TinyML
Machine learning on microcontrollers and ultra-low-power devices with just a few kilobytes of RAM – AI on a chip smaller than a coin.
TinyML brings machine learning to microcontrollers with kilobytes of RAM – AI inference on a chip smaller than a coin, battery-powered for years.
Explanation
TinyML runs on Cortex-M processors with <256KB RAM and <1mW power consumption. Frameworks like TensorFlow Lite Micro and Edge Impulse enable model deployment. Typical applications: keyword spotting, anomaly detection, gesture recognition.
Marketing Relevance
TinyML enables AI in battery-powered IoT devices, wearables, and sensors – deployment at billions of scale without cloud dependency.
Example
A microcontroller-based sensor detects machine anomalies in the factory via TinyML – runs 2 years on a coin cell battery without cloud connection.
Common Pitfalls
Extreme model constraints (often <100KB), limited frameworks and tooling, debugging on microcontrollers difficult, not suitable for complex tasks.
Origin & History
Pete Warden and Daniel Situnayake coined the term in 2019. TensorFlow Lite Micro and the TinyML Foundation were established in 2019/2020. Edge Impulse (2019) democratized tooling. Arduino Nano 33 BLE became the standard development platform.
Comparisons & Differences
TinyML vs. Edge AI
Edge AI runs on more powerful hardware (smartphones, Jetson); TinyML targets ultra-low-power microcontrollers with kilobytes of RAM.
TinyML vs. On-Device Inference
On-device inference includes smartphones and laptops; TinyML specializes in tiny microcontrollers with extreme resource constraints.