Machine learning is eating the world, but many who are just starting out in the field may be unsure of how to take their first bite. edX, a global learning platform founded by Harvard and MIT, are offering a Professional Certificate in Tiny Machine Learning (tinyML) — a four month, self-paced course culminating in deploying code to an Arm Cortex-M4-based Arduino with various onboard sensors and an attached camera. If you're just starting out with machine learning and microcontrollers, that might sound like a load of gobbledygook — so thankfully, Arduino has put together a "starter pack" with everything you need in one box: the Arduino Tiny Machine Learning Kit.
The kit is centered around the venerable Arduino Nano 33 BLE Sense, which already comes equipped with a bevvy of sensors, including a 9-axis IMU for measuring motion, humidity, temperature and barometric sensors for measuring your environment, a light color and intensity sensor, and gesture and proximity sensors. The board also packs a microphone, ideal for voice applications and wakeword detection, but it lacks one critical feature required to run the full gamut of TensorFlow Lite Micro examples: a camera! Fortunately, Arduino has bundled an OV7675 CMOS VGA (640x480) Image Sensor with the kit, and a Tiny Machine Learning Shield to interface with the Nano.
- Arduino Nano 33 BLE Sense board
- OV7675 Camera
- Arduino Tiny Machine Learning Shield
- USB A to Micro USB Cable
The Tiny Machine Learning Kit starts shipping in January 2021, and is available for pre-order directly from Arduino for just $49.99.