Practice with the Machine Learning Co-Processor

We are interested in building an Arduino + Machine Learning project for an upcoming Arduino Day (details to be announced).

Part of what we want is to be able to process large amounts of streaming information and interpret that data through the filter of a neural network. While any computer could do this in theory, adding a Co-Processor should speed things up a lot.

While normally co-processors are used for video feeds, let’s see if we can stream Arduino-based sensor data instead. The first part, however, will be working with the Google Coral to see how it want’s it’s data formatted.

Let’s see if we can train a basic neural network, and push it to the Google Coral – for on-going real-time classifications.