Enhancing Cloud Security with Deep Learning-Based Intrusion Detection in Cloud Computing Environments

Published: May 2, 2024

Cloud computing (CC) offers a wide range of on-demand resources and services over the internet. However, due to inherent vulnerability cloud's dispersed architecture, guaranteeing privacy security cloud assets remains tough task. Data service protection in systems is continuous serious concern. This article presents fresh solution this problem by utilizing Deep Learning (DL) techniques for intrusion detection systems. The present research analyses strategies used several Intrusion Detection Systems (IDS). As network infrastructures expand, so do risks, increasing need dependable detection. Developers have created numerous (IDS) response expanding issues that saturate modern computer networks. Improving datasets training testing these solutions requires equal focus as building defense Improved significantly enhance capabilities both offline online models. essay contributes growing corpus on CC conducting thorough examination publicly available network-based IDS datasets. It emphasizes incorporate cutting-edge DL approaches into increased security. continues transform digital world, findings study major implications safeguarding sensitive data key cloud-based ecosystems. Furthermore, they provide current anomaly caused insufficient normal patterns data. accuracy information two, five, twenty-three classes soft-max regression (SMR) feature learning perform similarly 5-class 23-class categorization, it was also found STL completed all categorization categories with above 98% accuracy.

Language: Английский

---

UAV Based YOLOV-8 Optimization Technique to Detect the Small Size and High Speed Drone in Different Light Conditions

Published: March 15, 2024

Public safety and other sectors are paying more attention to uncrewed aerial vehicle ( UA V) identification as UAVs become commonplace in commercial industrial settings. Methods for object detection by progressing rapidly, too. Nevertheless, researchers still need overcome substantial obstacles this field because of the tiny size drones, complicated airspace backdrops, fluctuating light conditions. Addressing these issues, work suggests a small UAV approach that utilizes improved YOLOv8. The first step is upgrade device's recognition capabilities marks addition high-resolution head According study, YOLOv8 network's large target prediction layer, feature extraction, fusion layers eliminated. Only four, eight, or sixteen-sampled maps stored prediction. upgraded network architecture, which feeds from third C2f layer straight into SPPF multi-scale extraction using 16-times down-sampling. After leaving Upsample-ConcatC2f module, fused immediately attached next component. Our boosts performance compared baseline model 11.9% accuracy, 15.2% recall, 9% mean average precision (MAP). decreased 57.9 ^% , number limits 60%. Also, our better suited engineering deployment real-world system applications, it shows apparent benefits comparative studies experiments self-built datasets.

Language: Английский

---

Intelligent Systems For Predictive Maintenance In Industrial IoT

2018 5th IEEE Uttar Pradesh Section International Conference on Electrical, Electronics and Computer Engineering (UPCON), Journal Year: 2023, Volume and Issue: unknown

Published: Dec. 1, 2023

The IIoT has changed the way businesses function by making available real-time data and insights from linked devices. Predictive maintenance is becoming increasingly important since it enables to keep tabs on their machinery fix any problems before they become costly breakdowns. This abstract explores use of recurrent neural networks (RNNs) in creation smart systems for predictive industrial IoT. goal foresee impending equipment failure take corrective measures good time avoid downtime. Downtime extra expenses are two potential outcomes traditional approach, which often reactive or scheduled based previous data. Using machine learning techniques, intelligent driven RNNs provide a possible answer. As subset artificial networks, excel at processing time-series data, them an obvious candidate include feedback loop that allows capture temporal relationships superior typical feedforward forecasting failures past sensor Data gathering preprocessing major hurdles RNN-based Massive volumes produced settings, this must be cleaned, processed, prepared training RNN models. In addition, solutions can only effective with high-quality, consistent When constructed properly, models capable patterns anomalies. These then inform crews critical pieces equipment. By utilising RNNs, not current included but also allowing dynamic adaptable upkeep tactics. Furthermore, successful deployment, necessary combine edge computing cloud-based solutions. order analyse real make prompt decisions, moves closer source. Cloud-based solutions, other hand, scale as needed, store over time. used together, these methods very responsive scalable systems. Intelligent may enhance dependability, shorten repair times, better distribute scarce resources. Predicting when needed helps save money avoiding unneeded maintenance. There advantages using such beyond just financial savings. sustainability lowers energy use, lessens waste, lengthens life machinery. Workplace safety improved result decrease possibility unanticipated failure.

Language: Английский

---

Deep Learning Techniques for Autonomous Navigation of Underwater Robots

2018 5th IEEE Uttar Pradesh Section International Conference on Electrical, Electronics and Computer Engineering (UPCON), Journal Year: 2023, Volume and Issue: unknown, P. 1630 - 1635

Published: Dec. 1, 2023

Underwater exploration, environmental monitoring, and infrastructure inspection are just few of the many applications where autonomous navigation underwater robots is a tough crucial task. The undersea world complicated ever-changing, making conventional systems ineffective. When it comes to effective robot navigation, deep learning, more especially Convolutional Neural Networks (CNNs), has emerged as potent technique. This study reviews use convolutional neural networks (CNNs) in context great detail. major difficulties covered, including vision issues, changing illumination, existence objects. article goes on investigate how might help solve these problems boost navigational accuracy. CNNs' capacity extract useful characteristics from unprocessed sensor data like sonar, LiDAR, camera pictures benefit. These capabilities improve robots' perception, allowing them avoid obstacles, identify objects, map their environments. Because flexibility responding novel situations, CNN s well-suited for real world. Additionally, this research explores methods instruction transfer learning that most applicable field navigation. Reduce burden collecting massive amounts harsh settings by applying what been learned simulations real-world circumstances.We also combine CNNs with other fusion Simultaneous Localization Mapping (SLAM) algorithms. synergies overall ability, safely independently traverse unfamiliar sometimes dangerous terrain. describes recent developments case studies which played critical role accomplishing objectives. It demonstrates approaches may enable adaptive intelligent decision-making processes, completely transform robots. In conclusion, developing technology can navigate autonomously. Using CNN-based techniques, shows potential enhancing efficacy, safety, autonomy robotic systems, well possibilities involved implementing such settings. paper's findings aid continuing attempts design smarter, competent, versatile

Language: Английский

---

Energy-Optimization Route and Cluster Head Selection Using M-PSO and GA in Wireless Sensor Networks

Wireless Personal Communications, Journal Year: 2024, Volume and Issue: unknown

Published: May 23, 2024

Language: Английский

---

Crop Yield Prediction Through Artificial Neural Networks

2018 5th IEEE Uttar Pradesh Section International Conference on Electrical, Electronics and Computer Engineering (UPCON), Journal Year: 2023, Volume and Issue: unknown, P. 89 - 94

Published: Dec. 1, 2023

Evolving technologies like Artificial Intelligence, Machine Learning and Data Science are being integrated with almost all sectors. Agriculture, one of the primary sectors is a no exception to this. Regular technological advancements happening in agricultural field ensures food security economical balance country. Variety parameters soil type, rainfall, temperature, market price influence yield crop. Having dependable information about previous crop patterns essential for making assessments on risk management predicting yields. Predicting yields difficult task decision makers. A good forecast model might be adopted by farmers choice what when plant. The increasing population changing environmental conditions impact yield. This paper discusses that predicts Crop Yield through Neural Networks which an important part Learning.

Language: Английский

---

Enhancing reliability in coronary artery disease prediction using ensemble learning for key attribute identification

Life Cycle Reliability and Safety Engineering, Journal Year: 2025, Volume and Issue: unknown

Published: May 20, 2025

Language: Английский

---

Modern 3d Compression Application in Medical Imaging Approach

2018 5th IEEE Uttar Pradesh Section International Conference on Electrical, Electronics and Computer Engineering (UPCON), Journal Year: 2023, Volume and Issue: unknown, P. 1492 - 1496

Published: Dec. 1, 2023

When medical pictures are compressed, valuable information is extracted from the data they contain that crucial for clinicians to understand clinically. The purpose of notion image compression enhance picture content by compressing two images, such as MRI and CT scans, in order provide doctors accurate helpful their clinical care. In this project, photos were combined using Discrete Wavelet Transforms (DWT) approach separate functional anatomical images. compressed has no color alterations both additional spatial attributes information. According experimental data, discrete wavelet transformations highest performance.

Language: Английский

---

SFML: A personalized, efficient, and privacy-preserving collaborative traffic classification architecture based on split learning and mutual learning

Future Generation Computer Systems, Journal Year: 2024, Volume and Issue: 162, P. 107487 - 107487

Published: Aug. 24, 2024

Language: Английский

---

Assessing the Impact of Data Imbalance on the Predictive Performance of Machine Learning Models

Published: Sept. 18, 2024

Language: Английский

---