AI biweekly: Federated Learning

Dear Artificial Intelligence Enthusiasts,

Companies could work together to build vastly improved customer profiles—if data-privacy concerns didn’t hold them back. Federated Learning (FL) is a way for companies to share insights - without exchanging data - and takes those concerns off the board. To quote my colleague Arun Ravindran , Chief Data Scientist at BCG GAMMA:

“Federated Learning enables companies to share data in a “closed-loop system” to build a common, powerful machine learning model — and do it without actually exchanging data. This single capability may soon enable companies to vastly improve customer insight while addressing such critical issues as data privacy, data security, data-access rights, and access to heterogeneous data.

In this edition of my AI biweekly newsletter, I want to explore FL. Where it is being used, how to set it up, how Machine Learning (ML) does profit from it, and what the hype is all about (for more information, check out the links below):

• Healthcare: Be it predicting clinical outcomes for patients or securing their privacy - Healthcare organizations can control how their datasets are used, while researchers can protect the confidentiality of their models and search queries.
• ?Improving?ML:?Qualcomm has been researching ways ?to make the training of ML algorithms at the edge less energy-intensive, which means it could happen at the edge.
• ?Staying ahead of the game:?FL holds tremendous potential ?to enable companies in virtually any industrial sector to gain insights that are all too often obscured in siloed data. With each passing day, more technology players expand their FL offers. Each day, more new players enter the FL market. And each day, FL continues to demonstrate superior efficacy and security.?

Let me know if you have explored FL and where you see its benefits. I’m curious to hear!

Until next time,

Sylvain Duranton

++++++++++++++++++++++++++++++++++++++++++++++++++++++

Six Links on Federated Learning:

Federated learning for predicting clinical outcomes in patients with COVID-19

A research paper that used data from 20 institutes across the globe to train a FL model, called EXAM (electronic medical record (EMR) chest X-ray AI model), that predicts the future oxygen requirements of symptomatic patients with COVID-19 using inputs of vital signs, laboratory data and chest X-rays - Nature

Enabling AI-driven health advances without sacrificing patient privacy ?

Secure AI Labs (SAIL) is addressing those problems with a technology that lets AI algorithms run on encrypted datasets that never leave the data owner’s system. Health care organizations can control how their datasets are used, while researchers can protect the confidentiality of their models and search queries. Neither party needs to see the data or the model to collaborate. - - MIT News

Qualcomm is researching machine learning at the edge ?

So far, ML at the edge has only involved inference, the process of running incoming data against an existing model to see if it matches. Training the algorithm still takes place in the cloud. But Qualcomm has been researching ways to make the training of ML algorithms at the edge less energy-intensive, which means it could happen at the edge (Read also: Training ML Models at the Edge with Federated Learning on developer.qualcomm.com ) - STACEY ON IOT

Design a federated learning system in seven steps ?

This article is meant to be a guide that will enable you to set up a scalable federated learning system. Because requirements may differ across users and use cases, this guide won’t provide you with all of the answers. However, it should equip you with key questions and considerations to help you design a system that works for you - Towards Data Science

Practical Differentially Private Clustering ?

In this post, the Google AI team considers the following basic problem: Given a database containing several attributes about users, how can one create meaningful user groups and understand their characteristics? Importantly, if the database at hand contains sensitive user attributes, how can one reveal these?group?characteristics without compromising the privacy of?individual?users? - Google AI Blog

Federated Learning: The Next Big Step Ahead for Data Sharing ?

Federated Learning has emerged in the background as a new paradigm for collaboration and partnership between enterprises. It enables companies to share data in a “closed-loop system” to build a common, powerful machine learning model — and do it without actually exchanging data. This single capability may soon enable companies to vastly improve customer insight while addressing such critical issues as data privacy, data security, data-access rights, and access to heterogeneous data. - GAMMAscope