Machine Learning on Android Demystified
summary
Compares cloud inference versus on‑device inference and shows concrete paths: use TensorFlow Lite for local models and Firebase ML Kit to host and auto‑update models with Wi‑Fi/charging download conditions; includes Kotlin snippets for loading a .tflite and running an Interpreter. Demonstrates quantization (8‑bit) that cuts model size and latency at a measurable cost in accuracy, mentions an optional GPU delegate for complex nets, Smart Reply limits (English, up to ten messages), and translation models ≈3 MB.
AI-written summary · about
chapters (30)
- 0:00Intro
- 1:35Intro To Machine Learning Machine learning is the area of computational science that focuses on analyzing and interpreting patterns and structures in data. It is about learning from data and getting better at predicting results overtime.
- 2:44Can my task be solved with ML?
- 4:24Overfitting
- 4:54Bad use cases for ML
- 7:16Machine Learning Process
- 9:20Cloud ML: Pros and Cons
- 12:04Mobile Hardware: Designed for Al
- 13:06On-device inference: Advantages
- 14:24On-device inference: Drawbacks
- 16:35Frameworks for ML in Android apps
- 17:42Android Neural Networks API
- 19:43Using TensorFlow Lite SDK
- 22:38TensorFlow Lite Converter: Examples
- 23:20Quantization: Changing precision
- 24:58Quantization: Model Size vs Latency
- 25:32Quantization: Model Size vs Accuracy
- 27:28Optimizing ML models for mobile
- 29:03Firebase MLKit: Overview
- 31:19Translating Text
- 33:43Smart Reply
- 36:28AutoML Vision Edge: Overview
- 38:29AutoML Vision Edge: Usage
- 39:06Using custom models with MLKit
- 40:02Custom Model: Implementation Steps
- 41:07Custom Models: Conditions
- 41:47Custom Models: Inspection
- 42:27Custom Models: Inference
- 42:50Cross-Platform Machine Learning Flutter
- 43:42Summary
related talks
39mMachine Learning on Android Demystified
39mAI/ML on mobile/android with TensorFlow
36mMachine Learning for Mobile Developers
29mFrom TensorFlow to ML Kit: power your Android application with ML
19mOn-device machine learning: TensorFlow on Android
39mMachine Learning on Your Device: The Options
15mSimplified Machine Learning on mobile, browsers, and edge devices with on-device ML
09mMachine Learning on the Edge