Machine Learning on Android Demystified
Tatyana Casino is a software engineer at WillowTree, a mobile innovation agency. Tatyana has been building mobile apps for a variety of leading brands and clients. She has been developing for Android for 5 years, has done cross-platform development with Xamarin and worked on native iOS projects as well. She is passionate about learning, writing great code and exploring emerging technologies.
Ever wondered what it takes to implement machine learning in your app? Look no further! In this talk I will suggest different ways to approach this. I am going to compare cloud-based services with local (on-device) machine learning, focusing mostly on the latter. On-device predictions are happening strictly on a mobile device, giving us the benefit of keeping the user’s data private and not depending on the network connection. Features like Google Lens Suggestions, Call Screening, Live Caption are all leveraging on-device ML. However, the ML models should be prepared and optimized for efficiency and performance on mobile – I’m going to talk about that as well. For actual implementation, we will look into how to use TensorFlow Lite SDK for pose estimation as an example. Then we will move on to Firebase MLKit Base APIs and using custom models with MLKit. Code examples will be in Kotlin. After attending this talk you will understand the capabilities and limitations of each of these frameworks. You will have a good idea of where to start, what is necessary to implement your ML idea in your app, and what are the potential issues to be aware of.