Authors - Shiva Kumar Bandaru, Upendra Pratap Singh Abstract - In a federated learning-based setup, parameter aggregation plays a pivotal role in obtaining global parameter estimates that assimilate the knowledge learned by the different clients. With an efficient parameter aggregation strategy, the global parameter estimates derived are more generalizable, accelerating the local client training in the subsequent communication rounds. In the proposed approach, we propose a novel m-ary improvisation-based parameter aggregation algorithm to obtain the global parameters. Specifically, after a threshold number of communication rounds has elapsed, the performance of the clients is evaluated on an independent test set, and the clients with better generalization are labeled as strong and do not participate in the next set of a threshold number of communication rounds. In this way, weak clients participate in the federated learning for more communication rounds; after the next set of threshold communication rounds has elapsed, the clients undergo a similar evaluation to be labeled as strong or weak again. The proposed algorithm ensures weak clients get more attention/exposure to learn the model parameters collaboratively. The global model trained on the BraTS2020 dataset in a federated learning-based framework reports the Dice coefficient, Jaccard index and pixel accuracy values of 0.8851, 0.8965, and 99.92%, respectively. Further, we show empirically that the training time for the different clients reduces from 180 minutes in the first phase of federated learning to only 64.8 minutes in the last phase, highlighting an accelerated training process. Consequently, the results reported by the proposed federated learning-based segmentation model highlight its usability for efficiently carrying out brain segmentation involving private and sensitive brain scans.
Authors - Saurav Kumar, Shivani, Rashmy Moray, Shikha Jain, Sridevi Chennamsetti Abstract - The aim of the study is to inspect the factors determining the use of web 3.0 Meta based banking services. Diffusion of innovation theory has been used to explain the influence of perceived factors on attitude and behavioural intention to use the meta based banking services. Structured questionnaire as primary source of data collection has been applied and data gathered was analyzed using Structural equation model as statistical technique to achieve the stated objectives. SmartPLS as statistical tool was employed in analyzing the data and the outcome reveal that compatibility, observability and trialability showed a significant impact on attitude towards usage intent of Web 3.0 based meta banking services. The study has proved to be significant in the field of banking on metaverse for various stake holders and policy makers and be helpful to understand the perception of the customers in the usage of web 3.0 based banking.
Authors - Shubham Kishor Kadam, Pankajkumar Anawade, Deepak Sharma, Anurag Luharia Abstract - Artificial Intelligence (AI) may be defined as utilization of computer systems in undertaking processes, which are typical of human intelligence. AI is an incomparably new and actively developing scientific direction, which can qualitatively change most of the social processes. In the context of the increased usage of AI, the different educational settings are applying this technology to create new perspectives in the sphere of pedagogy nowadays. Today it is utilized to sift through incalculable quantities of information in order to discover patters, which would help devise better and more appropriate policies and educational strategies than the existing ones. This paper determine the pertinence of the AI in consideration of education along with the challenges using AI in education.
Authors - Harishh N, Drisya Murali, Suresh M Abstract - The study explores the possibilities of green logistics and the adoption of biodegradable packaging in freight transportation, focusing on the impact on reducing packaging waste and bringing in sustainability. The research uses the Grey Influence Analysis (GINA) methodology to analyze the identified eleven significant factors, which impact the adoption of biodegradable packaging in freight transportation. The primary role of packaging is to protect products during storage and transport, reduce costs, and sustainable way of product distribution and safety. The study also highlights the importance of improving the material properties of packaging, which can mitigate or minimize adverse environmental impacts. The study's findings highlight the need for various perspectives in future studies and the need for a comprehensive understanding of the relationship between various factors influencing biodegradable packaging in freight transportation.
Authors - Utkarsha Wanjari, Shubham Kadam, Chhitij Raj, Pankajkumar Anawade, Deepak Sharma Abstract - The digital divide continues to be a global issue since it accounts for the marginalization between the group owning access to Information and Communication Technology (ICT) and those without access. This report looks at the crucial role of ICT in bridging this gap and ensuring integral social and economic development. ICT does hold tremendous transforming potential through its power to enrich education, modify healthcare delivery systems, and strengthen governance through digital inclusion. Economically, it propels innovation, expands access to global markets, and creates financial inclusion through digital tools. Though still highly significant, challenges persist in the form of infrastructure deficits, digital literacy gaps, and socioeconomic inequalities. Through case study examples and successful global initiatives, this report is shaped by best practices and strategies to work around these challenges. It draws attention to public-private partnership efforts, policy reform, and investment in ICT infrastructure and ICT training. Bridging the digital divide is not just technical but also a pathway to achieving equitable and sustainable development in an increasingly digitalizing world.
Authors - Vasudha V. Ayyannavar, Lokesh B. Bhajantri Abstract - The healthcare sector is rapidly evolving, making the continuous exchange of healthcare data essential for both patient care and maintaining operational efficiency. In today’s landscape, file and data synchronization is no longer optional but a crucial requirement. This work presents a real-time data synchronization system tailored for hospital records management, enabling seamless and secure communication among healthcare users. The system uses real-time synchronization to ensure that updates made on the server are instantly reflected across all connected clients. In this work, a robust architecture is developed to support both MySQL and MongoDB databases, offering flexible data storage. It associates with Node.js and Express.js, utilizing Socket Input and Output for real-time and bidirectional communications. On the front end, HTML, CSS, and JavaScript are combined with Bootstrap to create a responsive and user-friendly interface, allowing easy data input and retrieval by healthcare users. The proposed solution ensures conflict-free data dissemination across various devices and is compared against existing methods, analyzing key metrics such as synchronization time, memory usage, and data accuracy. Overall, the system aims to enhance hospital records management through a reliable, scalable, and intuitive real-time synchronization solution.
Authors - Ganesh Haricharan Mungara, Pranai Govind Soorneedi, Karthik Mungara, C.N.S.Vinoth Kumar Abstract - The proliferation of smartphones has transformed communication, work, and information access. However, this convenience has brought significant security challenges, particularly from malware that can compromise user data and privacy. Despite numerous antivirus applications, detecting and removing malware from Android devices remains a challenge. Current solutions of ten fail to detect sophisticated malware, necessitating the intervention of cyber security experts, which can compromise user privacy. This project aims to develop a tool that detects malware on Android devices based on installed applications, eliminating the need for users to install third-party software. The proposed solution leverages pattern matching by checking installed packages against a database of known malware. If a match is found, the tool indicates potential malware presence. This method offers a privacy-preserving approach, focusing on app behavior rather than relying solely on signatures, making it harder for malware to evade detection. The tool addresses the limitations of existing antivirus solutions, which often require extensive permissions and access to personal data. By providing a user-friendly interface and ensuring privacy, this project aims to enhance the overall security of Android devices. Future enhancements include incorporating machine learning models to improve detection accuracy and expanding the tool to other mobile platforms like iOS. This innovative approach offers a reliable and privacy-focused alternative for malware detection on Android devices.
Authors - Pratibha Verma, Sanat Kumar Sahu, Latika Tamrakar Abstract - Coronary Artery Disease (CAD) is a major crisis midst populace worldwide. So, we prerequisite a system that is effective for the identification of CAD problems. In this study we formed a model substance on the classification technique that can clarification the problem of CAD. The Ensemble Bagging classification method develops the creation of multiple classifier models and their mutual outputs to achieve a unified classification outcome. This technique has been implemented in the field of CAD using Artificial Neural Network (ANN) models. The ANN based models are Multi-layer Perceptron Network (MLPN or MLP), Radial Basis Function Network (RBFN), ensemble bagging –RBFN (EB-RBFN), and ensemble bagging MLP (EB-MLP). Our experimental outcomes indicate that the anticipated ensemble bagging model suggestively enhances dataset classification accuracy when compared to individual MLP and RBFN classifiers. This ensemble model consistently delivers more accurate and valuable classification results. Its implementation substantially improves CAD diagnostic accuracy, enabling the more precise identification of patients affected by this condition. These findings imply that the utilization of ensemble learning techniques, specifically ensemble bagging with ANN models, holds great potential in enhancing the precision of CAD diagnosis. This advancement has the potential to improve patient management and treatment outcomes.
Authors - Pampati Sreya, Yashaswi D, Stephen R, Gobinath R, Ramkumar S Abstract - Predicting stock prices remains a challenging problem due to the highly dynamic and non-linear nature of financial markets. Traditional statistical models like ARIMA and GARCH often fail to capture the complexities inherent in stock market data. This paper investigates the use of deep learning techniques, focusing on Convolutional Neural Networks (CNNs) and a hybrid CNN-LSTM ensemble model for stock price prediction in the Indian stock market. The CNN model efficiently extracts temporal patterns from sequential data, while the CNN-LSTM ensemble leverages temporal dependencies for improved long-term prediction accuracy. Historical data from Tata Motors, spanning over two decades, was used to train and evaluate the models. Experimental results highlight the CNN-LSTM ensemble's superior performance in capturing volatile trends and long-term dependencies, with a notable decrease in test loss compared to standalone CNN. This study underscores the effectiveness of hybrid deep learning architectures in enhancing prediction reliability, paving the way for more adaptive and robust financial forecasting systems.
Authors - Mohmed Umar, Jeevakala Siva Rama Krishna Abstract - In the era of complete digital connectivity, it is the need of the hour to keep the networks safe from a wide range of cyberattacks. Traditional Network Intrusion Detection Systems (NIDS) rely mainly on signature-based approaches; though highly efficient in identifying known threats, they suffer from weaknesses in discovering new and developing attacks, such as zero-day vulnerabilities. This results in higher false positives and lower detection efficiency. We present a novel NIDS based on the ensemble methods in machine learning, namely Random Forest and Bagging Classifiers, with which we may promise detection accuracy at the cost of a reduced level of false alarms. We conduct extensive evaluations based on systematic data preprocessing, feature selection, and model training against benchmark datasets like KDD Cup 99 and NSL-KDD. The system being considered achieves a detection accuracy of 99.81%, along with an F1 score of 99.82% and an AUC score of 99.81%, thus significantly surpassing the performance from traditional approaches. These results show the aptness of machine learning methodologies in enhancing network security, as it makes for a flexible and scalable solution suited for real-time deployment in extensive environments. Future work will focus on further developing the scalability of the system and minimizing latency to ensure seamless real-time operation.
