Neural Networks connection to biology
History of neural networks through biological neurons
Why "Neuron"? The Biological Connection
The term neuron in neural networks comes directly from its biological counterpart — the nerve cell in living organisms. Understanding this connection helps grasp the fundamental inspiration behind artificial neural networks.
What is a Biological Neuron?
A biological neuron (or nerve cell) is the basic unit of the nervous system that processes and transmits information. Key characteristics include:
- Cell Body (Soma): Contains the nucleus and performs metabolic functions
- Dendrites: Input receptors that receive signals from other neurons
- Axon: Output transmitter that sends signals to other cells
- Synapses: Connection points where signals are transferred between neurons
How Biological Neurons Process Information
- Input: Dendrites receive electrical/chemical signals from thousands of connected neurons
- Integration: The cell body sums all incoming signals
- Activation: If the combined signal exceeds a threshold, the neuron "fires"
- Output: An action potential travels down the axon to synapses
- Transmission: Signals cross synapses to trigger responses in connected neurons
The Birth of Artificial Neurons
In 1943, Warren McCulloch and Walter Pitts created the first mathematical model of a biological neuron. Their work established the foundation for what we now call neural networks.
The artificial neuron was explicitly designed to mimic the "all-or-nothing" firing behavior of biological neurons.
Key Similarities
| Biological Neuron | Artificial Neuron |
|---|---|
| Dendrites | Input connections |
| Cell body (soma) | Sum + activation function |
| Axon | Output |
| Synaptic strength | Weight parameters |
| Firing threshold | Bias/threshold value |
Why the Name Matters
The naming wasn't arbitrary — it reflected the design philosophy:
- Inspiration: The computational model was directly inspired by how brain cells process information
- Abstraction: While simplified, the core idea of weighted inputs passing through an activation function mirrors neural behavior
- Historical Context: Early AI researchers literally studied brain biology to understand how to create intelligent machines
Simplifications in Artificial Neurons
Modern artificial neurons are vast simplifications of biological neurons:
- Biological neurons use electrochemical signals; artificial ones use mathematical operations
- Real neurons have complex timing and plasticity mechanisms
- Biological synapses change strength dynamically (learning)
- Artificial neurons use backpropagation for weight updates