Explore federated machine learning with Python in this 44-minute EuroPython 2021 conference talk. Delve into privacy-preserving AI techniques that allow model training without compromising data privacy. Learn about horizontal vs. vertical federated ML, cross-device/cross-silo training, and centralized vs. decentralized approaches. Discover existing FedML implementations in Python and witness a live code walkthrough of a minimal centralized federated learning system using python-socketio and confluent kafka. Examine the challenges enterprises face when implementing federated ML, including non-IID data, network latencies, and computational limitations. Gain insights into this emerging field ripe with research and development opportunities, and understand its potential impact on data privacy and AI collaboration.
Overview
Syllabus
Introduction
Presentation Overview
Data Privacy
PrivacyPreserving Machine Learning
Edge Device Use Cases
Enterprise Use Cases
Horizontal Federated Learning
Vertical Federated Learning
CrossDevice Federated Learning
CrossSilo Federated Learning
Centralized Federated Learning
Decentralized Federated Learning
Federated aggregation
Ingredients
Socketio
Recipe
Updates
Client
Train Model
End Session
Demo
Problems with Federated Learning
Taught by
EuroPython Conference
Related Courses
-
Federated Learning - Machine Learning on Decentralized Data
5.0 -
Practical Privacy-Preserving Machine Learning in Python
-
AI Singapore's Journey into the World of Federated Learning
-
Federated Learning with Formal User-Level Differential Privacy Guarantees
-
Protect Privacy in a Data-Driven World - Privacy-Preserving Machine Learning
