From the course: Machine Learning in Mobile Applications

Unlock the full course today

Join today to access over 23,400 courses taught by industry experts.

Server-side vs. client-side ML

Server-side vs. client-side ML

From the course: Machine Learning in Mobile Applications

Start my 1-month free trial Buy for my team

Server-side vs. client-side ML

“

There are two basic architectures mobile applications use to work with machine learning. The first is the server-side architecture. In this case, all of the work is done on the server from creating the model to live evaluation of results. The typical methods to communicate with the server-side model is your web service that the application calls to send information they want evaluated, and then to get a result back. The information to be evaluated may be some text, an image, recording, almost anything. It is normally the job of the application to find out if the result is correct, and potentially use a web service endpoint to give feedback on the result's accuracy, with possible model updates. If that happens, the next request for the evaluation could be using an updated or smarter model. The second architecture is the client-side model. Like the server-side architecture, the work of creating and processing any updates to the model is done server-side. The mobile clients just don't…

Contents