Federated Learning for Privacy-Preserving AI in Cloud Environments: Challenges, Architectures, and Real-World Applications

Authors: Satyam Chauhan

DOI: https://doi.org/10.5281/zenodo.14607851

Short DOI: https://doi.org/g8xxr7

Country: USA

Full-text Research PDF File:   View   |   Download

Abstract: Federated Learning (FL) is a transformative paradigm that enables decentralized AI model training while preserving user data privacy. This paper explores the integration of FL with cloud environments, addressing its role in privacy-preserving artificial intelligence (AI) and its architectural components. A comprehensive analysis of core FL concepts, privacy-enhancing techniques, and cloud-specific considerations such as scalability and resource management are presented. Case studies in finance demonstrate FL’s effectiveness in areas such as credit risk assessment, anti-money laundering (AML), and fraud detection, showcasing significant improvements in model accuracy and compliance with data protection regulations. The study highlights critical challenges, including communication overhead, model heterogeneity, and security threats, and proposes future research directions to address these limitations. By combining theoretical insights with practical applications, this paper underscores the transformative potential of FL in enabling privacy-preserving AI within cloud environments.

Keywords: Anti-Money Laundering, Cloud Computing, Credit Risk Assessment, Cloud Scalability, Decentralized Machine Learning, Differential Privacy, Fraud Detection, Federated Learning, Privacy-Preserving AI, Secure Aggregation, Synthetic Data Generation.

Paper Id: 231982

Published On: 2022-06-02

Published In: Volume 10, Issue 3, May-June 2022

Cite This: Federated Learning for Privacy-Preserving AI in Cloud Environments: Challenges, Architectures, and Real-World Applications - Satyam Chauhan - IJIRMPS Volume 10, Issue 3, May-June 2022. DOI 10.5281/zenodo.14607851

Share this