AI4EOSC webinar series (3): Introduction to federated learning

Monday Apr 22, 2024, 2:00 PM → 4:00 PM Europe/Madrid

Judith Sainz-Pardo Diaz (IFCA), Marcin Plociennik (PSNC), Álvaro López García (IFCA)

Join us for an insightful exploration of "Introduction to Federated Learning," where we unravel the potential of this cutting-edge approach in distributed machine learning. Across four dynamic sessions, we'll cover everything from the basics and best practices to understanding threats and exploring friendly frameworks like Flower. Dive into the world of federated learning and discover its applications within the AI4EOSC platform through live demonstrations and engaging discussions. Whether you're new to federated learning or seeking to deepen your understanding, this workshop provides a comprehensive overview to empower your journey into this exciting field.

 

All presentations will contain a questions & answers section at the end.

 

The webinar will be live streamed, starting at 14:00 CEST  on YouTube in the following link: https://www.youtube.com/watch?v=v7blpOB92C8 

 

Q&A: https://app.sli.do/event/n26byMswE186YECEuHcLLi/live/questions 

2:00 PM → 2:30 PM Basics of federated learning. Tips and tricks

Begin your journey into federated learning with a comprehensive overview of the basics, accompanied by invaluable tips and tricks for success. Explore the principles behind federated learning, including its decentralized nature and collaborative model, and gain practical insights into optimizing model performance and mitigating common challenges. Whether you're a novice or seasoned practitioner, this session equips you with the foundational knowledge and strategies needed to excel in federated learning.

Speaker: Khadijeh Alibabaei (KIT)

2:30 PM → 3:00 PM Understanding threats in federated learning

Enter into the complexities of federated learning security as we uncover potential threats and vulnerabilities inherent in decentralized machine learning environments. From data privacy concerns to examples of privacy attacks by malicious adversaries, this session explores the challenges of securing federated learning systems. Gain insights into how to adequately measure privacy for safeguarding sensitive data and acquire a deeper understanding of the privacy threats present in federated learning.

Speaker: Alberto Pedrouzo-Ulloa (atlanTTic Research Center, Universidade de Vigo)

3:00 PM → 3:30 PM Flower: a friendly federated learning framework

Discover Flower, a user-friendly federated learning framework designed to simplify the development and deployment of distributed machine learning models. Learn about Flower's architecture, features, and capabilities, and explore how it streamlines the implementation of federated learning algorithms across diverse environments.

Speaker: Javier Fernandez-Marques (Flower Labs)

3:30 PM → 4:00 PM Federated learning in the AI4EOSC platform (demo)

Experience federated learning in action with a live demonstration showcasing its integration within the AI4EOSC platform. Explore real-world use cases and scenarios where federated learning enhances data collaboration, privacy preservation, and model scalability within the European Open Science Cloud (EOSC) ecosystem. Through hands-on examples and demonstrations, uncover how Flower and AI4EOSC empower researchers and developers to unlock the full potential of federated learning in collaborative settings.

Speaker: Judith Sainz-Pardo Diaz (CSIC)