Perceptron Demo: Cat vs. Not-Cat Classifier

How a Perceptron Works

A Perceptron is the simplest type of artificial neural network. It's a model of a single neuron that was developed in the 1950s and forms the foundation for more complex neural networks.

Key Components:

1. Inputs: Binary (0 or 1) values representing features or characteristics
2. Weights: Values that determine how important each input is for classification
3. Bias: A threshold adjustment that helps the model make better decisions
4. Activation Function: A mathematical function that determines the output

How to Use This Demo:

1. Toggle the input switches ON (1) or OFF (0) to set animal characteristics
2. Adjust the weights using sliders to change the importance of each characteristic
3. Adjust the bias to change the decision threshold
4. Select an activation function to see how it affects the output
5. Watch how the dial output changes based on your inputs and weights

Activation Functions:

• Step: Outputs 1 if the weighted sum is positive, 0 otherwise. The classic perceptron function.
• Sigmoid: Squashes the output between 0 and 1, with
• Sigmoid: A smooth S-shaped function that outputs values between 0 and 1, often used in neural networks.
• ReLU (Rectified Linear Unit): Outputs the input directly if it's positive, otherwise outputs 0. Popular in modern neural networks.
• Tanh: Similar to sigmoid but outputs values between -1 and 1.

The Formula:

The perceptron calculates its output using this process:

1. Multiply each input by its corresponding weight
2. Sum all these products
3. Add the bias
4. Apply the activation function to this sum

Mathematically: output = activation(∑(input_i × weight_i) + bias)