Running a neural network on a micro-controller might seem absurd, but it’s possible (and has some great uses!). In this workshop, we’ll train a neural network to recognize one of several spoken words, convert it to a TensorFlow Lite model, and load it onto an ARM micro-controller (STM32 Nucleo-L476RG board), where it will listen for and respond to the wake word in real-time. Prior knowledge of machine learning is not necessary for this workshop, but it can be helpful to understand how the neural network operates.
Instructor: Shawn Hymel
Shawn is an electrical and embedded engineer, freelance content creator, and instructor. He and Harris Kenny host the podcast, Hello Blink Show, where they discuss various aspects of starting a business from sales to marketing to hiring. From 2013 to 2018, Shawn worked for SparkFun Electronics designing open-source PCBs, writing firmware, teaching concepts on video, and creating tutorials. Shawn started his own company, Skal Risa, LLC, to help companies create compelling technical content in electronics and embedded systems. He is also an advocate for enriching education through STEM and believes that the best marketing comes from teaching. He can be found giving talks, running workshops, and swing dancing in his free time.
Read the article on Hackaday:
https://hackaday.com/?p=450862
All example code and slides:
https://github.com/ShawnHymel/ei-keyword-spotting
Workshop project page:
https://hackaday.io/project/175078-remoticon-tiny-ml
Music by Rich Hogben