From Server-Based to Client-Based Machine Learning

In recent years, mobile devices have gained increasing development with stronger computation capability and larger storage space. Some of the computation-intensive machine learning tasks can now be run on devices. To exploit resources available preserve personal privacy, concept client-based has been proposed. It leverages users' local hardware data to solve sub-problems only uploads results rather than original for optimization global model. Such an architecture not relieve burdens servers but also protect sensitive information. Another benefit is bandwidth reduction because various kinds involved in training process without being uploaded. this article, we provide a literature review progressive from server based client based. We revisit number widely used server-based methods applications. extensively discuss challenges future directions area. believe that survey will give clear overview guidelines applying practice.