Embedded Technologies Expo & Conference 2020

Advances in Deep Learning enable new vision-based capabilities. This session shows a fun, practical demo of what can be achieved with the right approach and technologies, even without a Ph.D. in Machine Learning. It starts at the hardware level and then goes on to train a deep neural network and to learn how to optimize it for embedded devices, leveraging Amazon SageMaker Neo. On the device, this session uses Linux and AWS IoT Greengrass to connect with a cloud-based system using MQTT.