🌐 Multi-Layer Perceptron (MLP)

If a neuron is an ant, the MLP is the whole ant colony.

🍰 The Layer Cake

An MLP organizes neurons into layers:

1. Input Layer: The raw data (e.g. pixels of an image).
2. Hidden Layers: Where the magic happens. Layer 1 might look for edges. Layer 2 combines edges into shapes. Layer 3 combines shapes into a face.
3. Output Layer: The final guess (e.g. "It's a Dog!").

🐍 Python Implementation

Notice how we use the exact same concepts (weights, bias, forward) from our single neuron, but now we use PyTorch's nn.Linear which is just a massive block of neurons working together!

import torch
import torch.nn as nn

class SimpleNeuralNetwork(nn.Module):
    def __init__(self, input_size, hidden_size, output_size):
        super().__init__()
        # Layer 1: A colony of 'hidden_size' neurons
        self.layer1 = nn.Linear(input_size, hidden_size)
        
        # Layer 2: The final output neurons
        self.layer2 = nn.Linear(hidden_size, output_size)
        
    def forward(self, x):
        # 1. Pass through Layer 1 and apply the ReLU bouncer
        hidden_state = torch.relu(self.layer1(x))
        
        # 2. Pass to final output layer
        final_output = self.layer2(hidden_state)
        return final_output

# A network to predict 2 categories (Dog/Cat) from 10 inputs
model = SimpleNeuralNetwork(input_size=10, hidden_size=64, output_size=2)
sample_data = torch.randn(1, 10) # 1 row, 10 features

print("Network Guess [Dog Score, Cat Score]:", model.forward(sample_data))

🗺️ Visualizing the MLP