Machine Learning Models on Mobile Devices

Nowadays, mobile devices have enough computing power to run pre-trained models on them. This results in an optimal use of the hardware and an increase in speed, because the data are not sent over the Internet, which also means more privacy. In my presentation I will show different ways to integrate machine learning models into an iOS and Android application.

Required audience experience

Programming skills in Python and Swift or Java are an advantage, but not mandatory

Objective of the talk

The participants will learn how to integrate a pre-training model into an iOS and Android application with CoreML and Tensorflow Lite, and how to re-train a model for own pictures and use it instead of the pre-trained model. All examples are shown in a small demo.

You can view Lars’ slides below:

Machine Learning Models on Mobile Devices from Lars Gregori

 

Track 2
Location: Euston Date: October 15, 2018 Time: 3:15 pm – 4:00 pm Lars Gregori, SAP CX