Applications in federated machine learning

Abstract

In our paper we figured out, that federated learning (FL) is a deep learning technique used in various industries, including medicine, agriculture, vehicles, retail and finance. It offers privacy, data ownership, localized model training, bandwidth efficiency, real-time learning, scalability and resilience to device failures. In medicine FL can improve patient’s representation, drug development, medical image analysis, sickness diagnosis and individualized treatment planning. In agriculture, FL can improve crop irrigation, fertilization, harvesting and monitoring animal health. In retail, FL can analyze customer behavior data, preserving privacy. As we understand, federated learning divides model training among local data sources using sensors like GPS, microphones, and cameras. But learning models can be hacked by various threats, including data poisoning attacks.