Data is the priceless resource driving many of the services we use daily on our devices, thanks to machine learning (ML) algorithms that allows developers to optimize the user's experience based on common user patterns as well as the specific user's behavior. However, training those models requires adapting to a data environment mostly dominated by smartphones and IoT devices which entails concerns about privacy, bandwidth and power limitations.

Federated learning (FL) enables developers to train machine learning models across many devices without centralized data collection, ensuring that only the user has a copy of their data, and is used to power experiences like suggesting next words and expressions in Gboard for Android and improving the quality of smart replies in Android Messages. It works by training the model separately on each device using on-device machine learning, then sending the update to the model, not the data, to the cloud where it is averaged with all the participating devices as an update to the model back at Google's servers.

This approach brings many benefits like increasing model accuracy, reducing latency and decreasing high-bandwidth communications as well as reducing power consumption. Careful scheduling ensures on-device training happens only when the device is idle, plugged in, and on a free wireless connection, so there is no impact on the phone's performance.

Peter Kairouz is a research scientist at Google, where he leads research efforts on distributed, privacy-preserving, and robust machine learning. Prior to joining Google, he was a postdoctoral research fellow at Stanford University, and before that, he was a PhD student at the University of Illinois Urbana-Champaign (UIUC). He is the recipient of the 2012 Roberto Padovani Scholarship from Qualcomm's Research Center, the 2015 ACM SIGMETRICS Best Paper Award, the 2015 Qualcomm Innovation Fellowship Finalist Award, and the 2016 Harold L. Olesen Award for Excellence in Undergraduate Teaching from UIUC.