Authors: Priya Palanimurugan, Mr Kalaiselvan S, Dr.Thulasi Bharathi M, M.sakthivel

Abstract: Mental health issues remain a global public health concern, especially among the youth and digitally active population. This study examines the effect of digital storytelling as an intervention tool to increase mental health awareness, reduce stigma and encourage positive behavior changes. A quantitative research design involving 300 graduate students aged 18–25 in diverse educational subjects was employed. Participants were divided into an intervention group, which reflects a control group that receives curate digital stories and traditional information-based materials reflecting real-life mental health experiences.Advanced statistical techniques were used to assess the results in three time points (Post, Post-up). Descriptive data briefly presented demographic data; Alpha of Cronback confirmed the reliability of the scale; Confirmation factor analysis (CFA) valid measurement construction; And the multi -comprehensive analysis of the covalent (mancova) identified important group differences. Repeated measures Anova and Structural Equation Modeling (SEM) further detected time-based reforms in the intention of mental health awareness and behavior, mediate by low stigma. Moderation and latent development analysis highlighted demographic effects and individual trajectory patterns. Conclusions suggest that digital storytelling improves mental health awareness and reduces stigma compared to traditional approaches (P <0.01). The narrative-based method was particularly effective among the pre-risk participants for high digital literacy and mental health materials. The study supports the integration of digital story stories in public health education and mental health advocacy programs. These results contribute to increasing evidence that creative digital equipment can change mental health communication, offering scalable, attractive and human-focused solutions.

Published by: vikaspatanker

Authors: Sakthivel S, Vikash P

Abstract: The rapid expansion of digital financial services has significantly transformed the global financial ecosystem by enabling fast, convenient, and seamless transactions. However, this transformation has also increased the vulnerability of financial systems to fraudulent activities such as credit card fraud, identity theft, phishing attacks, insider fraud, and money laundering. Financial fraud results in substantial economic losses, damages institutional reputation, and undermines customer trust in digital banking systems. Traditional fraud detection mechanisms primarily rely on rule-based systems and manual audits, which are reactive, inflexible, and often incapable of detecting complex and evolving fraud patterns in real time. Advancements in artificial intelligence (AI), machine learning (ML), and data analytics have paved the way for intelligent financial fraud detection systems capable of processing large volumes of transaction data efficiently. By learning patterns from historical transaction data and identifying anomalies, AI-driven systems enable early detection and prevention of fraudulent activities. This paper presents an AI-based financial fraud detection framework that integrates data preprocessing, feature engineering, and machine learning-based classification for real-time fraud analysis. The proposed system aims to improve detection accuracy, reduce false positives, and enhance the overall security of digital financial transactions. Experimental results and analysis demonstrate that intelligent fraud detection systems provide scalable, adaptive, and reliable solutions for modern financial environments.

Published by: vikaspatanker

Authors: Mahavani Kb, Bavithra Rs

Abstract: Federated learning has also become a paradigm shift to making machine learning collaborative and not centralized around sensitive data. Federated learning solves the increasing privacy, regulatory compliance, and data sovereignty concerns by preventing the transfer of model training to centralized model training clients, like hospitals, financial institutions, and IoT devices. Cloud platforms are critical to the operationalization of this paradigm as it offers scalable orchestration, secure aggregation, and communication-efficient frameworks. The paper discusses how cloud native federated learning systems decrease the amount of communication, enhance the model convergence, and provide more robust privacy guarantees without violating regulation of systems like GDPR and HIPAA. By applying federated learning to the medical diagnostic and financial fraud detection domains, the study shows that federated learning can be successful in providing a high level of model accuracy and strong privacy protection. The results indicate the significance of supporting federated learning by cloud-native infrastructure that will allow implementing privacy-safe AI solutions that can be widely adopted in regulated industries. From a privacy and regulatory perspective, cloud-based federated learning systems provide strong guarantees that align with data protection regulations such as the General Data Protection Regulation (GDPR) and the Health Insurance Portability and Accountability Act (HIPAA). By eliminating the need for raw data transfer, federated learning inherently supports privacy-by design principles. When combined with advanced privacy-preserving techniques such as differential privacy, secure multi-party computation, and homomorphic encryption, federated learning further strengthens its compliance with strict legal and ethical requirements. To demonstrate the effectiveness of cloud-native federated learning, this study applies the proposed framework to two critical application domains: medical diagnosis and financial fraud detection. Experimental results show that federated models achieve performance levels comparable to, and in some cases exceeding, those of traditional centralized models, while significantly enhancing data privacy and security. In medical diagnostics, federated learning enables collaborative training across multiple healthcare institutions without exposing sensitive patient records. Similarly, in financial fraud detection, federated learning facilitates cross institutional intelligence sharing without compromising proprietary or customer data.

DOI:

Published by: vikaspatanker

Authors: Dr Saumya Bahadur

Abstract: This paper examines the Laṅkāvatāra Sūtra, a foundational text of Yogācāra Buddhism, through the lens of adversarial pedagogy and compares it with contemporary machine learning models, particularly Generative Adversarial Networks (GANs). The Sūtra is notable for its dialogical structure, in which the bodhisattva Mahāmati poses questions, challenges, and objections to the Buddha, who systematically deconstructs these conceptual formulations. This adversarial exchange is not merely rhetorical but functions as a pedagogical process: erroneous views and dualistic constructs are generated, tested, refuted, and refined until the practitioner’s reliance on conceptual elaboration collapses. In this way, the teaching method itself resembles an adversarial learning model, where insight emerges through continuous confrontation with errors. In this paper the author explores the method of learning where adversarial views are used to engage in deep learning and transcendence. GANs provide a modern analogue: they consist of two competing networks—a generator that produces synthetic outputs and a discriminator that evaluates their authenticity. Through iterative feedback and critique, both models improve in tandem, eventually producing outputs indistinguishable from real data. Similarly, the Buddha’s adversarial dialogues expose the “generated illusions” of discriminative thinking, while the “discriminator” function is represented by wisdom (prajñā), which identifies and dismantles conceptual fabrications. The comparison highlights both parallels and divergences. While GANs aim at convergence toward increasingly realistic outputs within representational constraints, the Buddhist adversarial method seeks not fidelity to appearances but the transcendence of representational frameworks altogether, pointing toward non-dual realization and liberation from suffering. This contrast underscores how ancient epistemic practices may resonate with modern computational paradigms while also exceeding them in scope, embedding cognitive, ethical, and soteriological dimensions absent in machine learning. The paper thus proposes that reading the Laṅkāvatāra Sūtra as an adversarial pedagogy provides fertile ground for interdisciplinary inquiry, bridging Buddhist philosophy, cognitive science, and artificial intelligence research.

Published by: vikaspatanker

Authors: Esakkiammal N, Deebika S

Abstract: Artificial Intelligence (AI) chatbots have become integral to modern communication, providing services ranging from customer support to mental health assistance. Their rapid adoption raises critical questions about their psychological influence on users. This study investigates the impact of AI chatbots on human emotional well-being, emphasizing psychological mechanisms such as emotional regulation, social support, companionship, and dependency. Using a mixed-methods approach—combining surveys and semi-structured interviews—the study examines the extent to which chatbots contribute to emotional support and their potential to induce dependency or social withdrawal. Results suggest that while AI chatbots can positively influence emotional well-being by providing accessible support, they may also create challenges, such as emotional over-reliance and diminished real-world social engagement. The paper concludes with practical, ethical, and design recommendations for AI chatbot developers, emphasizing the importance of balancing technology with human-centric emotional care.

Published by: vikaspatanker

Authors: Mallesh Miryala

Abstract: In integration-heavy Salesforce environments, release reliability depends less on the deployment tool and more on engineering controls: correct metadata scoping and ordering, explicit boundary contracts, retry-safe synchronization, and operational feedback. This article presents an end-to-end delivery model that (i) represents metadata, code, and access controls as a dependency graph to build deterministic delta packages and select relevant tests; (ii) treats system boundaries as executable API and event contracts, verified in CI by both providers and consumers to prevent drift; and (iii) implements integrations as idempotent, retry-safe state machines using external identifiers, payload digests, and bounded deduplication windows. We show how layered quality gates—static analysis, targeted suites, contract checks, and observability signals—create a control loop that reduces change-failure rate and shortens recovery time. The result is an implementation-oriented guide, with algorithms, diagrams, and reference patterns for teams operating Salesforce alongside middleware such as MuleSoft and legacy systems including GIS, ERP, and data platforms.

Published by: vikaspatanker

Authors: Ms. Nirmala R G, Pratap K V, Nithilan I

Abstract: The growing adoption of electric vehicles (EVs), renewable-energy microgrids, and portable power systems has intensified the need for efficient and reliable battery management strategies. Conventional passive balancing circuits in lithium-ion battery packs dissipate excess energy as heat, resulting in low efficiency, poor scalability, and thermal stress. This paper presents an Active Cell Balancing Battery Management System (ACB-BMS) employing a bidirectional buck–boost converter topology integrated with an Extended Kalman Filter (EKF)-based state-of-charge (SOC) estimation algorithm. The system dynamically redistributes charge between cells, achieving faster equalization and significantly reduced energy loss compared with resistor-based methods. The EKF enables accurate real-time tracking of each cell’s SOC, improving safety and charge control under varying load and temperature conditions. A complete MATLAB/Simulink simulation model of the proposed system has been developed and validated, demonstrating superior voltage uniformity, faster balancing response, and enhanced energy efficiency. The proposed approach forms a practical foundation for next- generation intelligent BMS architectures suitable for electric vehicles and hybrid renewable-energy storage. Future hardware implementation is planned to extend the technology toward commercial-grade embedded platforms.

Published by: vikaspatanker

Authors: Hanumesha S T

Abstract: Percolation theory provides a mathematically elegant and practically powerful framework for modeling connectivity transitions in random media, with applications ranging from porous materials and composite conductivity to epidemics, network robustness, and transport in disordered systems. A central quantity is the percolation threshold p_c, the critical occupation probability at which macroscopic connectivity emerges with nontrivial scaling. Although p_c is known exactly for a few planar cases and lattices, many practical scenarios require estimation under finite-size, boundary, and uncertainty constraints. This paper develops a rigorous and computation-oriented methodology for percolation threshold estimation that couples (i) probabilistic inequalities and bracketing arguments (crossing probabilities, monotonicity, sharp-threshold heuristics, and finite-size scaling), with (ii) simulation-based estimators (spanning probability curves, union-find connectivity, confidence intervals, and extrapolation). We emphasize a "two-engine" approach: bounds that constrain plausible threshold locations and simulation that refines the estimate while quantifying uncertainty. We also introduce an uncertainty-aware parameterization using intuitionistic fuzzy sets and (hyper)graph abstractions to represent ambiguous occupancy mechanisms and heterogeneous coupling patterns; this is motivated by real settings where the effective "open probability" is not a crisp scalar but a range informed by measurement noise or multi-factor criteria. The final manuscript provides a Word-ready, mathematics-forward exposition, with figures and tables embedded to illustrate lattice configurations, spanning curves, scaling collapse, and probabilistic bracketing.

DOI: http://doi.org/10.5281/zenodo.18092138

Published by: vikaspatanker

Authors: Vishal Ranjan, Dr. Jyoti Yadav

Abstract: High-Performance Concrete (HPC) is an advanced form of cement concrete where ingredients are selected and proportioned to enhance various properties of the concrete in both fresh and hardened states. One key feature of HPC is its higher strength, which offers significant structural advantages. The primary components contributing to the cost of a structural member are concrete, steel reinforcement, and formwork. This paper compares these components when higher-grade concrete, specifically HPC, is used, and highlights how high-strength concrete provides the most economical solution for designing load-bearing members, particularly in carrying vertical loads to the building foundation through columns. The mix design variables critical to concrete strength include the water-cementitious material ratio, total cementitious material, cement-admixture ratio, and superplasticizer dosage, which are analyzed to achieve the desired high-grade concrete mix.

DOI: https://doi.org/10.5281/zenodo.18092124

 

Published by: vikaspatanker

Published by: vikaspatanker