The following correction pertains to [1]. Please update your copy of the Magazine accordingly. When working with a truncated Fourier expansion, the corrected form of Equation (6) is as follows: x M [n] = M 2 k=− M 2 c k e i 2πk N n , M < N. (1) This correction applies to the Equations (6), (7), (9), and (15) as well. Please update these equations in your copy accordingly.

Recently, Doppler tracking based positioning, navigation, and timing (PNT) with low Earth orbit (LEO) satellites has attracted considerable attention. However, as LEO satellites are not dedicated for PNT, we cannot obtain the crucial information, such as signals structure and clock synchronization from the LEO signals of opportunity (SOPs), bringing challenges to Doppler LEO-PNT. Moreover, Doppler signal acquisition is time-consuming, which implies that real-time target localization is not allowed. In this paper, LEO-PNT is studied from passive source localization point of view. A reduced-dimension MUSIC (RD-MUSIC) with uniform circular array (UCA) is proposed for target localization using LEO SOPs. UCA is adopted because it offers uniform coverage in all directions, resulting in highly effective signal detection from any angle-of-arrival. Additionally, both odd and even numbers of antenna elements are investigated. The devised method employs one-dimensional (1-D) search, and then establishes equations from which the target's coordinates can be solved in a computationally attractive manner. The efficiency and reliability of our RD-MUSIC are demonstrated via numerical simulations, where the data are generated by the systems tool kit (STK) of Analytical Graphics.

In order to improve the efficiency of automated warehouse robots in picking and packaging operations and enhance the control precision of their motion execution, this study incorporated machine learning algorithms into the design of the control system of automated warehouse robots. Mechanical learning and inductive learning were used to design machine learning methods applicable to these robots. Meanwhile, the learning process was optimized with the help of PID feedback adjustment, thus constructing an efficient execution system for the picking and packaging operations of automated warehouse robots. Taking the quantity of goods picked and packaged within the same period of time as the task indicator for the operations, simulation was carried out on the operation situations of robots adopting different machine learning methods. The results show that by using the combination of mechanical and inductive learning methods and the PID feedback adjustment approach, automated warehouse robots can handle more goods within the same period of time, significantly improving their motion execution efficiency in picking and packaging operations and providing strong support for the efficient operation of automated warehouses.

Real-time decision-making is a strength for industries where decision-making in the fast-changing environment has influenced decision-making due to advancing technologies. The current paper outlines a conceptual framework that leverages data science to improve real-time decision-making in flexible and expandable systems. The architecture combines machine learning and streaming analytics for high-velocity data streams, which provide business decisions in milliseconds. As a distributed cloud-based system, the proposed system guarantees high scalability and flexibility since the predictive models are updated in real time to adjust to the new data pattern. The proposed framework outperforms other multiple models by using the decision tree, deep learning, and ensembling; accuracy, speed, and resources results are more persuasive. Actual implementations substantiate that the system’s dynamic aspects increase the accuracy of the premeditated choices by 25% when compared to the normal unelastic models and, at the same time, slashes the time delay by 40%. The result of this research gives real-life insight into the effective integration of real-time data science into BI systems, making it perfect for building strong BI systems in the future. The paper’s results enrich the current discussion on real-time analytics by presenting a clear guideline for building larger-scale decision-making frameworks for domains that include finance, healthcare, and IoT-based industries.

Classifying child speech is challenging due to the limited availability of large, publicly accessible datasets recorded in consistent environments. This study aimed to develop a classifier to differentiate child and adult speech using datasets such as LibriSpeech, VoxCeleb, Common Voice, OGI Kids, CMU Kids, and MyST. Initial attempts using diarization and embeddings from ECAPA-TDNN, Whisper, and HuBERT led to misclassification due to dataset-specific clustering caused by recording artifacts. To address this, data augmentation with impulse responses was applied, improving generalization. A fine-tuned gradient boosting classifier trained on ECAPA-TDNN embeddings achieved 93.5% accuracy. Future efforts will explore enhanced augmentation and model tuning for further improvements.

WordPress, originally developed as a blogging platform, has grown into a powerful content management system (CMS) that underpins over 40% of websites worldwide. Its versatility, accessibility, and extensibility have made it an indispensable tool for software developers. This paper examines the critical role WordPress plays in software development, highlighting its contributions to rapid prototyping, scalability, integration with modern technologies, and democratizing software creation. By analyzing real-world case studies and exploring the extensive ecosystem of plugins, themes, and APIs, this paper underscores how WordPress fosters innovation and accelerates development timelines.

Global e-commerce has transformed international trade, enabling businesses and consumers to connect across geographical and cultural divides. However, cultural and economic disparities often hinder its full potential. This paper explores the role of artificial intelligence (AI) in overcoming these barriers. It examines AI-driven personalization, localization, and accessibility strategies that address language differences, economic constraints, and cultural nuances. By analyzing case studies and real-world implementations-including the innovative strategies of FoodKravings.com-this study highlights the transformative potential of AI in creating a more inclusive global e-commerce landscape. The findings underscore the importance of leveraging AI to foster trust, enhance user experiences, and drive sustainable growth in emerging markets.

The anonymization of medical texts in Spanish is essential for safeguarding patient privacy and ensuring compliance with regulations such as the General Data Protection Regulation (GDPR). This study explores the application of Large Language Models (LLMs) to automate the de-identification of Protected Health Information (PHI) in clinical texts, using foundational models like Gemma-2 (9B and 2B) and Llama (3.1 8B and 3.2 3B). Two benchmark datasets, DisMed and MEDDOCAN, were used to evaluate performance, providing a diverse range of clinical narratives and structured records for testing the models' efficacy. The methodology employed few-shot learning to guide LLMs in recognizing and tagging PHI entities such as names, dates, locations, and numerical identifiers. Custom evaluation metrics, including overlap-based and similarity-based methods, were introduced to address the limitations of traditional exact-match metrics, enabling a more nuanced analysis of model outputs. Post-processing steps ensured accurate and non-overlapping entity annotations to refine the anonymization process. Results showed that larger models, such as Gemma-2 9B and Llama 3.1 8B, achieved high F1 scores across most PHI categories, with particularly strong performance in standardized entities like addresses and locations. However, challenges remained in identifying diverse formats of dates and numerical data. The models demonstrated greater adaptability compared to traditional approaches, but occasional generation of hallucinated entities highlighted areas for improvement. Dataset quality and diversity significantly influenced performance, underscoring the importance of robust training data. The study concludes that LLMs are effective in automating the de-identification of Spanish medical texts, meeting the objective of providing a reliable and efficient anonymization solution. While computational costs and variability in certain outputs pose challenges, this work lays a strong foundation for future research to enhance the scalability and precision of these systems, facilitating secure data sharing and compliance in medical research.

Transfer learning, successful in knowledge translation across related tasks, faces a substantial privacy threat from membership inference attacks (MIAs). These attacks, despite posing significant risk to ML model's training data, remain limitedexplored in transfer learning. The interaction between teacher and student models in transfer learning has not been thoroughly explored in MIAs, potentially resulting in an under-examined aspect of privacy vulnerabilities within transfer learning. In this paper, we propose a new MIA vector against transfer learning, to determine whether a specific data point was used to train the teacher model while only accessing the student model in a whitebox setting. Our method delves into the intricate relationship between teacher and student models, analyzing the discrepancies in hidden layer representations between the student model and its shadow counterpart. These identified differences are then adeptly utilized to refine the shadow model's training process and to inform membership inference decisions effectively. Our method, evaluated across four datasets in diverse transfer learning tasks, reveals that even when an attacker only has access to the student model, the teacher model's training data remains susceptible to MIAs. We believe our work unveils the unexplored risk of membership inference in transfer learning.

This review explores the burgeoning field of virtual and augmented reality (VR/AR) applications for mental health, focusing on the integration of conversational artificial intelligence (AI) with animated photorealistic avatars created from photographs. This approach facilitates the development of personalized, accessible, and engaging therapeutic experiences across multiple platforms, including mobile phones and VR headsets. By combining the immersive qualities of VR/AR with the interactive capabilities of AI, we present a novel solution to addressing mental health needs, while also expanding to elderly care applications through computer vision technology. Special consideration is given to the potential for adoption in the Japanese market, which is characterized by a rapidly aging population and a growing interest in technological solutions for healthcare in a smart city context. This review draws upon existing research, clinical trials, and the development of the hAI! Friend MR application, highlighting its innovative cross-platform approach and discussing future directions and potential challenges.

The proliferation of Internet of Things (IoT) devices and the advent of 6G technologies have introduced computationally intensive tasks that often surpass the processing capabilities of user devices. Efficient and secure resource allocation in serverless multi-cloud edge computing environments is essential for supporting these demands and advancing distributed computing. However, existing solutions frequently struggle with the complexity of multi-cloud infrastructures, robust security integration, and effective application of traditional deep reinforcement learning (DRL) techniques under system constraints. To address these challenges, we present SARMTO, a novel framework that integrates an action-constrained DRL model. SARMTO dynamically balances resource allocation, task offloading, security, and performance by utilizing a Markov decision process formulation, an adaptive security mechanism, and sophisticated optimization techniques. Extensive simulations across varying scenarios-including different task loads, data sizes, and MEC capacities-show that SARMTO consistently outperforms five baseline approaches, achieving up to a 40% reduction in system costs and a 41.5% improvement in energy efficiency over state-of-the-art methods. These enhancements highlight SARMTO's potential to revolutionize resource management in intricate distributed computing environments, opening the door to more efficient and secure IoT and edge computing applications.

YouTube educational resources can help improve physics classroom instructional practices to promote learning. This study investigated the impact of integrating YouTube-based virtual labs on learners' engagement and understanding in physics classrooms. The study was conducted in a secondary school in Nairobi City, Kenya. The participants were secondary school physics students in classes from one to two. The study applied a mixed research approach. The study participants were grouped into two groups: an experimental group and a control group. Unlike the control group, the experimental group was subjected to YouTube-based virtual labs. The study relied on quantitative data from achievement tests and qualitative data from observations during Virtual labs integrated physics classroom and focus group discussions. The results from achievement tests revealed no significant difference in performance between the experimental and control groups. On the contrary, focus group discussion results showed that incorporating YouTube-based virtual labs in physics classrooms improves learners' understanding. Finally, research findings from focus groups and observations signified that integrating YouTube-based virtual labs in physics classrooms improves learners' engagement.

Millimeter-wave (mmWave) signals experience severe environmental path loss. To address this issue highly directional antennas with beamforming capabilities are used to align the transmit and receive antenna beams, allowing the signal to travel longer distances. However, advanced algorithms are needed to detect the angle-of-arrival (AoA) for better alignment of the antenna beams. Combining software-defined radios (SDR) with mmWave radio frequency (RF) antenna systems can enable researchers to develop advanced algorithms for real-world scenarios. Using low-cost mmWave RF front-end components and an SDR, we develop a testbed that uses open-source tools and high-level programming languages for beamforming algorithm development. In the testbed, we prototype a deep Q-network (DQN) algorithm for mmWave AoA detection. We evaluate the performance of the algorithm by fine tuning the reinforcement learning (RL) hyperparameters.

This paper is concerned with the investigation and analysis of a new operational and technical capability to assess GEO satellites from space-borne platforms using extremely high-frequency high-frequency radar operating at sub-THz frequencies. The concept of close monitoring and highly detailed imagery of GEO assets from all aspects, including those unattainable from the Earth is developed based on the analysis of two proposed orbital deployment scenarios. Accounting for orbital perturbation factors during an extended period of time, the ability to build multi-aspect ISAR imagery of the asset during single and multiple encounters is demonstrated, based on the mutual attitudes of the asset and the radar platform. A linearised model of the encounter geometry is presented and the approach to generate a sequence of ISAR image frames according to the geometry of the proposed scenarios is detailed. The simulation of ISAR frames at two frequency bands, centred at 75 GHz and 300 GHz produced in a developed metaheuristic simulator, GEIST, is demonstrated, to highlight the transition of scattering mechanisms and the change in visibility of particular features. Attitude-agnostic frame-to-frame image alignment and linear feature extraction using the Hough transform is then demonstrated on a sequence of simulated images.

Social networks, with their intricate web of connections, serve as a lens for studying human interaction patterns and community structures. The SNAP Social Circle dataset for Facebook provides a platform for such analysis, representing users as nodes and their interactions as edges. By employing modularity analysis, hierarchical clustering, and k-clique detection, this study investigates the complex social dynamics within the Facebook network. The results shed light on tightlyknit communities, influential nodes, and structural hierarchies, contributing to strategies for enhancing online engagement and community development.

The interactions between geometrical Doppler, beam pattern, and normalized radar cross section (NRCS) result in leakage of geometrical Doppler into the perceived geophysical Doppler. Starting from high-resolution synthetic aperture radar (SAR) data we model this leakage for synthesized real aperture radar (RAR) observations of ocean motion. The uncertainty introduced by leakage is ∼1 m s −1. Corrections proposed in this work, using only the synthesized lower-resolution RAR NRCS, decrease this to O(0.1 m s −1). Further reduction of instantaneous leakage may be achieved through temporal averaging, since the leakage-inducing NRCS gradients appear mostly atmosphere induced and decorrelate rapidly. The azimuth resolution and number of independent samples determine a system's sensitivity to leakage. C-band systems are inherently prone to suffer from greater leakage and worse corrections than their Ku-and Ka-band counterparts. With DopSCA, the propagated effect of leakage is only secondary compared to the pulse-pair uncertainty. In similar systems where pulse-pair uncertainty is suppressed, leakage will dominate instead.This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible.

While tobacco advertising innovates at unprecedented speed, traditional surveillance methods remain frozen in time, especially in the context of social media. The lack of large-scale, comprehensive datasets and sophisticated monitoring systems has created a widening gap between industry advancement and public health oversight. This paper addresses this critical challenge by introducing Tobacco-1M, a comprehensive dataset of one million tobacco product images with hierarchical labels spanning 75 product categories, and DEFEND, a novel foundation model for tobacco product understanding. Our approach integrates a Feature Enhancement Module for rich multimodal representation learning, a Local-Global Visual Coherence mechanism for detailed feature discrimination, and an Enhanced Image-Text Alignment strategy for precise product characterization. Experimental results demonstrate DEFEND's superior performance, achieving 83.1% accuracy in product classification and 73.8% in visual question-answering tasks, outperforming existing methods by significant margins. Moreover, the model exhibits robust zero-shot learning capabilities with 45.6% accuracy on novel product categories. This work provides regulatory bodies and public health researchers with powerful tools for monitoring emerging tobacco products and marketing strategies, potentially revolutionizing approaches to tobacco control and public health surveillance.

This study presents a novel approach to quantum communication by enabling the direct encoding of human-readable messages into quantum states for secure transmission. Unlike traditional methods such as cryptographic key sharing or amplitude-phase encoding, our technique maps binary representations of characters onto quantum states, combining simplicity with scalability. Leveraging the unique properties of quantum mechanics, this approach ensures exceptional security against eavesdropping and demonstrates practical viability. Through simulations on a 10-qubit system, we achieved reliable transmission of messages up to five characters with fidelity exceeding 95%. For longer messages (up to 10 characters), a two-part simulation approach maintained high fidelity, exceeding 85%, while demonstrating scalability. Additionally, tests incorporating symbols and numbers revealed the versatility of this encoding method. By bridging the gap between theoretical quantum communication and practical applications, this research opens pathways for secure military communication, robust IoT networks, and hybrid quantum-classical systems. The findings underscore the potential of quantum systems to transform secure messaging by offering a highly scalable, tamper-proof solution for real-world applications.