Browse Items (16488 total)
Sort by:
-
Emerging Cyber Threats in 5G and Beyond: A Wireless Communication Perspective
There are digital age threats to cybersecurity that cost organizations and businesses their infrastructure, operations, and even sensitive data. Cybersecurity risk management is important to ensure that organizational assets are not subjected to cyberattacks, data breaches, and any other vulnerability. This paper also looks at other significant risk-reduction strategies, including threat intelligence, models of risk assessment, encryption, access control systems. It also talks about how machine learning and artificial intelligence can help improve the way threats are detected and handled. Organizations should take a proactive and layered approach to security that integrates leading-edge security methods, with regulatory compliance processes and employee awareness programs. Also, regular security checking, incident response plans and consistent monitoring help in reducing risks and business continuity. Organisations must continue to adjust as cyber threats change, utilizing cutting-edge cybersecurity solutions to fortify their defenses. Organisations may create robust security structures in a continuously linked and threatprone digital environment by using the perspectives this research offers on efficient risk management techniques. 2025 IEEE. -
A Smart Academic Ecosystem Framework for Enhancing Digital Skills and Startup Success Among Potential Women Entrepreneurs
Lack of digital literacy is still a significant impediment for women entrepreneurs to participate and succeed in sectors that rely on technology infrastructure, as mentoring and incubation support is uneven, even for those who are skilled. Conventional academic programs tend to lack well-defined links between structured digital education and entrepreneurial careers, and, as a consequence, gaps between learning outcomes and startup success result. This paper introduces a Smart Academic Ecosystem (SAE) Framework, with tailored and interactive digital adaptation, AI-based mentorship, and dedicated startup incubation, pleading for a standard, one-size-fits-all system. The framework is implemented with a layered architecture comprising data ingestion (from learning management systems), feature stores (learner profiling), and algorithmic modules (knowledge tracing, contextual learning path recommendations, graph-based mentor matching, and venture readiness scoring). Fairness-enabling interventions and privacy-preserving analytics are built into the system to support fairness and trust. A pilot evaluation with early-stage women entrepreneurs identified substantial gains: digital skills mastery increased by more than 20%, startup initiation improved by 12 percentage points, and equity gaps in digital confidence and access were significantly narrowed. Findings emphasize that the SAE model leads not only to faster development of digital capability but also to higher chances of entrepreneurial success. This paper provides a replicable, standards-based, and computer-science-centred framework for academic institutions to encourage women entrepreneurs and innovation ecosystems. 2025 IEEE. -
Design and Analysis of Conformal Antenna for Wearable Devices
This paper presents the design, modeling, and bending analysis of a conformal antenna operating at 2.4 GHz for wearable devices. The antenna's performance was evaluated using HFSS, focusing on return loss, geometry optimization, and bending effects. The optimal resonant frequency of 2.4 GHz was achieved for a patch length (Lp) of 41.78 mm and a width (Wp) of 50.20 mm, achieving a return loss of -34.94 dB and an impedance of 53.5 ohms. The antenna's width had minimal impact on the resonant frequency. When optimized on a Teflon substrate without bending, the antenna demonstrated excellent resonance and impedance matching, with a Voltage Standing Wave Ratio (VSWR) of 1.63 at 2.4 GHz, indicating minimal reflection. The E-plane and H-plane radiation patterns were analyzed at 2.4 GHz, showing a peak gain of 7.38 dB at a theta angle of 0 degrees. Bending analysis revealed that increased bending causes a negative shift in the resonance frequency and affects impedance matching. The proposed antenna is flexible, low-cost, and suitable for wearable medical devices and other applications in the 2.4 GHz frequency band, with return loss, VSWR, radiation pattern, and impedance results all within acceptable ranges. 2025 IEEE. -
Pattern Reconfigurable Antenna Design Using Amc Array for Enhanced 5G Performance
This paper introduces a microstrip patch antenna integrated with a novel artificial magnetic conductor (AMC) to reconfigure the antenna radiation pattern for 5 G sub- 6 GHz applications. By adopting AMC technology, the proposed antenna exhibits beam steering, pattern reconfiguration, and enhancement in gain, suitable for the high data rate demands of 5 G networks. The design process involves the evolution of antenna elements, the design of AMC unit cell (AMC-UC) to operate at the desired 3.75 GHz, and the development of an AMC integrated antenna that is suitable to control the antenna radiation pattern, achieving significant enhancements in signal directionality and minimizing interference. The simulation results demonstrate improved performance parameters, such as return loss and gain, highlighting the potential of AMC-assisted reconfigurable antennas in advancing 5G network coverage and capacity. This work provides valuable information on the achievement of versatile and efficient antenna designs for next-generation wireless communications. 2025 IEEE. -
Compressed Spatio Temporal Graph Neural Networks for Multivariate Time-Series Forecasting
Precise traffic flow forecasting is crucial for efficient transportation management and traffic congestion alleviation. Existing models typically fail to process the intricate spatial- temporal relationships in traffic data and thus incur compromised prediction performance. In this work, we introduce a Compressed Spatial-Temporal Enhanced Graph Neural Network (Comp-STEMGNN) to overcome these limitations. Our model combines 1D convolution-based temporal compression with graph neural networks to compress redundant time-series information without compromising vital patterns. Graph convolutional layers and temporal convolutional blocks extract the spatial and temporal relationships and facilitate efficient learning from enormous sensor networks. Experimental comparisons on benchmark traffic datasets show that Comp-STEMGNN outperforms existing approaches in forecasting accuracy while enjoying substantial computational complexity reduction. These findings identify its potential in real-time traffic forecasting and intelligent transportation systems. 2025 IEEE. -
Application of Advanced Data Mining and Computer Vision Techniques in License Plate Recognition
As urbanization is expanding rapidly and vehicular traffic is on the rise, efficient and automated vehicle identification is a must. Smart transportation, safety monitoring, & police work all heavily rely on Automatic License Plate Detection (ALPD). Traditional heuristic-based image processing techniques are incapable of handling environmental variations; Artificial intelligence (AI) & machine vision solutions are therefore used. YOLO, Faster R-CNN, and SSD are some of the most effective CNN-based and object identification algorithms that demonstrate cutting-edge accuracy in license plate recognition. The paper studies the usage of deep learning algorithms with prepossessing and advanced localization methods for ALPDs' optical character recognition (OCR) and character segmentation. The research also studies integrating ALPD with edge computing and IoTs to develop real-time smart traffic solutions. It also examines machine learning techniques, deep learning innovations, and conventional methods, emphasizing how well various models perform in comparison in terms of accuracy, computational efficiency, and real-time processing power. By employing cutting-edge designs, this study seeks to increase the scalability and resilience of license plate recognition systems, which will support future urban development, security applications, as intelligent transportation management. 2025 IEEE. -
Enhancing Low-Power VLSI Design through AI-Based Simulation and Optimization
AI and ML techniques have dramatically influenced rapid developments in low-power VLSI design with fast advancements in device simulations and power optimization strategies. AI-based simulation tools are now used for accurate modeling of power consumption, improving thermal analysis, and quickening design iterations through the detection of inefficiencies and optimization of energy consumption. In fact, this work focuses on some AI-enabled methods of power reduction techniques such as voltage scaling, clock gating, and leakage current minimization with respect to a sustainable VLSI design. Moreover, a synthetic dataset is created to mimic the actual power consumption trend in VLSI circuits so that predictive modeling and regression techniques can be used for power estimation. Different regression models are used to check the predictive accuracy, and it was found that the highest R2 score was 0.85 by Linear Regression, while the worst was achieved by Decision Tree Regression at 0.50. Results of the correlation analysis and models by machine learning clearly indicate that the frequency and operating voltage are the major contributors to consumption power, while gate counts have a relatively insignificant contribution. Introduction of AI in VLSI simulation enables the enhancement of power efficiency while maintaining sustainability outcomes by optimizing energy usage and cost reduction in terms of computation. 2025 IEEE. -
Reinforcement Q Learning for Terrain-Energy-Aware Lunar Rover Navigation
Effective lunar navigation is difficult in rough terrain and scarce energy resources. Classical path-planning has difficulty with terrain adaptation and energy optimization. This work introduces a Reinforcement Learning (RL)-based solution for energy-optimal lunar rover navigation based on NavCam data from Chandrayaan-3. A Q-learning framework translates terrain characteristics - elevation, slope, and hazards - into a reward scheme, balancing safe travel, minimal energy consumption, and mission effectiveness. The RL agent learns to respond to varying conditions, punishing dangerous regions such as craters and slopes. Simulations on lunar grids demonstrate better energy efficiency and accuracy than traditional approaches. This research pushes autonomous planetary exploration forward, optimizing rover navigation with actual mission imagery for future lunar missions. 2025 IEEE. -
High-Gain Sequentially Rotated LHCP Metasurface Antenna Array for Uplink Ka-Band CubeSat Applications
This paper presents a compact circularly polarized 2 2 microstrip antenna array with a metasurface superstrate designed for Ka-band uplink CubeSat communication applications. The proposed antenna array operates at 28 GHz and consists of four-square patch elements arranged in a sequential rotation, connected through a sequential-phase feed network to achieve stable circular polarization. To enhance gain and axial ratio performance, a 6 6 rotated plus-shaped metasurface layer composed of periodic unit cells is placed above the antenna array.The antenna structure, with overall dimensions of 22 20 6.04 mm3, is designed and simulated using ANSYS HFSS. Simulation results demonstrate an impedance bandwidth of 8.12% (26.74-29.00 GHz) for |S11| < -10 dB and a 3-dB axial ratio bandwidth of 1.52% (27.42-27.84 GHz). The antenna array achieves broadside LHCP achieves a maximum gain of 14.8 dBi at 27.5 GHz with a half-power beamwidth of 9 along the Phi = 90 plane. The inclusion of the metasurface layer results in a gain improvement of approximately 5.2 dBi and with a peak gain of 15.5 dBi and a total efficiency greater than 95%. 2025 IEEE. -
Celestial Image Classification Using Attention And Boosting Mechanism
Astronomical image classification is vital in the comprehension of celestial objects, but deep learning models are severely challenged by the lack of labeled datasets. The novelty of the study is two-fold - the development of the dataset and a hybrid learning method that combines both transformer-based feature extraction and gradient-boosted decision trees to improve classification performance for celestial image classification. This study is a comparison of CNNs, transformers, and hybrid models in nebulae, galaxy, and star cluster classification using the dataset collected from the Hubble Space Telescope image archive. Through progressive data augmentation, the dataset was augmented from 603 images to 4,500 high-diversity training samples to enhance model generalization. This research explores various architectures, including ResNet-50, DenseNet-121, EfficientNetV2-S, DeiT (Data-Efficient Image Transformer), and hybrid models like DeiT-RF (Data-Efficient Image Transformer - Random Forest) and DeiT-XGBoost (DXg). DXg brings a novel fusion mechanism in which DeiT learns high-level spatial representations, adaptive dimensionality reduction fine-tunes feature selection, and XGBoost best classifies celestial objects. Such a unique combination of transformers and gradient boosting improves interpretability without sacrificing state-of-the-art performance. 2025 IEEE. -
Explainable AI for Secure and Trustworthy Autonomous Network Management
Rise of AI-driven autonomous networks for managing complex, dynamic infrastructures. While AI optimizes performance, it acts as a black box. This lack of transparency undermines trust and security, making it challenging to validate decisions, detect adversarial attacks, and understand why an AI model made a specific routing, security, or resource allocation decision. Security blind spots face significant challenges in detecting subtle adversarial manipulations or policy exploits because the reasoning behind the model's decisions is hidden. Additionally, poor diagnosability occurs when a network fault or performance degradation occurs, making root cause analysis slow and complex. Hence, the network operators are hesitant to cede control to systems whose actions they cannot verify or audit. Explainable AI (XAI) is critical for bridging this gap, ensuring management decisions are transparent, interpretable, and defensible. The proposed model makes real-Time management decisions. This model uses post-hoc techniques to generate explanations for each decision. It presents actionable insights and cross-references explanations against security policies and known threat patterns to flag anomalous reasoning. 2025 IEEE. -
Edge Computing and Real-Time Analytics as the Next Frontier for Big Data in IT Companies
Edge computing combined with real-Time analytics is rapidly transforming the way the IT companies can use big data to make smarter and quicker decisions. Centralized model of clouds traditional clouds are being put under stress owing to an explosive increase in data volume, latency sensitive applications as well as bandwidth limitations. Edge computing also makes computation much closer to data sources and allows real-Time analytics that can mitigate latency by orders of magnitude, increase data security and achieve instant insights at a scale. This paradigm shift gives power to IT firms to maximize operations, individualize services and allow agile reactions in a dynamic scenario like smart infrastructure, IoT implementations, and AI-based systems. Edge computing can also be used to offload processing to the edge thus reducing traffic in the core network as well as enabling distributed intelligence. Moreover, edge systems, coupled with AI models, make IT infrastructures perform predictive analytics at the source and become less dependent on backhaul links. This hybridizing process is a paradigm shift in the serious research of big data strategies, and in a future where competitive advantage will rest on latency, context-sensitivity, and localized smarts. 2025 IEEE. -
Machine Learning-Based Intrusion Detection Systems for 5G and beyond Networks
NextGen networks (5 G and beyond) have diversified their infrastructure. Traditional Intrusion Detection Systems (IDS) cannot effectively address the continuously evolving landscape of threats, which is why machine learning-based IDS has emerged as a crucial solution. This overview presents the trends in the application of machine learning techniques (deep learning and ensemble methods) for machine learning-based intrusion detection in 5 G and beyond networks. The important issues tackled encompass real-time anomaly detection, large-scale data processing, adaptive learning against unknown attacks, and detection outcomes. Specifically, we emphasize the promising combination of federated learning, reinforcement learning, and graph-based methods for deployment in distributed, resource-constrained network environments. We present a comprehensive overview of performance metrics such as accuracy, false positive rate, computational overhead, and scalability for each approach, highlighting the crucial trade-offs necessary for successful deployment in dynamic 5G scenarios. Furthermore, we prioritize privacy-preserving methods and secure model sharing. This abstract could further highlight that machine learning-based schemes for intrusion detection systems are important additions toward providing strong defences for cyberspace in 5 G and beyond. 2025 IEEE. -
Federated Learning for Privacy-Preserving Threat Detection in IoT-Enabled Networks
The advent of Internet of Things (IoT) devices at a continuous rapid pace has greatly increased the surface for cyberattacks to measure the effectiveness of threat detection mechanisms. Most conventional centralized threat detection frameworks require sending sensitive device data to a single central server for aggregation, with significant privacy risks and scalability challenges. Such challenges could be efficiently addressed with the use of Federated Learning (FL), an emerging decentralized paradigm of training machine learning models, through the collaboration of a large number of devices, such as IoT sensors, that store the data locally and do not share raw data. In this work, we integrate FL to propose a threat detection framework for preserving privacy in IoT-enabled networks. In this paper, we propose a system architecture in which edge devices perform local training of machine learning models on encrypted traffic and behavioral data and then periodically share only the model updates with a centralized aggregator. This approach ensures the privacy of the data, minimizes communication overhead, and improves detection capabilities for real-time threats. The efficacy of FL-based threat detection is examined through experimental evaluations on benchmark datasets of IoT attack traces, indicating that FL-based approaches achieve competitive accuracy versus prior centralized schemes while greatly mitigating risks of data leakage. We further address issues regarding heterogeneous device resources, communication efficiency, and adversarial attack resilience in this context. Our results indicate that federated learning is a very effective approach for providing IoT environment protection, as it securely balances privacy, scalability, and detection performance. 2025 IEEE. -
AI-Powered Analytics in IT Services and the Opportunities and Challenges for Scalable Growth
The technological advancement has witnessed high levels of efficiency in operations, decision-making and customer engagement because of the integration of AI enabled analytics into IT-based services. The paper examines the influence of AI-based solutions that are transforming IT service models, with focus on the effect on scalability, predictive maintenance, and intelligent automation. The studies to point out the opportunities presented by AI describe improved personalization of service delivery, fewer instances of downtime, and data-based optimization approaches. Nonetheless, it also addresses such vital issues as the privacy or interpretability of data and the infrastructural requirements of scaling AI solutions. It is suggested to replace these limitations with a hybrid framework resolving limitations by integrating the advantages of cloud-native frameworks with edge-intelligent systems. Experimental study conducted among medium sized IT companies revealed that the speed in delivering the services increased by 5 0%, 6 0% lower error rates and over 55% reduction in downtime. The results hint at the possibility of bearing AI-driven analytics-based IT services when the required control and strategic design are employed. 2025 IEEE. -
Deep Reinforcement Learning for Dynamic Resource Allocation in Intelligent Communication Networks
As intelligent communication implementations like 5G and IoT-enabled infrastructure meet technological advancement, provisioned network resource allocation dynamically and efficiently will be deemed crucial to accommodate diverse service demands and guarantee an optimized part of the network on demand. Resource management strategies based on traditional approaches often lack a sufficient response to the dynamic posture of network states and complex, heterogeneous environments. To address these challenges, deep reinforcement learning (DRL) has emerged as a powerful methodology wherein deep neural networks are employed to enable intelligent and adaptive decision-making based on dynamic network conditions. In this paper, we study the potential of DRL for dynamic resource management in intelligent communication networks. We build a DRL-driven agent that enables optimal allocation policy learning by interacting with high-dimensional, stochastic network environments with variable traffic loads, user mobility, and heterogeneous quality-of-service (QoS) requirements. Realistic simulation scenarios show that the proposed DRL framework outperforms conventional allocation heuristics in terms of throughput, latency, and fairness among users. We elaborate on the ramifications of explorationexploitation tradeoffs, convergence stability, and compute efficiency in the context of scale deployments. This way, our results prove that DRL is a potential candidate for dynamic resource allocation in future intelligent communication networks due to its better adaptability and performance. 2025 IEEE. -
Predictive Analytics in Wealth Management and the Role of Machine Learning for Investment Professionals
The paper focuses on the use of predictive analytics and machine learning (ML) to potentially transform the landscape of contemporary wealth management and provide financial decision-makers with cutting-edge tools that improve decision-making and portfolio management practices with clients. Conventional investment models are usually based on the past performance and unchanging risk evaluation, which is not flexible in a fluctuating market. As compared, the suggested ML-based framework would use the real-time data source dynamic, that is, market trends, behavioral financial indicators, and macroeconomic signals to capture the real-time forecast and provide individual investment advice. Multi-model ensemble comprising Random Forest, XGBoost, and LSTM networks were created to predict the asset performance and to evaluate investor risk profiles with high accuracy. Empirical analysis proved that the proposed system proved to be more accurate (95.3 %), have a higher precision (92 %), recall (94.1 %), and F1-score (92.1 %) as compared to current approaches. The strategy improves risk-adjusted returns, minimises human bias and decision lag. These results reinforced the useful application of ML in enhancing the strategic potential of investment professionals and establishing a more robust and flexible ecosystem of managing wealth. 2025 IEEE. -
Artificial Intelligence-Powered Stock Market Forecasting with Metaheuristic Feature Selection Techniques
This study proposed a hybrid stock market forecasting model which consists of Artificial Intelligence (AI) and metaheuristic feature selection algorithms to improve the accuracy in prediction and efficiency of the prototypical. It uses PSO (Particle Swarm Optimization) algorithm to pick the most relevant feature out of a pool of technical indicators and sentiment data and temporally learns the pattern using the LSTM (Long Short-Term Memory) network. The model yields better learning by diminishing noise and dimensionality and prevents over fitting. The efficiency of the anticipated system is seen through comparative analysis with such baseline models as SVM (Support Vector Machine), RF (Random Forest), and standard LSTM. This prototypical obtained MAE of 11.2RMSE of 18.18, and the mean absolute percentage error (MAPE) of 5.36 percent, with R2 of 0.91 and directional accuracy of 86.4 percent. The above results confirm the effectiveness of the suggested method, providing a solid and generalizable solution in terms of intelligent stock market prediction and investment decision support. 2025 IEEE. -
Enhancing Investment Advisory with Machine Learning for a New Era in Financial Services
The current financial service environment, where the volatility of markets and the need to offer flexible solutions is growing, is starting to challenge the traditional investment advisory models. This paper implements a new framework, which incorporates the most advanced methods of machine learning, to make investment advising a process driven by real data. This is unlike the current models which are overly dependent on historical trends or fixed risk profiles, our system allows us to use real time behavior analytics, sentiment analysis and dynamic portfolio optimization to give hyper personalized investment recommendations. The framework feeds the ensemble learning, attention-based neural networks, explainable AI (XAI) to make sure the transparency, regulatory, and investor trust. The innovation in particular is based on the constant interaction between client and adjustment of the model in terms of a ready and sensitive advisory intervention. The study will not only improve the relevance and precision of financial advice, but will with its informed automation of advisor-client relationship led to a redefinition of the advisor-client relationship. The insights guide to a world of advisory services where ML and machine learning complement strategic decision-making with unheard levels of specificity and individuality. 2025 IEEE. -
Machine Learning in Investment Analysis-Enhancing or Replacing Human Judgment
Machine Learning (ML) involvement in investment analysis is quickly revolutionizing the investment-based decisions through becoming highly accurate, quick, and embracing increased data processing capabilities. This paper is to research on whether ML is complementary or a possible replacement to human financial judgment. We run experiments over 1.2 million financial transactions between 150 firms comparing old style analyst recommendations and ML-models, including XGBoost, LSTM and Random Forest. The findings indicate that ML models outperformed prediction capability by 19.6 percent and lowered the volatility of the portfolios by 14.3 percent in 5-year investment. Also, the ML-Aided decision-making was better than human (only) approaches in 78 percent of the cases in markets with high volatility or that involved trading in complicated assets. The qualitative variables like regulatory policy changes and investor sentiment however were too difficult to decipher under the leadership of ML only. Our results indicate that ML supports rather than supers the human judgement and thus demonstrates a hybrid paradigm of decision making that resolves computational exactitude with context sensitive understanding in the modern investment scenarios. 2025 IEEE.