Data privacy and regulatory compliance are also crucial threats to the centralized machine learning practice in the digital financial environment. Federated Learning (FL) provides a referral-free solution that enables several financial units to collaborate in training machine learning models without sharing sensitive customer data. The paper investigates the architecture design, system issues, and performance enhancements of FL in distributed financial settings. By conducting a thorough analysis of communication restrictions, data heterogeneity, and security flaws, the paper identifies obstacles to the smooth implementation of FL. Other emerging technologies in the privacy-preserving methods, adaptive optimization algorithms, and hierarchical architectures relate specifically to the area of financial applications that are evaluated in the paper. Moreover, it presents new governance regimes and practical applications, underscoring the relevance of FL in fraud detection and credit rating assessment, as well as in investment approaches. The insights introduced not only demonstrate the viability of FL in the financial sphere but also outline a path towards the creation of safe, scalable, and regulatory-compliant AI ecosystems. Grounding technology in collaboration between institutions and legal frameworks, federated learning is one of the core infrastructures of the new era of safe, intelligent financial systems.
In today’s digital first landscape, enterprise IT systems form the backbone of mission critical operations, demanding resilience, reliability, and rapid crisis response. The complexity of distributed, cloudnative, and IoTdriven environments has rendered traditional monitoring tools inadequate for ensuring uninterrupted services. AI driven operational dashboards represent a paradigm shift by combining realtime data visualization with advanced machine learning, natural language processing, and predictive analytics. Unlike conventional dashboards that passively display metrics, these intelligent platforms actively detect anomalies, forecast failures, and recommend or even automate remedial actions. This reduces mean time to detect (MTTD) and mean time to resolve (MTTR), providing enterprises with faster, more informed decisionmaking during crises such as outages, cyberattacks, or traffic surges.Case studies of prominent industries show quantifiable enhancements in anomaly detection accuracy, alert noise reduction, and operational efficiency. There are obstacles, however, including data complexity of integration, explanation holes, model shift, and organizational resistance. The review integrates existing technological bricks, experimentation findings, and industrial practices alongside identifying research gaps and limitations. In delineating a theoretical framework and future research directions, it substantiates the imperative of AI driven dashboards as adaptive, reliable, and scalable solutions for enterprise IT resilience in a more dynamic world of operation.
The societal perspectives on energy use and sustainability are further complicated by incredible advancements in systems that inform through data-intensive information engineering. Artificial Intelligence (AI) provides data-rich capability in data-rich environments, while the energy system uses much less energy in aggregate as an energy system comparable to blockchain applications. As predictive and adaptive intelligence has proven to be more efficient, energy utilization is also evolving in AI and its resource and data management ecologies. The decentralized elements, such as openness and security of blockchain applications, also greatly add to contact reliability of the energy operative performance of the entire ecosystem by reducing wasted and/or non-usage of energy and/or other resource domains through information engineering systems. The current use of smart grids as important for smart grids considering edge computing/applications and thermal fluid systems applications for city microclimate controls for developing economies is included and analyzed. The paper is a thoroughly informative presentation of smart cities with AI, blockchain technology, and sustainable resource consumption and use, based on emergent literature to communicate meaningful contributions to the discourse area.
In the semiconductor industry, where reliability, yield, and performance are the key competitive factors, the ATE plays a pivotal role. It enables high-speed and highly accurate testing and prepares outputs such as a test log, waveform capture, diagnostic reports, and yield statistics necessary for successful failure analysis (FA). It also helps identify signatures of failure, isolate the defect sites, and correlate suspicious test-based abnormalities with their root causes. Parametric and functional aberrations, and thus ATE generation and verification, have become major tools for engineers to correlate complex defect behaviors with behavioral-phase-level defects that are generally overlooked if one simply relies on a traditional inspection system. Future integration of ATE data into advanced analytics and machine learning will also contribute to a proactive and prospective approach to FA. There remain numerous challenges today, however, in dealing with big data, delivering diagnostic resolution, and sustaining diagnostic methodology for new architectures such as 3D ICs and advanced nodes. This paper discusses the roles of ATE deliverables in FA and addresses the latest developments, bottlenecks, and future directions toward improving diagnostic resolution and product quality assurance.
The growing complexity and size of current VLSI designs have resulted in the placement and routing (PnR) steps of the design flow being highly constrained and time-consuming. Traditional manual constraint definition, validation, and iterative design refinements that are needed to close the design are a major setback on the timeline of design closure. This paper describes the in-depth analysis of constraint-conscious automation of VLSI PnR workflows using scripting and machine learning based models. The discussed methodology utilizes both custom scripting, in terms of Tcl and Python, and physical design tools, to mechanize the extraction, verification, and enforcement of physical design constraints through multiple floor-planning and layout iterations. Also, the concept of machine learning is presented to foresee the congestion hotspots, optimize cell positioning, and adjust the routing strategies using the historical design data. The application of supervised learning for congestion estimation and reinforcement learning to legal placement are compared in some of the most popular EDA toolchains. Experimental results indicate that combined scripting-ML can greatly lessen turnaround time, increase the quality of design, and improve turnaround time closure. Also, the framework permits speedy design-space exploration and constraint-re-targeting across projects. The significance of hybrid automation techniques in performing efficient, scalable, and constraint-compliant PnR has been brought out in this research, especially at advanced nodes where the interaction between constraints is highly nonlinear and design margins are very thin.
The high-velocity growth of off-label administration of glucagon-like peptide-1 (GLP-1) receptor agonists that were initially approved to be used as Type 2 diabetes medications has raised new safety issues and initiated regulatory procedures for developing adverse event label expansions. Since these agents are becoming popular in weight reduction, the treated population has become less homogenous, and therefore, the risk of adverse events that were not seen during the pre-approval trials is more probable. The paper critically evaluates the pharmacological justification of off-label use, the role of post-marketing surveillance in identifying safety signals, and the regulatory frameworks of various jurisdictions regarding the correction of the labeling of that product. It also deals with regulatory communication obligations of regulators, manufacturers, and prescribers to communicate changing risks; and with operational impediments to the successful implementation of label changes in a timely and consistent manner. It outlines the evidential challenges in demonstrating causality among various patient groups as well as recommending approaches to future regulation, including the development of dynamic labeling systems, streamlined global pharmacovigilance, and enhanced safety research. Analysis indicates that the adaptive, coordinated, and evidence-based strategies are essential to ensure the protection of public health and, at the same time, maintain access to therapeutics in the new reality of changing prescription patterns at a blistering pace.
Recent advancements in Integrated Metal Oxide Semiconductor (MOS) sensor arrays have made them valuable technologies for advanced environmental monitoring, as they can detect multiple gases simultaneously on a single chip. Specifically, MOS sensor arrays have the potential to detect carbon monoxide (CO), nitrogen dioxide (NO₂), methane (CH₄), ammonia (NH₃), and volatile organic compounds (VOCs). Applications such as urban air quality monitoring, industrial site safety, and various smart city infrastructures can utilize multi-gas sensors. This paper examines the design, fabrication, and field deployment of multi-gas MOS sensor arrays from the perspectives of device architecture, transduction material selection, microfabrication strategies, and signal processing. The major challenges, including issues of cross-sensitivity, selectivity, drift, and size reduction, have been covered. Additionally, ways to mitigate these challenges through methods such as operational temperature modulation, temperature-assisted nanostructured materials, and machine-learning-based calibration have been discussed. This work provides a comprehensive overview of important technologies, methods, and concepts, highlighting the real capabilities of integrated MOS sensor arrays. These arrays could be utilized to create reliable, scalable, and cost-effective multi-gas sensing systems for next-generation environmental monitoring.
This paper presents a novel correction framework for multivariate time series data that enhances data quality by integrating Autoencoder-based Generative Adversarial Networks (GANs) with correlation-aware analysis. With the widespread adoption of deep learning models for sensor-driven applications such as quality control and demand forecasting, data anomalies caused by sensor faults and network errors have emerged as a critical challenge, often degrading model performance. In multivariate time series, anomaly correction is particularly difficult because erroneous values must be distinguished from legitimate temporal variations while preserving inter-variable dependencies and temporal consistency. To address this challenge, we propose a data correction method that combines deep learning–based anomaly detection with correlation-driven correction. An Autoencoder-based GAN is employed to identify anomalous patterns in multivariate time series, while a window relevance matrix is introduced to guide precise correction. This matrix captures complex relationships among variables by jointly incorporating dynamic time warping and Pearson correlation coefficients within sliding windows. By leveraging both temporal alignment and statistical dependency, the proposed approach performs anomaly correction that maintains structural coherence across variables. Extensive experiments conducted on diverse multivariate time series datasets demonstrate that the corrected data consistently improves predictive performance compared to raw data. These results indicate that the proposed method effectively enhances data quality and model reliability, offering a robust solution for anomaly correction in multivariate time series applications.
Optimizing of drill hole spacing for coal resource classification is a method for optimizing and evaluating coal resources in a seam within a coal mining area. The Kelay Block 3 area is included in the concession of PT Berau Coal, Berau Regency, East Kalimantan. The main coal-bearing formation is the Labanan Formation. The Labanan Formation is known to have fairly thick and low-quality coal seams. Exploration drilling has been conducted at 69 drilling locations, consisting of 34 core drilling locations and 35 open-hole drilling locations. The coal resource classification in the study area falls into the moderate geological condition category according to SNI 5015:2019. This research was conducted to analyze the optimal borehole spacing in the Labanan Formation for coal resource evaluation and classification in the study area using the global estimation variance (GEV) approach. The results obtained from optimizing borehole spacing in the study area were taken at a spacing of 350 m for the measured category, 600 m for the indicated category, and 1200 m for the inferred category, based on the thickness variable. This is because the global estimation variance (GEV) method considers population and variation, which are values that approximate variance thru GEV analysis. Optimizing borehole spacing using the global estimation variance (GEV) method showed wider and more optimal results than the provisions of SNI 5015:2019 under the same geological conditions.
Grape leaf diseases cause serious problems for viticulture around the world by having a substantial impact on the output and quality of grapevine cultivation. Conventional manual diagnosis techniques are laborious, subjective, and frequently ineffectual in extensive field set-tings. In this study, we present the Inceptive Synergic Network Model (ISNM), a revolutionary deep learning framework for the precise, effective, and scalable categorization of grape leaf diseases. With a Scale-Invariant Feature Learning (SIFL) module to improve spatial in-variance and reduce superfluous background noise, ISNM combines the advantages of Mobile-Net and ResNet-50 as dual backbones for reliable multi-scale feature extraction. We also investigate how wavelet-based sub-band decomposition can enhance feature localization in a variety of illumination and disease-spread scenarios. Lightweight convolutional layers optimize the fused deep features, allowing for de-ployment on edge devices for real-time monitoring. The suggested ISNM outperforms baseline CNNs and other conventional designs in terms of precision, recall, and computational efficiency, achieving a state-of-the-art accuracy of 98.75% when tested on the Plant Village grape leaf dataset. With potential uses in precision farming, self-sufficient vineyard monitoring, and sustainable crop management, this work provides a scalable, interpretable, and useful approach to early grapevine disease diagnosis.
Gastrointestinal diseases affect over 40% of the global population and rank among the leading causes of cancer-related deaths worldwide. Wireless capsule endoscopy (WCE), though minimally invasive, generates 45,000-50,000 images per procedure, with expert endoscopists missing 22-28% of pathological cases due to examination complexity and varied disease presentations. Deep learning approaches like Con-volutional Neural Networks require large annotated datasets, which are rarely available in medical imaging. Capsule Networks (CapsNets) excel in limited-data scenarios through spatial-hierarchy inference but struggle with subtle features in complex medical images. We present a Boosted Capsule Network that incorporates dual enhancement mechanisms: modified feature boosting via intensity inversion to expose diverse low-level patterns, reducing the generalization gap by 45%, and class capsule amplification to sharpen decision boundaries and improve inter-class separation. Ablation studies on the Kvasir-V2 dataset show individual components achieve 89.3% (feature boosting) and 91.7% (class boosting), while their synergistic combination reaches 97.90% accuracy, a 15.8% improvement over baseline CapsNet (82.10%) and 1.1% over previous state-of-the-art (96.80%). The lightweight architecture adds zero trainable parameters for class boosting and minimal overhead for feature enhancement, enabling real-time clinical deployment. We experimentally deployed the model as a web-based prototype with human-in-the-loop uncertainty handling for confidence scores below 0.5.
The article outlines the key components of a high-quality information and advisory environment within an educational institution. Tasks are grouped, and the functions and principles for organizing the creation and development of effective learning in the information and advisory environment of an educational institution are prescribed. The role of cloud technologies in expanding the potential of an educational institution's information and advisory environment is demonstrated. An experimental study at the ascertaining stage found that the vast majority of respondents were not sufficiently ready to use the information and advisory environment; they lacked knowledge, skills, motivation, and abilities. Therefore, we have developed a set of necessary and sufficient pedagogical conditions and proposed innovative methods through which the selected components of an educational institution's information and advisory environment foster fruitful cooperation, operate in interrelation, and contribute to the high-quality provision of educational services. At the formative stage, in the experimental group, unlike the control group, a more significant and statistically significant increase was observed in all criteria corresponding to the components, which allowed us to say that the developed methodology, which was based on certain authorial pedagogical conditions for increasing the level of educational services and organizing an information and advisory environment, is quite effective in an educational institution.
This systematic literature review investigates the recent advancements in the applications of Artificial Intelligence (AI) and Radiomics in Contrast-Enhanced Mammography (CEM), focusing on their diagnostic, predictive, and prognostic value in breast cancer imaging. The study aims to synthesize current evidence on the integration of AI-based algorithms and radiomic approaches that enhance lesion detection, classification, and clinical interpretation. The review follows the PRISMA protocol to ensure methodological rigor and transparency. A comprehensive search was conducted across Scopus and PubMed databases using the keywords contrast, mammography, and artificial intelligence, retrieving relevant studies published in 2025. After applying inclusion and exclusion criteria, 36 primary studies were selected for qualitative synthesis. The analysis identified three major thematic domains: (1) AI architectures and classification/detection models, (2) Radiomics and multi-modality predictive/prognostic models, and (3) Segmentation, microcalcification, and data-tooling for detection. The findings revealed that hybrid and ensemble deep learning models significantly improved diagnostic performance, while radiomics-based approaches enhanced molecular subtype prediction, risk stratification, and treatment planning. Furthermore, advances in segmentation and synthetic data augmentation improved lesion localization and model robustness, supporting more accurate and reproducible image interpretation. Despite methodological progress, challenges persist regarding data standardization, model explainability, and clinical validation across diverse populations. The review concludes that integrating AI and radiomics within CEM holds substantial potential for transforming breast cancer diagnostics by improving precision, interpretability, and clinical decision-making. Continued development of standardized frameworks and multicenter validation is essential to ensure reliable, ethical, and clinically applicable AI adoption in breast imaging practice.
This paper investigates the application of Quantum Gravity Theory (QGT) to the dynamics of elliptical galaxies, with a primary focus on the well-observed systems M87 and M49. Based on the theoretical framework proposed by Wong et al. (2014), which integrates relativity theory and quantum theory, QGT offers a novel explanation for galactic dynamics without invoking dark matter. The theory posits that quantum gravitational effects, including the exchange of gravitons and antigravitons, produce an effective antigravity phenomenon in the outer regions of galaxies, mimicking the dynamical influence traditionally attributed to dark matter. We apply the QGT potential to model the kinematic data of M87 and M49, two massive elliptical galaxies with extensive observational constraints. We derive the QGT-modified Jeans equation, incorporating the QGT potential ( ) to predict velocity dispersion profiles. By rigorously accounting for baryonic mass—including the intra-cluster medium (ICM)—we achieve fits to data spanning 0.5–150 kpc using the stellar mass-to-light ratio as a free parameter. Bayesian analysis reveals decisive statistical superiority over NFW (ΔBIC > 13) and MOND models (ΔBIC > 9). These results suggest that QGT, rooted in the fundamental principles of modern physics, offers a compelling alternative explanation for the dynamics of elliptical galaxies. By successfully modeling M87 and M49 without dark matter, this study challenges the necessity of the dark matter hypothesis in these galactic systems and opens new avenues for exploring quantum aspects of gravity at astrophysical scales
The growing adoption of Healthcare Internet of Things (HIoT) systems has improved patient monitoring and clinical efficiency. Still, it has also exposed hospitals to ransomware attacks that can disrupt life-critical operations. Conventional intrusion detection systems (IDS) such as Snort, Suricata, and Kitsune struggle to provide tamper-proof evidence or prioritize device-specific risks, limiting their effectiveness in clinical environments. To address this gap, we propose a blockchain-assisted framework that integrates hybrid anomaly detection, Byzantine fault-tolerant forensic logging, and severity-aware mitigation. Suspicious traffic is first evaluated using anomaly scoring with EVT-based thresholding, then immutably recorded on a blockchain ledger via PBFT consensus, ensuring tamper-resistance. Smart contracts enforce mitigation decisions based on severity scores that account for anomaly magnitude, device criticality, and network exposure, thereby guaran-teeing the rapid protection of life-support devices while minimizing unnecessary disruption to low-risk equipment. An experimental evalua-tion of the N-BaIoT, ToN_IoT, and CIC-IDS2017 datasets shows that the framework achieves a detection accuracy of up to 98%, a false-positive rate of 1.1%, and an average latency of 6.8 ms, outperforming baseline IDS solutions. Security analyses confirm resilience against log tampering, obfuscation-based evasion, and denial-of-service flooding, while throughput scalability exceeded 2200 TPS across hospital nodes. By combining blockchain accountability with clinically aware mitigation, this framework provides a robust, real-time defense against ransomware in HIoT environments, advancing the state of cybersecurity for patient-centered healthcare systems.
Smart mobility services generate large volumes of sensitive location and identity data, raising critical concerns related to privacy leakage, security vulnerabilities, and trust in large-scale urban deployments. To address these challenges, this paper proposes a blockchain-based privacy-preserving framework for smart mobility services that integrates geo-indistinguishability, pseudonymous authentication, Zero-Knowledge Proofs (ZKPs), and Proof-of-Authority (PoA) consensus into a unified architecture. The framework ensures end-to-end privacy by combining calibrated location obfuscation with decentralized transaction validation and immutable auditability, thereby mitigating both inference-based attacks and reliance on centralized trust. The proposed framework was evaluated using the TAPAS Cologne mobility dataset, comprising 1,000 simulated vehicles and 20 block-chain validators. Experimental results demonstrate that adversarial inference accuracy is reduced to below 12%, while approximately 75% navigation utility is preserved at balanced privacy budgets. Security analysis confirms robust protection against tracking, replay, Sybil, and collusion attacks, with replay attack success rates reduced from 70% to 2% through the enforcement of timestamps and nonces, along with cryptographic verification. Performance evaluation demonstrates that the framework achieves high throughput (1,200 transactions per second) with sub-second latency (0.8 seconds) under realistic transaction loads. Storage growth is optimized to 2.1 GB per million transactions, and the PoA consensus mechanism achieves approximately 30% lower energy consumption compared to Proof-of-Stake-based designs. In addition, resilience ex-periments confirm Byzantine fault tolerance under up to 30% malicious validator participation, without service degradation. Overall, the results demonstrate the practical feasibility of deploying the proposed framework in real-world smart mobility ecosystems that require simultaneous privacy preservation, scalability, and energy efficiency. The framework represents a significant step toward trustwor-thy, privacy-aware, and sustainable smart-city mobility infrastructure, providing a robust foundation for next-generation decentralized mo-bility services.
This paper introduces a unified analytical and computational framework for controlling and characterizing chaos in generalized fractional-order dynamical systems. Despite advances in fractional calculus, existing methods face persistent challenges: high computational costs, limited stability criteria, and a lack of quantitative chaos measures for generalized operators. To bridge these gaps, we propose Apc-GM, a novel adaptive predictor-corrector algorithm that enhances numerical accuracy by 45% and reduces computation time by 23% compared to classical approaches. Our framework integrates three core pillars: a generalized controllability theory, an extended Lyapunov stability analysis valid for all orders α>0, and a quantitative chaos criterion based on generalized Lyapunov exponents. Numerical validation on a 4D fractional Lorenz-Stenflo system demonstrates 92% chaos suppression efficiency under optimal control, while maintaining consistency with classical results. This work provides robust, ready-to-use tools for modeling, analysis, and control of complex fractional-order systems in engineering and applied sciences. Keywords: fractional calculus, chaos control, adaptive numerical methods, nonlinear dynamics, stability analysis, Lyapunov exponents, predictor-corrector methods.
The landscape of digital marketing is transformed by Artificial Intelligence (AI) technologies that improved personalization, efficiency, and customer engagement. However, research on AI adoption is limited, especially in Saudi retail small and medium-sized enterprises (SMEs). This study addresses this gap by examining how the perceived usefulness of AI in different digital marketing stages influences adoption intention. Drawing on the Technology Acceptance Model (TAM) and the RACE framework (Plan, Reach, Act, Convert, Engage), five hypotheses were developed to test the effects of perceived usefulness on adoption intention across these stages. Data were collected through surveys from 450 decision-makers in Saudi retail SMEs. Structural Equation Modeling (SmartPLS) was used for analysis. The results support four hypotheses. Perceived usefulness in the Plan, Reach, Convert, and Engage stages positively influenced adoption intention (H1, H2, H4, H5). The Act stage showed no significant effect (H3). These findings highlight that AI’s perceived value differs across marketing stages. The study contributes to theory by extending TAM in a multi-stage marketing context. It shows that adoption intention is shaped by how useful AI is perceived across customer journey stages. Practically, the results guide managers to focus AI investments in stages where perceived usefulness drives stronger adoption, especially Reach, Convert, and Engage. For policymakers, the findings emphasize the need for targeted support and training to enhance AI readiness in Saudi SMEs.
The article clarifies the meaning of "integration" in education and outlines its types. The role of the integrated information educational envi-ronment in higher education is substantiated; in particular, key provisions for organizing it are presented. Ways to prepare students for ap-plying integrated forms of educational organization in higher education through various types of disciplinary integration are presented. Dur-ing the ascertaining experiment, the diagnostic methods used revealed that respondents in the EG and CG lacked skills and analytical abili-ties, as well as self-control, reflection, self-regulation, and professional self-improvement. During diagnostics, we found that more than 70% of future specialists in socionomic specialties are not ready for integrated training in the information educational environment, indicating the need to study the problem outlined above. Summarizing the results of the ascertaining stage of the study, we emphasize the need to develop, substantiate, and implement new content, pedagogical conditions, practical means, forms, and methods. The data obtained during the forma-tive experiment allow us to conclude about the effectiveness and efficiency of implementing pedagogical conditions, because in the EG, where pedagogical conditions were introduced, the respondents increased their initial level of readiness for the specified problem.
Congenital heart defects and structural cardiac abnormalities including atrial septal defect (ASD), Patent Foramen Ovale (PFO), Ventricular Septal Defect (VSD), Patent Ductus Arteriosus (PDA), and left atrial appendage anomalies represent a significant global health burden, often requiring timely diagnosis and intervention to prevent long-term morbidity and mortality. The development of transcatheter occlusion devices has transformed the treatment of many disorders, providing less invasive, safer, and more effective alternatives to traditional surgical procedures. This review delves into the historical progression and recent technological advancements in occluder device design, materials, and deployment methods. It demonstrates the transition from early metal-based devices, such as nitinol frameworks, to next-generation biodegradable occluders made of polylactide, polydioxanone, and polycaprolactone. These materials have excellent biocompatibility, facilitate tissue integration, and prevent the long-term difficulties associated with permanent implants. Innovations such as 3D/4D printing, shape-memory polymers, and hybrid devices are pushing the development of safer and more patient-specific solutions. Despite positive preclinical and clinical results, there are still hurdles in optimizing degradation rates, mechanical strength, and long-term effects. This study gives a thorough assessment of current and emerging occlusion technologies, focusing on their potential to improve procedural success, patient safety, and the future landscape of structural heart disease treatment.
