🧠 The Artificial Neuron

The foundational building block of all Deep Learning!

🐜 The Ant Analogy

A single ant is pretty dumb. It can only do a few simple things. But millions of ants working together can build complex bridges and cities. An Artificial Neuron is our ant. By itself, it only does one tiny piece of math. But when you connect millions of them, they can write Shakespeare or recognize faces.

🧮 How it works

1. Inputs ($X$): The data it receives.
2. Weights ($W$): How much it "cares" about each input.
3. Bias ($b$): A baseline starting value.
4. Output ($Y$): $Y = (X \times W) + b$

🐍 Python Implementation

Let's build a single neuron from absolute scratch using consistent terminology that we will carry throughout the entire course!

import torch
import torch.nn as nn

class SingleNeuron(nn.Module):
    def __init__(self, input_size):
        super().__init__()
        # 1. Weights: How much we care about each input
        self.weights = nn.Parameter(torch.randn(input_size, 1))
        # 2. Bias: Our baseline value
        self.bias = nn.Parameter(torch.zeros(1))
        
    def forward(self, x):
        # 3. The Math: (Inputs * Weights) + Bias
        output = (x @ self.weights) + self.bias
        return output

# Test our single ant!
neuron = SingleNeuron(input_size=3)
sample_input = torch.tensor([[1.0, 2.0, 3.0]])

print("Neuron Output:", neuron.forward(sample_input))

🎨 Visual Representation