Recording & Slides

Peter Kairouz, Brendan McMahan, Virginia Smith | Federated Learning Tutorial · SlidesLive

Tutorial Slides

Overview

Federated learning (FL) is a machine learning setting where many clients (e.g. mobile devices or whole organizations) collaboratively train a model under the orchestration of a central server (e.g. service provider), while keeping the training data decentralized. Similarly, federated analytics (FA) allows data scientists to generate analytical insight from the combined information in distributed datasets without requiring data centralization. Federated approaches embody the principles of focused data collection and minimization, and can mitigate many of the systemic privacy risks and costs resulting from traditional, centralized machine learning and data science approaches. Motivated by the explosive growth in federated learning and analytics research, this tutorial will provide a gentle introduction to the area. The focus will be on cross-device federated learning, including deep dives on federated optimization and differentially privacy, but federated analytics and cross-silo federated learning will also be discussed. In addition to optimization and privacy, we will also introduce personalization, robustness, fairness, and systems challenges in the federated setting with an emphasis on open problems.

Goals

Our primary objective is to highlight to attendees the recent advances in FL and the huge intersection between FL and the research interests represented by NeurIPS attendees. We want to demonstrate that FL research is widely interdisciplinary, approachable, and contains many critical problems that would benefit greatly from the expertise of conference attendees. By the end of the tutorial, the attendees will be able to:

1. Compare the difference between various distributed machine learning and federated learning paradigms, and understand the defining characteristics of each one of them.

2. Understand the unique capabilities, constraints, and challenges facing FL and FA.

3. Identify high impact open problems facing FL and FA.

4. Design methods to improve optimization, privacy, robustness, and fairness properties of federated systems.

5. Describe the complex tensions and interplay between data heterogeneity, optimization accuracy, robustness, privacy, and fairness.

6. Apply privacy-preserving technologies (such as data minimization, differential privacy, and secure aggregation) to distributed optimization and machine learning.

Resources

1. Advances and Open Problems in Federated Learning

• sign up to Now Publisher's alert list to get the final version for free when it's published!

2. Federated Learning: Challenges, Methods, and Future Directions

Organizers