AI-on-mobile means faster results and a better, more reliable user experience than is possible by shuffling data back and forth to the cloud.
In this talk I’ll show how TensorFlow and TFLite make deployment to mobile devices easy, and some of the techniques like pruning and quantisation that help keep things lean and efficient when deploying to low-power hardware.
I’ll cover some practical use cases such as image search and recommender systems, and discuss deploying MobileNets in a smartphone app to help users make better recycling choices. We’ll see that by keeping the user central to the workflow we can deliver high-quality results, while also learning from their expertise.
1. Learn about deploying deep neural networks in mobile apps using TensorFlow
2. Understand how to provide a great user experience with AI on mobile
An understanding of the basic principles of machine learning would be helpful, but not essential.