The worldwide cause of mortality is cardiovascular heart disease. The automatic prediction of heart disease can be made to possible for accurate detection in initial stage. In recent year, the artificial intelligence approaches giving promising outcomes in predicting various types of cardiovascular conditions. The main focous of this work is to implementation of various machine learning techniques used to predict cardiovascular heart disease (CHD) using electrocardiogram (ECG) datasets. ECG provide the electrical Signal from the heart that identify the presence of disease or not. The preprocessing method are used for improving the quality of ECG signals and extract the features from ECG of patients. There are several well-established machine learning techniques, including support vector machine (SVM) and K-nearest neighbour (KNN)., logistic regression and decision tree classifier used for prediction of the disease. So, our finding of this paper will provide the new understanding regarding CHD prediction using different machine learning techniques. The Decision Tree-based machine learning model demonstrated excellent performance, achieving 98% accuracy, 96% precision, 100% recall, and an F1-score of 97%, which is better than rest of other comparative machine learning models. Finaly expermental results shows that decision tree approach providing better outcome amongs all the algorithms with respect to all above mensioned parameter.

Stock price prediction is a branch of financial forecasting that provides valuable insights for investors, traders, financial institutions, and other stakeholders. This study examines the performance of three deep learning algorithms—LSTM, GRU, and SimpleRNN—in forecasting stock prices based on past data. In contrast to research that emphasizes global efficiencies, this study assesses model performance according to dataset characteristics. The dataset, which includes daily stock prices from 2009 to 2019, focused on closing prices. Data preprocessing included Min-Max Scaling and sliding window methods. Hyperparameters such as batch size, learning rate, dropout rate, and L2 norm strength were modified, with the training and test sets making up 70

Early and accurate detection of skin malignancies is critical for improving patient outcomes. However, deep learning (DL) models often fail to generalize because of variations in imaging protocols, differences in lesion appearance, and the uneven distribution of clinical data across institutions. Although federated learning (FL) enables collaborative model training without sharing raw patient data, its practical deployment is hindered by non-identically distributed (non-IID) data and inconsistent participation from slow or unreliable clients (laggards). We propose RePFL, a lag-resilient personalized federated learning (PFL) framework for skin cancer classification. RePFL integrates late-stage feature embedding alignment to reduce representation divergence while preserving client-specific personalization, a computationally efficient stochastic trace-based Fisher Information Matrix (t-FIM) score to estimate the contribution of each client without constructing the full Fisher matrix, and semi-asynchronous staleness-aware aggregation to incorporate delayed client updates while reducing the impact of lagging clients. Experiments on the HAM10000 dataset demonstrate that RePFL achieves classification accuracies of 87.50% and 86.86% under two non-iid settings, outperforming established FL and PFL baselines. The framework preserves raw-data locality but does not provide formal privacy guarantees, and its explanation maps are evaluated using computational faithfulness and stability metrics rather than clinical validation. Overall, RePFL provides a practical and robust framework for multi-institutional collaborative learning in resource-constrained healthcare environments.

The rapid scaling of microprocessor technologies has led to unprecedented advancements in processing speed and transistor density. However, as device dimensions shrink below the nanometre scale, traditional electrical on-chip interconnects face critical limitations including increased signal delay, power consumption, and reduced bandwidth. These challenges have emerged as major bottlenecks in the performance and energy efficiency of modern microprocessors, especially in multi-core and many-core architectures. To overcome these limitations, optical interconnects have gained significant attention as a high-performance alternative for on-chip and chip-to-chip communication. Optical interconnects offer key advantages such as higher data rates, lower latency, reduced power dissipation, and immunity to electromagnetic interference. The integration of optical components like waveguides, modulators, photodetectors, and silicon photonics within the chip architecture has the potential to revolutionize interconnect design by addressing the shortcomings of conventional electrical wiring. This paper explores the evolution of on-chip interconnects, comparing the performance, scalability, and integration challenges of electrical and optical solutions in scaled microprocessors. It also discusses emerging hybrid interconnect architectures that combine the strengths of both technologies. A comparison is made between electrical and optical interconnects for different design criteria, based on projections about the future of optical devices. Around a tenth of the length of the chip's edge is the crucial dimension beyond which optical connectivity becomes preferable to electrical interconnect at the 22 nm technological node.

The rapid growth of medical imaging data presents significant challenges in diagnostic accuracy, data privacy, and computational efficiency. Traditional centralized AI models struggle with scalability and pose risks to patient confidentiality due to data aggregation. Moreover, heterogeneous medical data across institutions complicates the development of robust diagnostic tools. To address these issues, we propose MID-large language model (LLM), a novel framework that integrates LLMs with a blockchain-based federated learning (FL) system for medical image analysis. It also ensures the security and privacy of sensitive medical data across decentralized networks. MID-LLM uses verification mechanisms to ensure the global model's integrity. It also employs aggregation techniques to reduce bias and improve training efficiency. Experiments on the BraTS 2020 dataset show that MID-LLM outperforms traditional FL, achieving higher Dice scores with improved computational efficiency. These results highlight MID-LLM's potential to enhance diagnostic accuracy while offering a scalable, secure solution for AI in healthcare.

The secure and efficient storage and sharing of medical images have become increasingly important due to rising security threats and performance limitations in existing healthcare systems. Centralized systems struggle to provide adequate privacy, rapid access, and reliable storage for sensitive medical images. This paper proposes a decentralized medical image-sharing framework to address these issues by integrating blockchain technology, the InterPlanetary File System (IPFS), and edge computing. Blockchain technology enforces secure patient-centric access control through smart contracts that enable patients to directly manage their data-sharing permissions. The IPFS provides decentralized and scalable storage for medical images and effectively resolves the storage limitations associated with blockchain. Edge computing enhances system responsiveness by significantly reducing latency through local data processing to ensure timely medical image access. Robust security is ensured by using elliptic curve cryptography (ECC) for secure key management and the Advanced Encryption Standard (AES) for encrypting medical images to protect against unauthorized access and data breaches. Additionally, the system includes real-time monitoring to promptly detect and respond to unauthorized access attempts to ensure continuous protection against potential security threats. System results demonstrate that the proposed framework achieves lower latency, higher throughput, and improved security compared to traditional centralized storage solutions, which makes our system suitable for practical deployment in modern healthcare settings.

Improved data sharing between healthcare providers can lead to a higher probability of accurate diagnosis, more effective treatments, and enhanced capabilities of healthcare organizations. One critical area of focus is brain tumor segmentation, a complex task due to the heterogeneous appearance, irregular shape, and variable location of tumors. Accurate segmentation is essential for proper diagnosis and effective treatment planning, yet current techniques often fall short due to these complexities. However, the sensitive nature of health data often prohibits its sharing. Moreover, the healthcare industry faces significant issues, including preserving the privacy of the model and instilling trust in the model. This paper proposes a framework to address these privacy and trust issues by introducing a mechanism for training the global model using federated learning and sharing the encrypted learned parameters via a permissioned blockchain. The blockchain-federated learning algorithm we designed aggregates gradients in the permissioned blockchain to decentralize the global model, while the introduced masking approach retains the privacy of the model parameters. Unlike traditional raw data sharing, this approach enables hospitals or medical research centers to contribute to a globally learned model, thereby enhancing the performance of the central model for all participating medical entities. As a result, the global model can learn about several specific diseases and benefit each contributor with new disease diagnosis tasks, leading to improved treatment options. The proposed algorithm ensures the quality of model data when aggregating the local model, using an asynchronous federated learning procedure to evaluate the shared model’s quality. The experimental results demonstrate the efficacy of the proposed scheme for the critical and challenging task of brain tumor segmentation. Specifically, our method achieved a 1.99% improvement in Dice similarity coefficient for enhancing tumors and a 19.08% reduction in Hausdorff distance for whole tumors compared to the baseline methods, highlighting the significant advancement in segmentation performance and reliability.

The motion of an autonomous ship is different from that of ground and aerial robots due to its maneuvering and environmental constraints. As a result, many techniques have been introduced for autonomous ship path planning. This paper presents a novel technique for global and local navigation planning of autonomous ships under complex static and dynamic constraints. Our technique, termed safety-enhanced path planning (SPP), has been developed to avoid potential collisions with underwater obstacles near seaside areas. SPP pre-processes the map to preserve the shape of visible obstacles and mark a safety-outline around the shores. Subsequently, an offset safety line (OSL) is drawn about the original shore to protect the ship when passing close to threat-defined offshore areas. The global path is produced with an enhanced A* multi-directional algorithm, considering the kinematic constraint of the ship. To ensure optimal path quality, the global path is further refined with a smoothing filter to improve consistency and smoothness. Additionally, local navigation is introduced to help the autonomous ship avoid collisions with other obstacle ships. Local offset trajectories are produced with 4th and 5th degree polynomials along longitudinal and lateral coordinates in time t. Distance closest point approach (DCPA) is utilized for early obstacle prediction to help the ship maneuver in complex dynamic obstacle avoidance scenarios. The trajectory set is filtered with an efficient cost policy to obtain the best trajectory for dynamic collision avoidance. We conduct simulations in MATLAB and compared with other maritime path planning methods to verify the effectiveness of our approach.

Biomedical image analysis plays a crucial role in enabling high-performing imaging and various clinical applications. For the proper diagnosis of blood diseases related to red blood cells, red blood cells must be accurately identified and categorized. Manual analysis is time-consuming and prone to mistakes. Analyzing multi-label samples, which contain clusters of cells, is challenging due to difficulties in separating individual cells, such as touching or overlapping cells. High-performance biomedical imaging and several medical applications are made possible by advanced biosensors. We develop an intelligent neural network model that can automatically identify and categorize red blood cells from microscopic medical images using region-based convolutional neural networks (RCNN) and cutting-edge biosensors. Our model successfully navigates obstacles like touching or overlapping cells and accurately recognizes various blood structures. Additionally, we utilized data augmentation as a pre-processing method on microscopic images to enhance the model's computational efficiency and expand the sample size. To refine the data and eliminate noise from the dataset, we utilized the Radial Gradient Index filtering algorithm for imaging data equalization. We exhibit improved detection accuracy and a reduced model loss rate when using medical imagery datasets to apply our proposed model in comparison to existing ResNet and GoogleNet models. Our model precisely detected red blood cells in a collection of medical images with 99% training accuracy and 91.21% testing accuracy. Our proposed model outperformed earlier models like ResNet-50 and GoogleNet by 10-15%. Our results demonstrated that Artificial intelligence (AI)-assisted automated red blood cell detection has the potential to revolutionize and speed up blood cell analysis, minimizing human error and enabling early illness diagnosis.

Background: Infectious diseases like COVID-19 pose major global health threats. Robust surveillance systems are needed to swiftly detect and contain outbreaks. This study investigates the integration of Blockchain technology and machine learning to establish a secure and ethically sound approach to tracking infectious diseases. Methods: We established a Blockchain-based framework for the collection and analysis of epidemiological data while upholding privacy standards. We employed encryption and privacy -enhancing technologies to gather information on case numbers, locations, and disease progression. Artificial neural networks were employed to scrutinize the data and pinpoint transmission patterns. A prototype was specifically designed to work with COVID-19 data from specific countries. Results: The Blockchain system enabled reliable and tamper -proof data gathering with enhanced transparency. The evaluation showed it allowed cost-effective tracking of infectious diseases while upholding confidentiality safeguards. The neural networks effectively modeled disease spread based on the Blockchain data. Conclusions: This research demonstrates the viability of Blockchain and machine learning for infectious disease surveillance. The system strikes a balance between public health concerns and personal privacy considerations. It also addresses the challenges of misinformation and accountability gaps during disease outbreaks. Ongoing development can lay the foundation for an ethical framework for digital disease tracking, ensuring both pandemic preparedness and response capabilities are upheld. (c) 2024 AGBM. Published by Elsevier Masson SAS. All rights reserved.

This paper presents a comprehensive study of 3D point cloud Federated Few-Shot Learning (3DFFL), focusing on addressing challenges such as limited data availability and privacy concerns in point cloud classification for applications such as autonomous vehicles. We introduce a novel approach that integrates Federated Learning with Few-Shot Learning techniques, with a special emphasis on optimizing network architectures for 3D point cloud data. Our method capitalizes on the strengths of PointNet++ for feature extraction and ProtoNet for classification, all within a federated learning framework to ensure data privacy and collaborative learning. Significantly, the approach is augmented with the use of attention and SoftMax layers, enhancing the feature extraction and classification processes. Extensive experiments on the ModelNet40, ShapeNet, and ScanOnjectNN datasets validate our method's accuracy and adaptability in handling 3D point cloud classification, especially in privacy-sensitive and collaborative scenarios. This study not only demonstrates the potential of integrating attention mechanisms and SoftMax layers in 3DFFL but also lays a robust foundation for future advancements in this evolving field, particularly in technologies dependent on 3D data processing.

LPWAN has partially replaced traditional wired networks in fields such as smart industry, smart healthcare, smart home, etc., due to its low power consumption, high reliability and low cost. LPWAN can achieve long-distance and low-power data transmission without increasing bandwidth, thus meeting the energy efficiency, cost-effectiveness and security requirements of IoT devices. However, low-power IoT also faces some challenges due to design limitations. For example, the reliability of connection under harsh environment and communication interference conditions, and ensuring the long life of devices. To solve this problem, in addition to improving hardware aspects, we also seek to use machine learning methods to make devices run under highly intelligent scheduling conditions, so as to optimize device connection reliability and energy utilization. To this end, this paper proposes a data acquisition, denoising, prediction and transmission optimization method for low-power sensor networks. First, by collecting sound data, video data and light data using temporal flow for modality alignment we achieve data denoising and prediction. Second, by predicting the transmission efficiency of sensors under different temperature humidity and illumination conditions we dynamically adjust sensor power and bandwidth according to transmission loss changes to maximize data transmission efficiency. Finally, we deployed the multimodal Transformer method on edge sensor nodes, combining the data transmitted from image, temperature, and humidity sensors. This approach improved the reliability of data transmission for the sensor devices. Experimental results show that compared with existing methods our proposed method has significant improvement in delay, wavelet denoising efficiency, packet delivery ratio and transmission efficiency.

Assessing the importance of spreaders in networks is crucial for investigating the survival and robustness of networks. There are numerous potential applications, such as preventing outbreaks, spreading viruses on computer networks, viral marketing, and sickness spreading. These problems are usually unable to be solved by many heuristics with low time complexity. A number of tentative heuristics have been proposed for different application scenarios. However, we still lack an efficient heuristic to solve this type of problem effectively, for instance, low rating accuracy or high time complexity. To deal with this issue, this paper proposes a new heuristic called Locality-based Structure System (LSS), which is based on local information rather than global information of nodes in a network to determine the importance of spreaders. The proposed LSS takes into account the k-shell, degree, and number of triangles in a network. First, connectivity factors are computed based on the properties of nodes connected to them. Then, each node's contribution to the importance of lines is computed. Finally, the degree and k-shell of nodes, as well as their contribution to the importance of lines, are taken into account. The proposed LSS is validated on a set of real and synthetic complex networks, where the simulation results under the standard SIR model and Kendall correlation coefficient indicate that LSS can efficiently identify influential spreaders in numerous types of networks without requiring any advanced parameter settings.

Addressee detection (AD) enables robots to interact smoothly with a human by distinguishing whether it is being addressed. However, this has not been widely explored. The few studies that have explored this area focused on a human-to-human or human-to-robot conversation confined inside a meeting room using gaze and utterance. These works used statistical and rule-based approaches, which tend to depend on specific settings. Further, they did not fully leverage the available audio and visual information or the short-term and long-term segments, and they have not explored combining important conversation cues—the facial and audio features. In addition, no audiovisual spatiotemporal annotated dataset captured in mixed human-to-human and human-to-robot settings is available to support exploring the area using new approaches.

Purpose The majority of machine component failures are caused by load conditions that change with time. Under those circumstances, the component can function effectively for a long time but then breaks down unexpectedly and without warning. Therefore, the study of fatigue considerations in design becomes important. Also, to determine the component's long-term tenability, fatigue behavior must be investigated. This paper aims to investigate the fatigue life of aluminum 6061-T6 alloy under uniaxial loading using experiments and finite element simulation. Design/methodology/approach Both base metal (BM) and friction stir welding (FSW) configurations have been used to analyze fatigue behavior. The experimental tests were carried out using Instron-8801 hydraulic fatigue testing machine at frequency of 20 Hz and load ratio of 0.1. The yield strength, ultimate tensile strength, amplitude stress and fatigue life were used as input in simulation analysis software. Based on the findings of the tensile test, the maximum stress applied during the fatigue testing was estimated. Simulated and experimental results were also used to plot and validate the S-N curves. The fracture behavior of specimens was also examined using fractographic analysis. Findings The fractured surfaces indicate both brittle and ductile failure in the specimens. However, dimples dominated during the final fracture. The comparison between experimental and simulation results illustrates that the difference in fatigue cycles increases with an increase in the yield strength of both BM and FSWed specimens. This disparity is attributed to many factors such as scratches, rough surfaces and microstructural behavior. Aluminum 6061-T6 alloy is considered a noteworthy material where high strength with reduced weight contributes to the crash-worthy design of automobile structures. Originality/value The current study is significant in the prediction of the fatigue life of aluminum 6061-T6 alloy using experiments and simulation analysis. A good correlation was found when the experimental and simulation analysis were compared. The proposed simulation analysis approach can be used to anticipate a component's fatigue life.

This paper presents a spider web-shaped antenna for near-field UHF RFID applications. The proposed design consists of concentric decagons, and open-ended microstrip lines etched on the circular substrate. The proposed design features a uniform and fairly strong electric field distribution in the nearfield region. In addition, this antenna exhibits impedance matching ranging from 902 MHz - 925 MHz and further poses low gain characteristics, that is required for most of the nearfield application in order to avoid misreading of other tags. Moreover, this design provides symmetric current distribution throughout the structure, thereby solves orientation sensitivity problems of low-cost linearly polarised tag antennas. The measurement results demonstrate tag can successfully read expensive jewellery items, tagged pills and tag placed in different orientations. Therefore, the proposed reader antenna is a good candidate for near field RFID, healthcare and IoT Applications.

Stock market forecasting has drawn interest from both economists and computer scientists as a classic yet difficult topic. With the objective of constructing an effective prediction model, both linear and machine learning tools have been investigated for the past couple of decades. In recent years, recurrent neural networks (RNNs) have been observed to perform well on tasks involving sequence-based data in many research domains. With this motivation, we investigated the performance of long-short term memory (LSTM) and gated recurrent units (GRU) and their combination with the attention mechanism; LSTM + Attention, GRU + Attention, and LSTM + GRU + Attention. The methods were evaluated with stock data from three different stock indices: the KSE 100 index, the DSE 30 index, and the BSE Sensex. The results were compared to other machine learning models such as support vector regression, random forest, and k-nearest neighbor. The best results for the three datasets were obtained by the RNN-based models combined with the attention mechanism. The performances of the RNN and attention-based models are higher and would be more effective for applications in the business industry.

Online clustering of short text streams has become significant due to the popularity of news and social media platforms. The objective of online clustering is to maintain active topics (clusters) by automatically detecting new topics and forgetting outdated ones. Most existing approaches exploit static and high dimensional semantic term representation of the text to enhance the clustering quality. While these approaches use inference procedures that depend on a fixed batch size to reduce the number of clusters related to a given topic and bring it closer to the actual number of topics. This paper proposes a non-parametric Dirichlet model with episodic inference (EINDM) to cluster the evolving short text stream by introducing a window-based low-dimensional semantic term representation which captures the contextual relationships between words. In addition, an episodic inference procedure is introduced to reduce the cluster sparsity in the model. Furthermore, a novel "word specificity" measure is proposed based on neighborhood terms for evolving contexts for individual terms. Extensive empirical evaluation demonstrates that EINDM yields the best performance, in terms of NMI, homogeneity, and cluster purity, compared to recent state-of-the-art clustering models.

Vehicle re-identification is one of the essential application of urban surveillance. Due to enormous variation in inter-class and intra-class resemblance creates a challenge for methods to distinguish between the same vehicles. Additionally, varying illumination and complex environments create significant hurdles for the existing methods to re-identify vehicles. We present a multi-guided learning method in this paper that uses multi-attribute and view point information, while also enhancing the robustness of feature extraction. The multi-attribute sub-network learns discriminative features like, i.e. color and type of vehicle. Moreover, the view predictor network adds extra information to the feature embedding and To validate the effectiveness of our framework, experiments on two benchmark datasets VeRi-776 and VehicleID are conducted. Experimental results illustrate our framework achieved comparative performance.

This article presents a bio-inspired circularly polarized ultrahigh-frequency (UHF) radio frequency identification (RFID) tag antenna for metallic and low-permittivity substances. This tag design is based on a leaf-shaped radiator, two shorting stubs, and slots etched on F4B substrate. Initially, the tag antenna is designed using a characteristics mode analysis (CMA) by analyzing the first six characteristic modes (CM) and characteristic angles. The width of orthogonal slots is varied to get the resonance of CM modes in the required US RFID band. Moreover, the edges are blended to get orthogonal current distribution, which is necessary for circular polarization. In addition, the proposed tag design is optimized further using CST Microwave studio and an RFID chip is exploited as a capacitive coupling element (CCE) to run CM modes with the orthogonal current pattern. This tag can also be tunable to European RFID (EU) band (866–868 MHz) by changing the length of shorter diagonal slot. The tag design offers a read range of 7–4.5 m on $100\times100$ mm2 metals plate and low-permittivity substrates (for the 902–928-MHz band). In the EU band, the corresponding read ranges are 5.7 and 3.5 m above metal and low-permittivity objects, respectively. This circularly polarized tag antenna is advantageous in terms of cost, circular polarization feature, and ease of fabrication due to the absence of vias, shorting pins, and matching circuits. Therefore, this tag design is suitable for labeling various low-permittivity objects, industrial conveyer belt applications, baggage handling systems, and Internet of Things (IoT) applications.