Explore hardware acceleration techniques for Machine Learning inference on Android devices in this 21-minute conference talk by Adrien Couque at Paris Android Makers by droidcon 2023. Discover how the increasing processing power of Android devices enables on-device machine learning capabilities previously thought impossible. Learn to leverage specialized hardware such as GPUs, DSPs, and NPUs to accelerate ML processes and enable real-time experiences like augmented reality filters and live speech transcription. Gain insights into using Acceleration delegates and new TensorFlow Lite Android APIs to implement ML hardware acceleration effectively. Delve into the world of TensorFlow Lite on Android and unlock the potential for high-performance machine learning applications on mobile devices.