Yazar "Kumari, R." seçeneğine göre listele

Listeleniyor 1 - 2 / 2
  • Küçük Resim Yok
    Öğe
    Advancing Medical Recommendations With Federated Learning on Decentralized Data: A Roadmap for Implementation
    (Institute of Electrical and Electronics Engineers Inc., 2023) Kumari, R.; Sah, D.K.; Gupta, S.; Cengiz, K.; Ivkovic, N.
    This proposal presents a road-map for implementing federated learning for personalized medical recommendations on decentralized data. Federated learning is a privacy-preserving technique allowing multiple parties to train machine learning models collaboratively without sharing their data. Our proposed framework incorporates differential privacy techniques to protect patient privacy. We discuss several evaluation metrics, including KL divergence, fairness, confidence intervals, top-N hit rate, sensitivity analysis, and novelty to evaluate the performance of the federated learning system.These metrics collectively serve as a robust toolbox for assessing the performance of the federated learning system. The proposed framework and evaluation metrics can provide valuable insights into the system’s effectiveness and guide the selection of optimal hyperparameters and model architectures. Our framework incorporates differential privacy methods to safeguard patient information effectively. IEEE
  • Küçük Resim Yok
    Öğe
    Optimizing Resource Utilization Using Vector Databases in Green Internet of Things
    (Institute of Electrical and Electronics Engineers Inc., 2023) Kumari, R.; Sah, D.K.; Cengiz, K.; Nauman, A.; Ivkovi?, N.; Mihaljevi?, I.
    With the rapid proliferation of Internet of Things (IoT) devices and the ever-increasing volume of sensor data, optimizing resource utilization has become crucial for building sustainable and efficient IoT systems. In this study, we propose a novel approach for optimizing resource utilization in Green IoT through efficient storage and retrieval in vector databases. Our approach leverages specialized data structures, including k-d trees and ball trees, to achieve improved storage efficiency and accelerated retrieval performance for high-dimensional sensor data. We conducted extensive experiments to evaluate the effectiveness of our proposal, comparing it with traditional database approaches. The results demonstrate significant improvements in storage efficiency, with vector databases requiring considerably less storage space compared to traditional databases. Moreover, our approach enables fast and accurate retrieval of high-dimensional sensor data, reducing query times and enhancing real-time data analysis and decision-making capabilities. The technical achievements of our proposal offer promising prospects for the development of sustainable and efficient IoT systems in various domains, such as environmental monitoring, healthcare, and smart cities. Our work contributes to advancing the field of Green IoT by addressing the challenges of resource utilization and query performance through efficient storage and retrieval in vector databases. © 2023 IEEE.