Generative AI Applications In Consumer Engagement And Brand Communication
Authors: Chandana M, Vivek R Venkatesh
Abstract: GenAI is revolutionizing the way consumers interact with the brands and engage in conversation with them. This paper examines the ways in which brands use GenAI to develop personalized, co-creative, and immersive consumer experiences. Based on the results of a thorough review of the existing literature and current application practices, the paper highlights three main mecha-nisms through which GenAI can be used to increase engagement: personalization of content; collaboration on the content creation process between the brand and the consumer; and use of conversational AI assistants. It is found that despite the fact that GenAI is very effective at in-creasing personalization and engagement rates, its success heavily depends on the presence of human control and the ability to maintain balance between novelty of technology and emotional engagement. Moreover, consumer acceptance is shown to be age-dependent, with younger gener-ations being more comfortable with AI interactions. The paper provides recommendations on using GenAI in the communication strategy. Generative AI, Consumer Engagement, Brand Communication, Personalization, Human-AI Collaboration, Digital Marketing
Federated Deep Learning-Based Privacy-Preserving Cyber Defense Architecture For Distributed IoT Ecosystems
Authors: Dr. Vimala Roselin J, Jyothi V Kulkarni
Abstract: The explosive growth in IoT devices has created new cybersecurity challenges with an estimated 40 billion devices by 2030, with about 50% being vulnerable to cyber-attacks. The traditional centralized Intrusion Detection Systems (IDS) are faced with significant challenges such as com-munication bottlenecks, point-of-failure, and issues with privacy protection of data. In this paper, a novel cyber defense architecture is presented utilizing federated deep learning approach that incorporates hybrid deep neural networks along with privacy protections for distributed IoT networked devices. The proposed architecture makes use of a three-layer system architecture comprising IoT devices, edge gateways, and cloud cooperation using CNN, BiLSTM, and auto-encoders using a federated learning framework. Evaluation results have demonstrated the perfor-mance superiority of the proposed architecture with an accuracy of more than 99% in detecting anomalies, a 67% reduction in communication overhead, and effective resilience to adversarial attacks due to use of differential privacy.
AI-Driven Talent Acquisition Framework For Predictive Employee Performance Assessment
Authors: Surjya Narayan Sahoo
Abstract: The use of AI in talent acquisition is a significant paradigm shift that enables organizations to identify, evaluate, and select candidates based on their capability to excel in the required job roles. This paper describes an innovative talent acquisition framework built around AI technology and involving machine learning, natural language processing, and predictive analytics to improve the employee performance evaluation process. Specifically, the pro-posed talent acquisition framework encompasses resume scanning, soft skill assessment via conversational AI, and employee performance prediction using One-Class SVM classifiers. Based on quantitative analyses of different organizational databases, the framework pro-vides 95.28% accuracy in evaluating the performance of potential candidates, 40% decrease in recruitment process time, and 35% increase in performance prediction accuracy com-pared to conventional approaches.
Automated Detection Of Lung Diseases From Chest X-Ray Images Using Vision Transformers
Authors: Sakshi Shukla, Surjya Narayan Sahoo
Abstract: Lung diseases continue to be one of the top causes of deaths worldwide, which means that there is a need for efficient and fast diagnosis procedures. In this paper, we provide an overview of recent research in automated detection of lung diseases from chest X-rays with Vision Transformer (ViT). Traditional deep learning solutions such as Convolutional Neural Networks (CNNs) have proven themselves to be very effective. However, they suffer from the inability to model the crucial dependencies between distant spatial elements. Vision Transformers represent a major breakthrough in deep learning through the use of self-attention mechanism to model global relationships between the different parts of the image. This paper provides an overview of recent developments in the area of ViTs, covering such topics as hybrid CNN+ViT architectures, dual-stream attention fusion, and gaze guided architectures. Comparative analysis of the results shows that hybrid approaches, where global relationships are combined with traditional local feature extraction methods, produce the best results with accuracy rates above 98%.
CAD-Based Design Optimization Techniques For Sustainable Manufacturing
Authors: Ms. Amisha Malviya, Mr. Rahul Khobragade, Mr. Rahul Ghotkar
Abstract: The increasing demand for environmentally responsible industrial production has accelerated the adoption of sustainable manufacturing practices across various engineering sectors. Modern manufacturing industries are continuously seeking innovative approaches to minimize material consumption, reduce energy usage, lower production costs, and improve product quality without compromising functional performance. Among the emerging technological solutions, Computer-Aided Design (CAD) integrated with design optimization techniques has become a fundamental component in achieving sustainable manufacturing objectives. CAD enables engineers to create accurate three-dimensional digital models, perform virtual simulations, evaluate multiple design alternatives, and optimize product structures before physical manufacturing begins. This digital-first approach significantly reduces material waste, shortens product development cycles, mini-mizes production errors, and supports environmentally friendly manufacturing processes. Conse-quently, CAD-based design optimization has become an essential methodology for developing high-performance products while promoting economic, environmental, and social sustainability. Sustainable manufacturing aims to balance industrial productivity with environmental conserva-tion by utilizing resources efficiently throughout the product life cycle. Traditional product devel-opment methods often relied on repeated physical prototyping, trial-and-error experimentation, and manual design modifications, resulting in increased material waste, excessive energy con-sumption, longer production times, and higher manufacturing costs. In contrast, modern CAD software integrated with optimization algorithms enables designers to analyse product perfor-mance virtually, identify inefficient design elements, and implement improvements during the early stages of product development. This proactive engineering approach reduces unnecessary manufacturing operations and supports sustainable resource utilization across diverse industrial applications including automotive, aerospace, biomedical engineering, consumer electronics, renewable energy systems, and advanced mechanical manufacturing. Recent technological ad-vancements have transformed CAD from a simple drafting tool into an intelligent engineering platform capable of integrating simulation, optimization, artificial intelligence, cloud computing, digital twins, and additive manufacturing technologies. Advanced CAD environments incorporate Finite Element Analysis (FEA), Computational Fluid Dynamics (CFD), topology optimization, parametric modelling, generative design, multi-objective optimization, and life cycle assessment tools to evaluate product performance under various operating conditions. These integrated capa-bilities enable engineers to optimize product geometry, structural integrity, thermal performance, manufacturability, and environmental impact simultaneously. As a result, sustainable product development has become more efficient, accurate, and cost-effective than traditional engineering approaches.
Beyond Accuracy: A Lifecycle-Integrated Ethical Evaluation Framework for NLP and Deep Learning Systems
Authors: Battula Sravani, Dr. Alugolu Avinash
Abstract: Recent breakthroughs in Natural Language Processing (NLP) and deep learning have enabled applications such as intelligent assistants and real-time content generation. However, with these applications being increasingly woven into the fabric of our daily lives, we must ensure that they are being used responsibly and ethically. Contemporary ethical issues are significantly more complex than traditional concerns, such as transparency in decision-making, protection of person-al data, environmental sustainability of large- scale models, and safe use in social settings. This paper discusses the integration of ethical principles into the development of NLP systems. This Paper discuss how interpretable and privacy-respecting models can be developed, how negative impacts can be mitigated, and how accountability in automated decision-making can be ensured. Moreover, we propose a framework to help developers in the application of these principles throughout the AI development lifecycle, from data acquisition to model evaluation and deploy-ment. By combining ethical principles and technological advancements, this research aims to ensure the development of NLP systems that are both extremely effective and socially responsi-ble. This paper offers real-world guidance to researchers and developers, urging them to develop NLP technologies for the betterment of humanity while ensuring safety, trust, and positive social impact.
AI-Driven Financial Fraud Detection Framework For Digital Payment Ecosystems
Authors: Dr. Gundupagi Manjunath, Dr. N Chandan Prashad
Abstract: The increasing digitization of financial services has led to the transformation of payment systems through improved convenience but also increased vulnerability to fraudulent attacks. Rule-based approaches to detecting such threats, limited by their static nature and preconceived threshold limits, prove inadequate to combat evolving threat models. In this paper, we propose a new ap-proach to fraud detection through artificial intelligence that utilizes ensemble machine learning and deep neural networks to create a zero-trust architecture security mesh. Our system utilizes Ran-dom Forest and XGBoost classifiers along with LSTM sequence aware networks and adaptive learning capability for addressing concept drift. The experiments conducted on our algorithm based on the PaySim synthetic transaction data set demonstrate significantly better fraud detection performance, with an accuracy of 99.96%, precision of 91.84%, and recall of 89.12%. Our ap-proach lowers false positive rates by 21% compared to traditional gradient boosting baselines while sustaining sub-120ms inference time.
Turbocharger Failure in Diesel Engines: A Review of Failure Mechanisms and Operating- Environment Factors with Reference to Kerala
Authors: Premdas T. P., Dr. Rajeesh C. R.
Abstract: The turbocharger is a critical component of modern diesel powertrains, enabling higher specific power output, improved fuel economy, and reduced emissions by forcing pressurised air into the combustion chamber. Because it operates at extremely high rotational speeds and thermal loads while depending on a thin, uninterrupted film of lubricating oil, the turbocharger is also one of the more failure-prone auxiliaries in an engine, and its breakdown often causes disproportionate downtime and repair cost relative to its size. Six cases attended in Consumer Dispute Redressal Commissions, Kalpetta, Wayanad, related to vehicle disputes in connection with car engine sei-zure due to turbocharger failure motivate this review. The paper reviews the published literature on turbocharger failure mechanisms — lubrication-related wear, foreign object and particulate damage, thermal and overspeed failures, and fuel- or oil-contamination-induced degradation — drawn from automotive, industrial, and marine diesel applications. It then examines operating-environment factors that are particularly relevant to Kerala: a tropical, high-humidity, coastal climate; heavy reliance on road transport (including the state-run KSRTC bus fleet) and in-land/marine diesel propulsion for fishing and inland-waterway vessels; monsoon-season operat-ing conditions; and recurring concerns over diesel fuel quality and adulteration in the wider Indian context. The review finds that while the generic failure modes of turbochargers are well docu-mented internationally, peer-reviewed, region-specific studies quantifying turbocharger failure rates, root causes, and maintenance practices for Kerala's automotive and marine diesel fleet are scarce. The paper synthesises the existing literature into a conceptual framework linking operating environment to failure mode, and identifies this scarcity as a clear research gap that a field-based, Kerala-focused study could address.
Study of Multi-Waste Hybrid Epoxy Composites using Chicken Feather, Banana Fiber and Egg Shells
Authors: Associate Professor P B Bhajantri, Associate Professor M M Ganganallimath, Basavaraj Japannavar, Kalmesh Shiragabaraga, Nagaraj Petnekar, Nandish Jadav
Abstract: The growing accumulation of poultry, agricultural, and food-processing waste has created an urgent need for sustainable pathways that convert such residues into useful engineering materials. This study examines the development of multi-waste hybrid epoxy composites reinforced with three commonly discarded bio-wastes: chicken feather fiber, banana plant fiber, and eggshell particulate. Chicken feather, a keratin-rich fiber byproduct of the poultry industry, offers low density, good thermal insulation, and hydrophobic behaviour; banana fiber, obtained from pseu-do-stem or leaf sheath residue of banana cultivation, provides cellulose-based tensile reinforce-ment; and eggshell, a calcium-carbonate-rich particulate waste, functions as a rigid filler that im-proves stiffness and reduces water uptake. This paper reviews fabrication routes such as hand lay-up, compression moulding, and vacuum-assisted lay-up used to combine these three waste streams within an epoxy matrix, and compares their individual and combined effects on tensile, flexural, impact, hardness, water-absorption, and thermal properties. The economic and environ-mental value of converting these waste materials into automotive, packaging, construction, and consumer-good components is discussed, along with the processing challenges of fiber treatment, filler dispersion, and interfacial bonding. The review concludes that a ternary hybridisation strate-gy-combining a protein-based fiber, a cellulose-based fiber, and a mineral particulate-offers a promising balance of strength, stiffness, and sustainability that no single waste stream can achieve alone.
Design and Development of an Eco – Friendly Solar Powered Grass Cutting Machine
Authors: Associate Professor Kiran D Aswale, Associate Professor M MGanganallimath, Ravikumar Nagarabetta, Prajwalsing Rajaput, Samarth Mali, Prashant M Babali
Abstract: The increasing demand for environmentally sustainable technologies has encouraged the development of solar-powered grass cutting machines as an alternative to conventional petrol- and electricity-operated grass cutters. Traditional grass cutters contrib-ute to fuel consumption, greenhouse gas emissions, noise pollution, and high operating and maintenance costs, creating the need for cleaner and more energy-efficient solutions. This review paper presents a comprehensive analysis of recent studies published between 2020 and 2026 on the design, development, fabrication, and performance evaluation of solar-powered grass cutting machines. The review was conducted through a systematic examination of selected peer-reviewed journal articles focus-ing on machine design, component selection, working principles, fabrication techniques, and performance characteristics. The reviewed studies reveal that most systems employ photovoltaic solar panels, rechargeable batteries, DC motors, rotary cutting blades, and lightweight chassis to achieve efficient, economical, and environmentally friendly grass cutting. The findings demonstrate that solar-powered grass cutters effectively reduce fuel dependency, operating costs, maintenance requirements, and environmental pollution while providing satisfactory cutting performance under suitable solar conditions. However, the review identifies significant research gaps, including limited battery backup, reduced efficiency under low solar irradiance, inadequate optimization of blade and chassis design, and insufficient long-term field performance evaluation. Addressing these challenges can improve system efficiency, reliability, and the practical adoption of sustainable solar-powered grass cutting technologies.
Review of Fuel Cell System Technology: From Stack Design to System Integration
Authors: Adeleye, S. A., Umoru, H. O., Towolawi, A. A
Abstract: The world is fast embracing a cleaner energy through the drive from different fuel cells technology such as proton exchange membrane fuel cells (PEMFCs), solid oxide fuel cells (SOFCs), phosphoric acid fuel cells (PAFCs), alkaline fuel cells (AFCs), and molten carbonate fuel cells (MCFCs), alongside other components technologies such as bipolar plates and membrane electrodes, air compressors, ejectors, high-power modular integration technologies, and control technologies. As a result of this recent advance in fuel cell technologies, it has led to potential applications in aerospace, transportation, and portable and stationary power generation due to high efficiency and low emissions as compared to fossil fuel. Fuel cell type also varies comparatively, based on efficiency, operating temperature, lifetime, energy/power density, and cost. Recent findings show that PEMFCs have the highest mass power density, compared to SOFCs, which makes them suitable for portable applications such as aircraft. PEMFCs and AFCs are suitable for low-temperature applications and are highly efficient. SOFCs and MCFCs are better for high-temperature operations. SOFCs are robust and suitable for high power demands, while MCFCs are advantageous for high-power output. Hydrogen fuel cells promise to decarbonize transportation and aviation sectors with the advantages of lower weight, compactness, and quick startup times. However, challenges remain around renewable hydrogen production/infrastructure and aircraft integration, besides hydrogen storage, water management inside fuel cells, and operational robustness under varying pressures. Generally, for all fuel cell types, more focus should be given to enhancing the stability and efficiency of fuel cell materials and reducing their cost.
Design and Implementation of a Hybrid Explainable AI Frame-work for Intelligent Vlsi Physical Routing Optimization
Authors: Vikas Raturi, M. Zahid Alam
Abstract: Optimizing the physical routing of electronic circuits is one of the hardest problems in semiconductor design, and it has a direct impact on circuit performance, power dissipation and manufacturability. Billion-transistor designs run up against the fundamen-tal computational limits of traditional routing algorithms. We present a new hybrid explainable artificial intelligence (XAI) framework combining graph neural networks (GNNs), attention mechanisms, and reinforcement learning to jointly optimize physical routing while providing interpretability. We conducted our empirical study by evaluating the framework on 47 bench-mark circuits used in both industrial and academic settings, ranging from 10K to 500K transistors. The framework can accom-plish a routing completion rate improvement of 18.7% over the state-of-the-art, with an average reduction of 23.4% in the time utilized for routing. The explanation module integrated into the design of LIME helps designers obtain actionable insights on routing decisions with 94.2% fidelity to used optimization objectives. Our extensive experiments show that this hybrid ap-proach gives a good trade-off between solution quality and computational efficiency whilst making transparency essential for adoption in the industry. The framework demonstrated the ability to perform well despite changes in circuit complexity, differ-ences in design rules, and routing constraints. Statistical analysis shows the approach is beneficial everywhere, from congestion metrics to wire length optimization to design rule compliance. This work connects black-box AI tools to industries that demand explainable and trustworthy optimization tools in VLSI design automation.
Graph Neural Network-Based Financial Fraud Detection in Real-Time Digital Payment Systems
Authors: Assistant Professor Anitha K, Assistant Professor Saranya B
Abstract: With the rise in adoption of real-time payment systems such as UPI, Pix, and FedNow comes a unique set of problems related to detecting fraud within strict latency requirements, and where fraudsters are acting in a networked manner. The conventional methods of rule-based and tabular machine learning work by treating transactions as individual entities and cannot account for the relationships between them and how it plays into the nature of today’s fraud schemes such as mule networks and authorized push payment scams. In this paper, we present a novel approach using graph neural networks to solve the problem of real-time fraud detection by representing the payment network as a dynamic and heterogeneous graph consisting of accounts, devices, merchants, and transaction flows. Our approach combines GraphSAGE for inductive representation learning and LightGBM for tabular feature extraction and can perform sub-millisecond predictions while being able to update in a streaming fashion. Our experiments on real transaction data show improvements over baseline methods in terms of fraud detection recall from 0.71 to 0.84 with the false positives being less than 5%.
Edge AI-Based Smart IoT System for Intelligent Monitoring, Anomaly Detection, and Real-Time Data Analytics
Authors: Assistant Professor Dr. S. Muthukumaran, Assistant Professor Dr. A. Victoria Anand Mary, Assistant Professor J. Antony Daniel Rex, Assistant Professor P. Simon Vasantha Rooban
Abstract: The proliferation of the Internet of Things (IoT) in various fields such as critical infrastructures, health care, and industry has brought about the need for intelligent systems that can process huge volumes of data with low latency. In this paper, we pro-pose an Edge AI based intelligent IoT framework which utilizes lightweight deep learning models coupled with distributed edge computing for real-time monitoring, fault detection, and analysis of data. In this work, we use a hybrid CNN-LSTM model on resource-constrained edge devices such as Raspberry Pi and NVIDIA Jetson Nano. Our experimental results show that the proposed system is able to detect faults with an accuracy of 92.0%, with response time below 150ms, a decrease in latency of 52% and a reduction of bandwidth utilization by 38% as compared to the cloud approach. We obtain a high precision of 91.9% and F1 score of 90.8% in physiological time series data using federated learning.
Artificial Intelligence and Innovation: Transforming Modern Entrepreneurial Ventures
Authors: Assistant Professor Dr Sandeep R.S
Abstract: The incorporation of Artificial Intelligence (AI) into entrepreneurial ventures is a paradigm shift in entrepreneurship and how new ventures are started, developed, and scaled. The aim of this paper is to analyze the role of AI in transforming different aspects of entrepreneurial ventures from recognizing opportunities through to strategy implementation. From a review of recent studies (2021 – 2026) and through quantitative adoption analysis, it has been established that AI is not only an operational tool that is incorporated to help in increasing operational efficiency but is also used to change the way business models operate. The results indicate that the use of G-AI in ventures increases venture viability by 11%, where there is complementary rather than substitution between human expertise and AI.
Privacy-Preserving Machine Learning in Distributed Systems
Authors: Assistant Professor V.Priyanandhini, Assistant Professor Vandana M
Abstract: Distributed machine learning frameworks encounter important privacy risks owing to the presence of data and model parame-ters across untrusted networks and nodes who may have malicious intents. In this paper, we describe our privacy-preserving framework which combines label retention on the master node, feature obfuscation using Lagrange interpolation based encod-ing, and differential privacy noise injection. Our methodology deals with privacy risks arising out of honest but curious work-ers, external snooping attacks, and heterogeneity problems like stragglers in a single solution. Our experimental analysis con-ducted on MNIST, CIFAR-10, Fashion MNIST, and simulated health care databases indicates that our framework offers com-parable accuracy within 1.6 percent reduction against non-private baseline solutions, besides ensuring strong privacy. It com-pares favourably with federated learning along with differential privacy in terms of 2-3 percent gain in accuracy and outpaces federated learning solutions based on homomorphic encryption in terms of computation time. The framework offers stable results under up to 40 percent straggler worker conditions.
Design And Development Of Mechatronic System For Automatic Blackboard Cleaning
Authors: Dr. M C Goudar, M MGanganallimath, Abhilash M Mernal, Doddanna M Hugar, Mustaf Kolhar, Vikas P Rathod
Abstract: The conventional method of cleaning blackboards is time-consuming and often produces considerable chalk dust, which can affect classroom cleanliness and the health of students and teachers. This project presents the design and development of a mechatronic system for automatic blackboard cleaning that integrates mechanical, electrical, and control components to automate the cleaning process. The proposed system consists of a motordriven cleaning mechanism, suitable wiping material, guide rails, sensors, and a microcontroller-based control unit. When activated, the system moves the cleaning mechanism systematically across the blackboard surface to remove chalk marks with minimum human effort. The design aims to provide uniform clean-ing, reduce dust generation, improve operational efficiency, and minimize cleaning time. The developed prototype is evaluated based on cleaning effectiveness, operating time, power consumption, and reliability. The proposed system offers a simple, economical, and user-friendly solution for educational institutions and demonstrates the practical application of mechatronics and automation in classroom environments.
DAHT: A Domain-Aware Hierarchical Transformer For Longitu-dinal Android Malware Detection Under Concept Drift
Authors: Ms.L.Catherine Lemoria, Dr.T. Miranda Lakshmi
Abstract: Android malware detection has achieved strong results on many benchmark datasets, but those results are often obtained from random train-test splits that mix applications from different periods. Such a split is useful for measuring ordinary predictive generalization, but it does not answer a practical question: how well does a detector trained on older applications work when it encounters applications released later? This paper studies that question through a Domain-Aware Hierarchical Transformer (DAHT) for static Android malware detection. DAHT converts a 4,561-dimensional feature vector into 32 learned tokens, projects them into a 64-dimensional embedding space, and processes them with a compact two-layer, four-head Transformer encoder before global average pooling and binary classification. The evaluation uses the Longitudinal Android Malware Dataset (LAMDA) and a strict forward-in-time protocol. Both DAHT and an XGBoost baseline are trained once on 2013, 2014, 2016, 2017, and 2018 data and then evaluated, without retraining, on 2019, 2021, 2022, 2023, and 2024. The results do not show universal superiority: XGBoost has higher accuracy and macro-F1 in four of the five test years. DAHT leads in 2024, reaching 99.50% accuracy and 0.6654 macro-F1, compared with 98.75% and 0.5802 for XGBoost, while DAHT records a ROC-AUC of 0.9139. The findings therefore support a narrower conclusion. DAHT is a useful architecture to examine under temporal distribution shift, but its advantage is conditional rather than general. The study also shows why longitudinal evaluation and class-balanced metrics are important when assessing malware detectors
Software Defect Prediction Using Machine Learning: A Comparative Analysis of Ensemble and Neural Classification Algorithms
Authors: Arjun Singh Tomar, Aashish Kumar Tiwari
Abstract: Ensemble and neural classifiers are widely believed to outperform single-model classifiers for software defect prediction, but the size of that advantage, and the training-data volume required to realise it, are not always quantified in comparative studies. This paper benchmarks six such models — Bagging, Random Forest, AdaBoost, Gradient Boosting, XGBoost, and a shallow multilayer perceptron — on static code metrics pooled from four NASA/PROMISE datasets, using a common preprocessing pipeline with SMOTE-based training-fold resampling. XGBoost achieved the best overall accuracy, precision, recall, and F1-score (89.4% accuracy, 85.1% F1-score) while training faster than Gradient Boosting, and a learning-curve analysis shows that its advantage over Random Forest grows with training set size rather than being constant across data volumes. The multilayer perceptron did not outperform the strongest tree ensembles at the data volumes examined. These results suggest that boosting-based tree ensembles remain the most practical choice for defect prediction from static metrics, with the specific choice between Random Forest and XGBoost best guided by the amount of historical defect data a project has accumulated.
