For a science fair in 2019 I helped two of my children build an instructables.com project using matchboxes and colored beads.
The project is based on the matchbox computer described by Martin Garner.
In his article Martin explains that learning machines are computers that improve with experience.
This learning machine improves over time through reinforcement learning.
Reinforcement learning is a training method based on rewarding desired behaviors and/or punishing undesired ones (by removing a colored bead in this case).
Matchbox Learning Machine
For the same science fair I helped my children to configure a Raspberry Pi to track objects using the YOLO v3 Tiny model.
Details of the Raspberry Pi setup are provided in Xiang Zhai's article Deep Learning with Raspberry Pi -- Real-time object detection with YOLO v3 Tiny!
The YOLO v3 Tiny model was trained using Supervised learning. A labelled set of photographs was used to train the model.
The people at the science fair were less interested in the computer vision demonstration.
YOLO
