AI/ML on mobile/android with TensorFlow
summary
Practical roadmap for running neural nets on Android: use TensorFlow Lite with hardware delegates for speed, apply quantization to shrink models and lower power at modest accuracy cost, and use transfer learning plus data augmentation (≈300 photos can generate thousands) to avoid training from scratch. AutoML/Neural Architecture Search can produce mobile-optimized models quickly (orders-of-magnitude time savings). Also mentions on-device encrypted inference schemes to protect model and data, and real-time object recognition demos.
AI-written summary · about
chapters (9)
- 0:00Intro
- 2:39Deep Learning
- 16:47Machine Learning
- 21:32Visual AI
- 27:38Deployment
- 31:24Machine Learning vs Deep Learning
- 33:01Security
- 34:56Landscape
- 36:02Manufacturing
related talks
39mMachine Learning on Android Demystified
47mMachine Learning on Android Demystified
15mSimplified Machine Learning on mobile, browsers, and edge devices with on-device ML
46mFire Up Your Smartphone Neurons with On-Device Deep Learning
32mAI for Mobile and IoT Devices: TensorFlow Lite
39mCreating smaller, faster, production-worthy mobile machine learning models for Android
23mAI is ready for you. Are you ready for AI?
47mMyths and Reality of Android Machine Learning