Designing AI-enabled video surveillance systems has become far more challenging than simply deciding the number of cameras or estimating storage. Once AI inference is performed on cameras or edge devices, every design decision influences several others. For example, increasing the pixel density needed for facial recognition affects lens selection, bitrate, storage capacity, and network bandwidth, while adding more AI analytics changes edge-computing requirements and software licensing. As a result, conventional CCTV planning methods are no longer sufficient for modern AI deployments. This paper presents the architecture of IndoAI’s integrated planning toolkit, which follows a three-stage workflow ie Design, Size, and Verify, using nine specialised calculators for camera placement, infrastructure sizing, and deployment validation. Alongside these tools, an Appization-based AI Agent License Sizing and ROI Predictor estimates software licensing requirements and deployment economics. Building on recent advances in Vision Language Models (VLMs) and platform economics, we propose a VLM as the natural-language interface to this planning framework. Users can describe surveillance requirements using text, photographs or floor plans instead of manually entering technical parameters. The VLM converts these inputs into structured data for the calculator suite. We also present two systems under development: a VLM-driven video intelligence pipeline that analyses surveillance footage to generate timestamped event detections, evaluated using the UCF-Crime dataset and a VLM-assisted Bill of Quantities generator that produces annotated camera layouts together with storage, bandwidth and licensing estimates for deployment planning.
Large-scale Large Language Model (LLM) inference systems deployed in distributed cloud environments face significant challenges in maintaining low latency, efficient GPU utilization, and energy-aware scheduling across heterogeneous hardware. Traditional Kubernetes-based orchestration frameworks are not optimized for the dynamic memory and compute characteristics of transformer workloads, often resulting in resource fragmentation and increased scheduling latency. This paper proposes Adaptive Resource Orchestration (ARO), a telemetry-driven framework designed for distributed LLM inference in heterogeneous GPU clusters. ARO introduces a Rack Affinity Group (RAG) hierarchical indexing mechanism that reduces scheduling search complexity to O(log N * M) while enabling topology-aware resource placement. The framework integrates multi-objective optimization to balance inference latency, energy efficiency, and GPU utilization. Experimental evaluation on a 32-node heterogeneous GPU cluster (H100, A100, and T4) demonstrates significant improvements over baseline Kubernetes scheduling approaches, including up to 84% reduction in P99 latency and 112% improvement in inference energy efficiency (68 tokens/J). These results highlight the importance of hardware-aware orchestration for improving performance and energy efficiency in distributed AI infrastructure.
This study examines technological strategies implemented to prevent examination malpractices in higher education, focusing on secure platforms, biometric authentication, online proctoring, and adaptive assessment designs. Employing a hybrid research framework, it integrates quantitative surveys, qualitative interviews, pilot program observations, and platform analytics to evaluate efficacy and user experience across diverse academic disciplines. Findings indicate that multi-layered security configurations combining biometric verification, AI-assisted proctoring, randomized question delivery, and blockchain-based audit trails reduce misconduct incidence while maintaining fairness and user acceptance. The research highlights the importance of balancing automated detection with human oversight, ensuring infrastructural readiness, and providing transparent communication to address privacy and usability concerns. Results also reveal discipline-specific variations in adoption and effectiveness, with STEM fields showing greater gains linked to interactive assessment formats. The integration of generative AI tools presents both opportunities for enhanced feedback and challenges related to potential exploitation, underscoring the need for careful design and monitoring. Hence, the outcomes suggest that thoughtfully implemented technological interventions can support academic integrity by aligning security measures with pedagogical objectives and stakeholder engagement
The early and accurate detection of foliar diseases is crucial for improving crop yield and ensuring sustainable agricultural practices. This study presents a Plant Foliar Disease Identification Model (PFDIM) for mung bean (Vigna radiata) leaves using a Support Vector Machine (SVM) classifier, specifically designed for operation in uncontrolled environmental conditions. A custom dataset of 1,307 field-captured images was developed, encompassing both healthy and diseased samples affected by Cercospora Leaf Spot, Powdery Mildew, and Yellow Mosaic Virus. The proposed model integrates preprocessing techniques such as image resizing, augmentation, HSV-based segmentation, and Histogram of Oriented Gradients (HOG) feature extraction to enhance robustness and accuracy. Through Grid Search optimization with 5-fold cross-validation, the model achieved 100% training accuracy and 89% testing accuracy, demonstrating strong generalization across variable lighting, background, and noise conditions. Furthermore, a user-friendly interface, titled Leaf Disease Classification, was developed to allow real-time detection and classification through simple image uploads, enabling accessibility for farmers and researchers. The results confirm that SVM provides reliable, scalable, and cost-effective performance for disease identification in real-world agricultural environments. This research contributes a practical, AI-based solution for early mung bean disease detection, supporting precision agriculture and sustainable crop management practices.
Software-defined radio (SDR) is a promising non-invasive approach for human activity recognition. While the deep learning methods in SDR-based HAR are of growing interest, the comparison of different model architectures has a lack of systematic empirical evidence describing the relative performance of different model architectures with the same signal conditions. Accordingly, this investigation performs an empirical evaluation of several deep learning architectures and classical machine learning architectures based on a publicly available SDR dataset. The publicly available University of Glasgow dataset, which comprises SDR devices and Universal Software Radio Peripheral (USRP) models X300/X310, was utilised to collect the data on the aforementioned activities and subsequently preprocessed and fed into a classifier. Five classifiers were systematically instantiated and evaluated: Convolutional Neural Network (CNN), one-dimensional Residual Network (1D ResNet), Long Short-term Memory (LSTM) network, Decision Tree and a Conditional Generative Adversarial Network (cGAN)-based classifier. Performance metrics were measured through overall classification accuracy since the preprocessing regimes and training regimes were consistent for all models. Experimental results show that the cGAN-based model achieved the highest accuracy of 96.4%, and CNN and Decision Tree show the close accuracy of 95.36% and 94.1%, respectively. Again, the performance of 1D ResNet was 86.2%, and that of LSTM was comparatively less at 75%. These results highlight the power of convolutional and adversarial models in learning discriminative signal features from the signal representations of the SDR, which, compared to purely sequential architectures, such as LSTM, demonstrate its limitation of the complex dynamics of radio frequency signals.
A foundational principle in the field of artificial intelligence asserts that there is a trade-off between a model’s explainability and its effectiveness. This trade-off significantly influences model selection for critical applications. This study presents a meta-analysis of 21 advanced NLP models from 2019 to 2023, encompassing encoder, encoder-decoder, and decoder architectures. A quantitative explainability framework was developed, grounded in architectural features, parameter efficiency, and the availability of interpretability tools. Our analysis revealed no significant correlation between explainability and performance across architectures, which contradicts common assumptions (Spearman’s rho mathbf { xi } _ { 1 } = mathbf { xi } _ { – } 0 . 1 6 0 , mathrm { p } = 0 . 4 8 9 ) ). The degree to which a model can be ex-plained is primarily predicted by the model’s intricacy ( rho = - 0 . 9 5 1 , p < 0 . 0 0 1 ) , though the model’s architectural family moderates this effect. Encoder-based models effectively circumvent the trade-off by achieving higher levels of ex-plainability without compromising performance. These results demonstrate that architectural design, rather than mere performance optimization, signifi-cantly influences interpretability. We hereby propose a formal explainability evaluation method and provide evidence-based recommendations for selecting models for specific use cases. Our contribution to the expanding corpus of research on interpretable AI challenges the prevailing assumption that performance and explainability are inherently incompatible. Furthermore, it offers practical guidance for developing transparent, high-performing natural language processing (NLP) systems.
This paper is describing main ideas and content of the new book which is now in preparation. The aim of this book is to review and explain how the distributed world, international one especially, is organized and investigate potential applicability of the developed and already tested in numerous applications high-level Spatial Grasp Model and Technology (SGT). This tech, actually representing a completely new system paradigm, can manage complex systems with a holistic spatial manner effectively covering physical and virtual dimensions, their interrelations, and integration as a whole. The book will brief different multidimensional areas with examples of practical solutions in them and their combinations in a high-level Spatial Grasp Language (SGL), the key element of SGT. This can allow for the creation and distributed management of very large spatial networks expressing different dimensions, which can be self-analyzing, self-optimising, and self-recovering in complex terrestrial and celestial environments. Also organize dynamic multi-networking solutions effectively supporting global evolution, security, prosperity, and integrity.
Modern financial systems depend on complex digital data pipelines that support regulatory reporting, enterprise analytics, and AI-driven decision-making. However, trust in these systems cannot be achieved through policy and audit alone, particularly in environments characterised by automated transformations, distributed ownership, and high-velocity data movement. This paper introduces a governance-embedded finance digital data foundation that unifies stewardship, privacy-by-design, lineage, and data observability as core architectural capabilities. The proposed framework shifts governance from static oversight to continuous verification by making control effectiveness observable at runtime. Using enterprise-scale case evidence from real-world financial data ecosystems, the paper demonstrates how observable governance improves audit readiness, strengthens provenance for explainable analytics, reduces time-to-detect and time-to-remediate data issues, and increases stakeholder confidence in AI-driven insights. The paper concludes by outlining research directions for real-time governance observability, AI-assisted stewardship, cross-domain validation, and standardised governance maturity models.
Cybersecurity threats pose serious risks to the world, particularly in many educational institutions wherein they were handling, managing, and controlling large volume of sensitive data. Human errors remain a leading cause of data breaches, highlighting the importance of assessing the employees’ level of cybersecurity awareness. This study aimed to determine the level of cybersecurity awareness of BTECH employees in terms of knowledge, skills, and attitude, and to examine whether there are significant difference and relationship among selected variables. Descriptive-quantitative research design with comparative and correlational components was utilized in this study. Data were collected from 190 teaching and nonteaching employees using a purposive sampling method. Gathering of data was accomplished through an adopted-modified and validated survey questionnaire. To analyze the data, percentage, frequency, weighted/composite mean, t-test, one-way ANOVA, and Pearson Product-Moment Correlation were used as statistical tools. Findings revealed a moderate level of cybersecurity awareness, with skills ranking highest (WM = 3.88), followed by attitude (WM = 3.47) and knowledge (WM = 3.29). A significant positive relationship was observed between knowledge, skills, attitude, and overall cybersecurity awareness. No significant differences were found across most demographic variables except for position. To significantly enhance and strengthen the cybersecurity knowledge, skills, and attitude of BTECH employees, this study concludes that proposing a cybersecurity training program is essential to improve cybersecurity awareness and reduce institutional vulnerability to cyberattacks.
Las empresas generan grandes volúmenes de datos a través de sistemas transaccionales, pero no siempre logran centralizarlos ni analizarlos de manera efectiva para la toma de decisiones estratégicas. En la industria alimentaria del consumo masivo, y particularmente en el sector del café en Ecuador, el mercado y la competencia exigen un enfoque basado en datos para identificar tendencias, oportunidades de crecimiento y analizar con mayor precisión el comportamiento del consumidor. En este contexto, una empresa dedicada a la producción y comercialización de café enfrenta la necesidad de mejorar su gestión comercial mediante el uso de la inteligencia de negocios. El principal problema radica en la dispersión de información proveniente de diversas fuentes, lo que dificulta el acceso a informes actualizados y comprensibles. Por ello en esta investigación se desarrolla un modelo de inteligencia de negocios que facilite la consolidación de datos y optimice la gestión comercial. Finalmente, al evaluar su implementación con el área estratégica comercial de la empresa, se validó que esta solución contribuye a la toma de decisiones más seguras, ágiles y confiables basadas en información precisa y oportuna
This article investigates the model-theoretic independence of an analytic and hypercomputational strengthening of the P vs NP problem from Zermelo-Fraenkel set theory with the Axiom of Choice (ZFC). By elevating to Π11 analytic statement, we prove ZFC independence. The Necessary Transference Principle [8] then extends this to the classical Π02 case via physical model exclusion. Specifically, we show that in the constructible universe (L), the failure of Σ11- uniformization leads to L ⊨ [P*] ≠ [NP*], whereas a generic extension (M_G) constructed via a Σ1k-complete oracle (O_G) yields M_G ⊨ [P*] = [NP*]. To address this independence, we introduce a novel Foundational-Physical Paradigm. By applying Landauer’s Principle and Kolmogorov complexity analysis, our analysis suggests that realizing O_G as a physical oracle would require entropy dissipation that scales beyond plausible energy bounds of the observable universe. Consequently, we formally propose Axiom X (The Axiom of Bounded Computation), which constrains the mathematical universe to physically admissible computational structures. We demonstrate that within the extended system ZFC_X, the independence is converted into a formal proof of [P*] ≠ [NP*], conditional on accepting Axiom X as a physically motivated foundational principle. This work provides a unified framework bridging descriptive set theory, computational complexity, and the fundamental laws of thermodynamics.
Accurate verification of general election outcomes is essential for understanding political dynamics and democratic processes. However, traditional forecasting methods such as opinion polls and voter surveys often suffer from sampling bias, delayed responses, and limited ability to capture rapid sentiment shifts. This study aims to propose and evaluate a transformer-based political sentiment analysis framework using online news articles to verify the outcome of the 2024 United Kingdom General Election. A dataset of 2,299 online news articles published between January and July 2024 was collected from seven major UK news outlets. A semi-supervised labelling approach was applied to assign sentiment polarity toward the Conservative Party and the Labour Party. To handle long-form political texts exceeding transformer token limits, articles were summarised before being processed. A fine-tuned BERT-base (uncased) model was trained to classify sentiment into positive, neutral, and negative categories. Model performance was assessed using accuracy, precision, recall, and F1-score, while temporal sentiment aggregation was conducted to compare predicted trends with the actual election outcome. The fine-tuned BERT model achieved strong sentiment classification performance, demonstrating robustness in identifying political sentiment from structured news media. Temporal analysis revealed a predominance of negative sentiment toward the Conservative Party, while the Labour Party received significantly higher levels of positive sentiment, particularly in the weeks leading up to the election. These aggregated sentiment trends successfully aligned with and verified the actual election outcome. The findings confirm that online news articles are a reliable alternative data source for political sentiment analysis and election outcome verification. Additionally, the study demonstrates the effectiveness of domain-adapted transformer-based models, such as BERT, in capturing meaningful sentiment shifts that reflect real-world electoral results.
Warehouse management plays a critical role in modern supply chain operations. With the rapid expansion of global trade, e‑commerce platforms, and complex logistics networks, organizations increasingly rely on advanced technological systems to manage warehouse activities efficiently. A Warehouse Management System (WMS) is a specialized software platform designed to control, coordinate, and optimize warehouse operations such as receiving goods, inventory management, order picking, packing, and shipping. This study explores the role of warehouse management systems in improving warehouse efficiency and overall supply chain performance. The research uses a qualitative approach based on secondary data collected from academic journals, books, and industry publications. Existing literature was analyzed to identify the functions, advantages, and challenges associated with implementing WMS technologies. The findings suggest that warehouse management systems significantly improve inventory accuracy, increase operational efficiency, reduce human error, and enhance order fulfillment speed. Furthermore, modern WMS platforms integrate emerging technologies such as cloud computing, automation, artificial intelligence, and the Internet of Things to create smart warehouse environments. Despite these benefits, organizations may face several challenges during implementation, including high initial costs, employee training requirements, and system integration complexities. The study concludes that warehouse management systems are essential tools for modern logistics management and will continue to play a significant role in improving supply chain efficiency in the future.
What you will read in the next pages is an anthology-worthy element and should be taken seriously. In this work, we attempt to describe a place that is unknown and therefore difficult to define, in the sense of arriving in an unknown place, which therefore will not be easy to describe. In the path that takes us from non-existence to being and becoming. Placed in this location, it becomes interesting that different things or people coexist that have not yet been presented in the era in which we connect and go to browse. A very important aspect is the determination of the window in which one says he had a dream where he saw a car that does not yet exist today but perhaps will exist in the future, or a deceased person or one who will exist in the future; So, once the time window is set, which isn't easy, and determining it is an important aspect in which brainwaves will then be selected. This selection is easy and precise when you're awake and thinking about something and the time window is very clear, so a small interaction with the person is enough. It's a different story when you're dreaming. Shannon's theorem, used in electrical engineering and telecommunications, helps select and "correlate" signals. Why this theorem was invented, crucial for effectively sampling signals and being able to fully reconstruct them, is one of those mysteries that leads us to ask ourselves at lunch with the family: Do I have a colleague? We must therefore perfect the start and stop of recording during the dream by focusing on where we are interested, useful for discovering what we think and that our brain suggests things that we thought were lost in our historical culture and instead suddenly... what we need emerges. We would therefore try to obtain feedback from the Afterlife in order to reproduce the brain waves seen in the dream phase thanks to a helmet connected to proactive and not just generative Artificial Intelligence.
Generative artificial intelligence now drafts our correspondence, condenses our reading, and solves our problems faster than we can articulate them. A growing body of evidence suggests, however, that the very tools that lift short-term output can quietly erode the cognitive capacities they appear to augment. This paper synthesizes four recent empirical studies—a randomized controlled trial of AI in high-school mathematics, a field experiment with elite management consultants, a classroom trial of an AI physics tutor, and a neurophysiological study of AIassisted writing—and argues that the headline question is badly posed. The decisive variable is not whether people use AI but how they use it. We adopt cognitive debt, a construct describing the trade of present convenience for future capability, as an organizing lens, and develop a threetier framework for cognitively sustainable use. Tier 1 delegates non-developmental tasks without guilt; Tier 2 protects domain expertise by deploying AI adversarially rather than substitutively; Tier 3 preserves the desirable difficulties on which durable learning depends. Each tier is grounded in established findings from cognitive psychology. We close with implications for individuals and organizations, and with the limitations of an evidence base that remains early and uneven.
The increasing unpredictability of global supply chains necessitate advanced technological solutions for disruption mitigation. It explored the integration of Artificial Intelligence (AI) and Machine Learning (ML) in project management to enhance supply chain resilience. AI-driven risk identification and forecasting enable organizations to anticipate disruptions and proactively manage risks, while machine learning models optimize supply chain operations through predictive analytics and anomaly detection. The application of AI in decision-making and real-time supply chain adaptation further enhances agility, leveraging scenario planning, digital twins, and AI-powered automation in logistics. Additionally, the convergence of blockchain with AI and ML has introduced unprecedented transparency in supply chain operations. Blockchain-integrated AI enhances real-time tracking, while smart contracts automate compliance, ensuring greater accountability across global supply networks. However, despite these advancements, significant challenges persist. Issues such as data quality and bias in AI-based forecasting, high implementation costs, cybersecurity risks, ethical concerns, and resistance to AI adoption hinder widespread deployment.
This article examines the transformative integration capabilities of SAP S/4HANA Finance Cloud with logistics modules, highlighting the architectural foundations and business value of unified financial-operational data management. The article explores how S/4HANA's Universal Journal (ACDOCA) creates a single source of truth that eliminates traditional boundaries between finance and logistics domains, enabling real-time transaction processing and eliminating reconciliation requirements. The article analyses critical integration scenarios across inventory management, procurement, sales, and manufacturing processes, detailing the technical enablers that facilitate seamless connectivity, including embedded analytics, API management, and cloud platform integration. Implementation considerations, including master data harmonisation, account determination configuration, change management, and third-party integration, are examined through the lens of documented organisational experiences. The article concludes by evaluating quantifiable business benefits and exploring emerging technologies such as artificial intelligence, blockchain, and IoT that are shaping the future evolution of finance-logistics integration in enterprise systems.
The enhancement of IoT applications for low-power and long-range communication requires developing communication techniques that consume a small amount of power while transmitting at longer distances. LoRa backscatter is a promising solution for such applications. In this work, we will develop a model that helps to estimate the communication range between a LoRa backscatter tag and a receiver. The developed model has been tested by simulation using Python, and the results are validated by comparing the achieved range using our model with state-of-art LoRa backscatter works. We have also extended the model to account for the effect of SNR loss due to direct interference from the transmitter and inter-tag interference from neighbouring tags in concurrent LoRa backscatter systems.
Cyber security remains a pressing concern for small and medium-sized enterprises (SMEs), particularly in developing nations where cultural, structural, and leadership barriers hinder proactive security adoption. This study investigates the unique organizational and national culture dynamics influencing cyber security behaviors in Ecuadorian SMEs, exploring how leadership traits, risk perception, and employee attitudes impact cyber security readiness through qualitative interviews with SME leaders across multiple sectors. Using analytical frameworks such as Hofstede’s cultural dimensions, McClelland’s motivation theory, and dialectical organizational theory, the research identifies five primary themes: symbolic implementation of cyber security policies, low formalization of risk response mechanisms, lack of return-on-investment frameworks, leadership styles that either support or hinder risk planning, and widespread resistance to behavioral change. Despite increasing exposure to cyber threats, many SMEs remain reactive rather than strategic, often lacking formal policies, employee training, or incident response plans. Leadership personality traits play a significant role in shaping organizational prioritization of cyber security, with achievement-oriented leaders driving more effective strategies. The study concludes that improving cyber security in Ecuadorian SMEs requires a multi-faceted approach involving cultural transformation, leadership development, and localized application of global standards. National campaigns, leadership training, stronger CSIRT partnerships, and practical tools like ROI calculators are essential to advancing digital resilience. This research provides an evidence-based framework for policymakers, educators, and business leaders aiming to enhance cyber security practices within the specific constraints of Latin American SMEs, recommending future exploration of industry-specific adaptations and scalable public-private partnerships to foster sustained cyber readiness.
This article presents a comprehensive analysis of ethical frameworks for artificial intelligence-driven decision systems, addressing the critical need for responsible AI deployment across industries. As AI systems increasingly influence high-stakes decisions in healthcare, finance, criminal justice, and other sectors, the imperative for robust ethical guidelines becomes paramount. The article examines four core ethical principles-fairness, transparency, privacy, and accountability-that form the foundation of responsible AI development, exploring their interconnected nature and practical implementation challenges. Through analysis of implementation strategies and best practices, the article demonstrates how organizations can translate abstract ethical principles into operational reality through systematic approaches including ethical impact assessments, technical implementation of fairness-aware algorithms, organizational governance structures, and meaningful stakeholder engagement. Industry-specific case studies from healthcare, financial services, criminal justice, and autonomous vehicles illuminate both successful implementations and cautionary tales, revealing how ethical considerations vary across sectors based on decision contexts, stakeholder impacts, and regulatory environments. The article further examines the rapidly evolving regulatory landscape, comparing divergent approaches across jurisdictions including the European Union's comprehensive risk-based framework, the United States' fragmented sector-specific regulations, and emerging international standards. By synthesizing insights from computer science, philosophy, law, and social sciences, this analysis provides practitioners with actionable guidance for developing AI systems that balance technological innovation with human values, ensuring that AI advancement enhances rather than diminishes human dignity, autonomy, and societal well-being. The healthcare sector exemplifies both the transformative potential and ethical complexity of AI decision systems. Kumar's analysis of AI and cloud data engineering in healthcare demonstrates how these technologies are fundamentally reshaping medical decision-making processes, from diagnosis to treatment planning [12]. The integration of cloud infrastructure with AI systems enables unprecedented data processing capabilities, but also amplifies concerns about patient privacy, algorithmic transparency, and the equitable distribution of healthcare resources [12]. This convergence of advanced technologies in healthcare settings underscores the critical importance of robust ethical frameworks that can guide responsible innovation while protecting patient welfare and autonomy.
