What is TinyML?
• Tiny machine learning (tinyML) = ML on IoT devices
• TinyML is the ML applied on edge/local IoT devices (like on a fitness band), instead of processing in the cloud
Why TinyML?
• Edge nodes generally have low computing capabilities so traditional ML algorithms needed a makeover (compressed + more efficient with less code)
• TinyML also solves the issues of data security, privacy, latency, storage, and energy efficiency
Examples
• Edge microprocessors to process TinyML: Arduino, NodeMCU, ESP8266, Raspberry Pi, etc.
• Software to make TinyML models: Google’s TensorFlow Lite
Use cases
• Real-time healthtech
• Livestock wearables
• Emergency response
• Remote monitoring
• AR glasses