What is it?
Robin is an interactive, perceptron-based robot built using the Arduino IDE and an Arduino R4 WiFi. The project introduces the fundamentals of artificial intelligence by demonstrating how a simple artificial neural network can be trained to make decisions.
How does it work?
At its core, Robin uses a perceptron—the simplest form of an artificial neural network capable of binary classification. Users train the robot by showing it different colours and using physical buttons to indicate preference. Pressing the green button trains Robin to “like” a colour, while the red button trains it to “dislike” a colour. During the display phase, when a colour is brought close to Robin’s ultrasonic “eye,” the robot responds emotionally: it becomes happy if the colour is liked and sad if it is disliked. This demonstrates how training data influences machine behaviour.
Who is it for?
Robin is designed for students, beginners, and anyone curious about how computers learn. It is especially suitable for children and first-time learners exploring AI and machine learning concepts through hands-on interaction.
Why did your research lead to this idea?
Research into neural networks and tools like ChatGPT highlighted how complex AI systems are built from simple foundations. This led to the idea of creating a tangible, approachable example that breaks down machine learning into an understandable experience.
What is the value?
The project demystifies AI by turning abstract concepts like neurons and binary classification into a playful, physical interaction.
What were your learnings?
We learned how training data shapes machine behaviour, how perceptrons function, and how physical interaction can make complex computational ideas more accessible and engaging.
