-Adam an improvement 1.0% 0.29% dataset, 0.93% 1.68% CIFAR-10 respectively.

Language: Английский

---

Deep Learning: A Comprehensive Overview on Techniques, Taxonomy, Applications and Research Directions

Published: Aug. 2, 2021

Deep learning (DL), a branch of machine (ML) and artificial intelligence (AI) is nowadays considered as core technology today's Fourth Industrial Revolution (4IR or Industry 4.0). Due to its capabilities from data, DL originated neural network (ANN), has become hot topic in the context computing, widely applied various application areas like healthcare, visual recognition, cybersecurity, many more. However, building an appropriate model challenging task, due dynamic nature variations real-world problems data. Moreover, lack understanding turns methods into black-box machines that hamper development at standard level. This article presents structured comprehensive view on techniques including taxonomy considering types tasks supervised unsupervised. In our taxonomy, we take account deep networks for discriminative learning, unsupervised generative well hybrid relevant others. We also summarize where can be used. Finally, point out ten potential aspects future generation modeling with research directions. Overall, this aims draw big picture used reference guide both academia industry professionals.

Language: Английский

---

Enhancing mental health with Artificial Intelligence: Current trends and future prospects

Journal of Medicine Surgery and Public Health, Journal Year: 2024, Volume and Issue: 3, P. 100099 - 100099

Published: April 17, 2024

Artificial Intelligence (AI) has emerged as a transformative force in various fields, and its application mental healthcare is no exception. Hence, this review explores the integration of AI into healthcare, elucidating current trends, ethical considerations, future directions dynamic field. This encompassed recent studies, examples applications, considerations shaping Additionally, regulatory frameworks trends research development were analyzed. We comprehensively searched four databases (PubMed, IEEE Xplore, PsycINFO, Google Scholar). The inclusion criteria papers published peer-reviewed journals, conference proceedings, or reputable online databases, that specifically focus on field offer comprehensive overview, analysis, existing literature English language. Current reveal AI's potential, with applications such early detection health disorders, personalized treatment plans, AI-driven virtual therapists. However, these advancements are accompanied by challenges concerning privacy, bias mitigation, preservation human element therapy. Future emphasize need for clear frameworks, transparent validation models, continuous efforts. Integrating therapy represents promising frontier healthcare. While holds potential to revolutionize responsible implementation essential. By addressing thoughtfully, we may effectively utilize enhance accessibility, efficacy, ethicality thereby helping both individuals communities.

Language: Английский

---

A review on machine learning techniques for secured cyber-physical systems in smart grid networks

Energy Reports, Journal Year: 2024, Volume and Issue: 11, P. 1268 - 1290

Published: Jan. 9, 2024

The smart grid (SG) is an advanced cyber-physical system (CPS) that integrates power infrastructure with information and communication technologies (ICT). This integration enables real-time monitoring, control, optimization of electricity demand supply. However, the increasing reliance on ICT infrastructures has made SG-CPS more vulnerable to cyberattacks. Hence, securing from these threats crucial for its reliable operation. In recent literature, machine learning (ML) techniques and, recently, deep (DL) have been used by several studies implement cybersecurity countermeasures against cyberattacks in SG-CPS. Nevertheless, achieving high performance state-of-the-art constrained certain challenges, including hyperparameter optimization, feature extraction selection, lack models' transparency, data privacy, attack data. paper reviews advancement using ML DL It analyzes constraints need be addressed improve achieve implementation. various types cyberattacks, requirements, security standards protocols are also discussed establish a comprehensive understanding context will serve as guide new experienced researchers.

Language: Английский

---

THE PREDICTION OF HEART DISEASE USING MACHINE LEARNING ALGORITHMS

Science Journal of University of Zakho, Journal Year: 2024, Volume and Issue: 12(3), P. 285 - 293

Published: July 14, 2024

Heart disease threatens the lives of around one individual per minute, establishing it as foremost cause mortality in contemporary era. A wide range individuals over globe has encountered intricacies associated with cardiovascular illness. Various factors, such hypertension, elevated levels cholesterol, and an irregular pulse rhythm hinder early identification a disease. In cardiology, similar to other branches Medicine, timely precise cardiac diseases is utmost importance. Anticipating onset heart failure at appropriate moment can provide challenges, particularly for cardiologists surgeons. Fortunately, categorisation forecasting models assist medical business real applications data. Regarding this, Machine Learning (ML) algorithms techniques have benefited from automated analysis several datasets complex data aid community diagnosing heart-related diseases. Predicting if patient early-stage primary goal this paper. prior study that worked on Erbil Disease dataset proved Naïve Bayes (NB) got accuracy 65%, which worst classifier, while Decision Tree (DT) obtained highest 98%. article, comparison been applied using same (i.e., dataset) between multiple ML algorithms, instance, LR (Logistic Regression), KNN (K-Nearest Neighbours), SVM (Support Vector Machine), DT (Decision Tree), MLP (Multi-Layer Perceptron), NB (Naïve Bayes) RF (Random Forest). Surprisingly, we 98% after applying LR, MLP, RF, was best outcome. Furthermore, by classifier differed incredibly received work.

Language: Английский

---

Artificial intelligence for hydrogen-enabled integrated energy systems: A systematic review

International Journal of Hydrogen Energy, Journal Year: 2024, Volume and Issue: unknown

Published: Aug. 1, 2024

Hydrogen-enabled Integrated Energy Systems (H-IES) stand out as a promising solution with the potential to replace current non-renewable energy systems. However, their development faces challenges and has yet achieve widespread adoption. These main include complexity of demand supply balancing, dynamic consumer demand, in integrating utilising hydrogen. Typical management strategies within domain rely heavily on accurate models from experts or conventional approaches, such simulation optimisation which cannot be satisfied real-world operation H-IES. Artificial Intelligence (AI) Advanced Data Analytics (ADA), especially Machine Learning (ML), ability overcome these challenges. ADA is extensively used across several industries, however, further investigation into incorporation hydrogen for purpose enabling H-IES needs investigated. This paper presents systematic literature review study research gaps, directions, benefits ADA, well role

Language: Английский

---

The role of IT in sustainable environmental management: A global perspective review

International Journal of Science and Research Archive, Journal Year: 2024, Volume and Issue: 11(1), P. 1874 - 1886

Published: Feb. 18, 2024

This review offers a comprehensive overview of the intricate relationship between Information Technology (IT) and sustainable environmental management on global scale. As world grapples with challenges, understanding pivotal role IT in fostering sustainability becomes increasingly imperative. The begins by acknowledging pressing issues faced globally, including climate change, resource depletion, biodiversity loss. It highlights potential to serve as transformative force addressing these challenges practices across various industries. explores how contributes through improved monitoring, data collection, analysis. delves into technologies such Internet Things (IoT) devices, sensors, satellite imaging providing real-time data. information enables better decision-making, management, implementation eco-friendly practices. Furthermore, examines promoting energy efficiency reducing carbon footprints. discusses adoption smart grids, systems, software solutions that contribute optimizing consumption impact. also concept "green IT," emphasizing importance adopting environmentally friendly design, production, disposal equipment. initiatives aimed at minimizing electronic waste circular economy within industry. Additionally, perspective sheds light facilitates collaboration knowledge sharing among nations organizations. underlines significance international cooperation leveraging for technology achieving goals. In conclusion, this underscores multifaceted globally. emphasizes innovation, driving collaborative efforts towards more resilient future.

Language: Английский

---

A comprehensive review of machine learning's role in enhancing network security and threat detection

World Journal of Advanced Research and Reviews, Journal Year: 2024, Volume and Issue: 21(2), P. 877 - 886

Published: Feb. 17, 2024

As network security threats continue to evolve in complexity and sophistication, there is a growing need for advanced solutions enhance threat detection capabilities. Machine learning (ML) has emerged as powerful tool this context, offering the potential detect mitigate real-time by analyzing vast amounts of data. This comprehensive review explores role machine enhancing detection. The begins providing an overview current landscape challenges faced traditional approaches. It then delves into fundamental principles its application security. Various techniques, including supervised learning, unsupervised deep are discussed detail, highlighting their strengths limitations context Next, examines different aspects security, intrusion detection, malware anomaly behavioral analysis. Case studies real-world examples presented illustrate effectiveness learning-based approaches identifying mitigating threats. Furthermore, discusses considerations associated with deploying environments, such data privacy, model interpretability, adversarial attacks. Strategies addressing these improving robustness models explored. Finally, outlines future research directions opportunities leveraging Areas federated explainable AI identified promising avenues further investigation. In summary, provides insights By capabilities algorithms organizations can strengthen defenses against cyber better protect networks sensitive

Language: Английский

---