What is federated learning? | TechTalks

Jirsak/Shutterstock.com

The demand for large amounts of information is one of the most difficult aspects of machine learning. Obtaining training datasets for machine learning models entails confidentiality, security, and computational concerns that businesses would prefer to avoid. “Federated learning” is one strategy that can help with some of these issues. The fundamental concept of federated learning is to use user data to train a machine learning model without having to send the data to cloud servers. The problem is that federated learning isn’t applicable to every machine learning application. If the model is too huge to operate on the user’s device, the developer will have to come up with other ways to protect the user’s privacy.   

Source: https://bdtechtalks.com/2021/08/09/what-is-federated-learning/ 

RELATED ARTICLES

Subscribe to our newsletter