A layman’s introduction to Privacy-Preserving Federated Learning
PART 1: Introduction to Federated Learning
What is Federated Learning?
Federated learning is a technique that enables distributed clients to collaboratively learn a shared machine learning model while keeping their training data localized. For example, individual hospitals can collaboratively train a machine learning model in order to improve their predictive accuracy for a given task.
Federate Learning Workflow
Let us assume that there are several hospitals that want to develop a shared machine learning model. Figure 1 depicts the basic workflow of a federated learning environment. First, each hospital trains a local model, rather than sending raw data to a centralized server. Hospitals act as remote clients and communicate with a central server at regular intervals to learn a global model. For every iteration, the hospitals send their local model to the server. The server computes and sends back the global model, ∆W, to individual hospitals. This process repeats until convergence occurs or some sort of stopping criterion is achieved.
In the next part, I’ll introduce the necessity for privacy-preserving federated learning.
For detailed information please take a look at -
- https://github.com/vaikkunth/PrivacyFL
- PrivacyFL: A simulator for privacy-preserving and secure federated learning (Vaikkunth Mugunthan, Anton Peraire Bueno, Lalana Kagal) https://arxiv.org/abs/2002.08423
