slsgd
SLSGD: Secure and Efficient Distributed On-device Machine Learning
Xie, Cong, Koyejo, Sanmi, Gupta, Indranil
Edge devices/IoT such as smart phones, wearable devices, sensors, and smart homes are increasingly generating massive, diverse, and private data. In response, there is a trend towards moving computation, including the training of machinelearning models, from cloud/datacenters to edge devices [1,24]. Ideally, since trained on massive representative data, the resulting models exhibit improved generalization. In this paper, we consider distributed on-device machine learning. The distributed system is a server-worker architecture.