Introduction to Federated Learning: Leveraging the Flower Framework

Home /
Schedule /
Introduction to Federated Learning: Leveraging the Flower Framework

Introduction to Federated Learning: Leveraging the Flower Framework

Emerging Technologies

Short Talk

Intermediate

By Emeka Obiefuna

ML Engineer

Abstract:

The exponential growth of machine learning (ML) hinges on data, but its acquisition often raises privacy concerns due to the potential exposure of sensitive information. Thus, the challenge lies in balancing the potential benefits of machine learning with the risks of compromising user privacy.

This talk explores Federated Learning (FL), a revolutionary approach that enables machine learning advancements while safeguarding individual privacy by decentralizing the learning process.

Key topics covered:

The Data-privacy Conundrum: Dive into the inherent tension between data-driven ML and individual privacy.
Demystifying Federated Learning: Gain a clear understanding of FL's core principles and how it decentralizes model training across multiple devices while keeping data local.
Practical FL Mechanisms: Explore diverse techniques like federated averaging, client-server architecture, and data partitioning, along with their benefits and trade-offs.
Real-world Applications: Discover how FL empowers organizations in various sectors like healthcare, finance, and social networks to analyze data responsibly.
Challenges and Future Directions: Acknowledge current limitations (for example, communication overhead and model convergence) and delve into emerging research addressing them.

By attending this talk, the audience will:

Understand the foundational principles and practical applications of Federated Learning.
Appreciate its critical role in ensuring responsible and secure ML practices.
Gain a critical perspective on the challenges and future directions of FL in the evolving data landscape.

GO BACK