A concept · Inside the Machine · Computing
Neural Networks
Layers of simple units learning patterns from data.
A photo app sorts a decade of holiday pictures by face and gets almost all of them right. Nobody at the company sat down and wrote a rule for what our aunt's face looks like. The same week, a chatbot at work drafts a decent email from three bullet points. When we ask how it did that, the honest answer from the people who built it is that they cannot point to the line of code. So where did the skill come from?
What Neural Networks means
A neural network is a stack of layers made of very simple units, each passing a number to the next, and the whole stack is tuned on examples until its outputs match what we wanted. Show it a million labelled photos and it ends up sorting faces; nobody writes the sorting. Owning the idea changes how we treat the tools. We stop asking a chatbot to explain the rule it followed, because there is no rule to recite, only a mapping that happened to work on what it saw. We start asking instead what it was trained on and whether our case looks like that. Before: the machine knows. After: the machine has been shaped, and the shape has edges.
Not to be confused with
A neural network is not the same thing as machine learning, deep learning or a language model. Machine learning is the whole field of learning from data. A neural network is one family of methods within it. Deep learning is neural networks with many layers, and a language model is one thing built that way.
Neural Networks examples: where it shows up
- At work: the tool that drafts our reports was never told the rules of our industry; it absorbed a pattern from other people's writing, and it will follow that pattern into places our rules forbid.
- At home: the photo app that finds our children in ten years of pictures learned faces from millions of strangers first.
- In the news: when a headline says an AI "decided" something, no decision was written down anywhere; a trained mapping produced an output, and nobody can fully read why.
Where it sits
Concept 1 of 4 in Inside the Machine, a journey in the course Understand AI.
See where Neural Networks sits on the sky →
Nearest on the sky
A filled dot marks a star that is already open to walk.
Knowing it isn't owning it.
Connect Neural Networks to your mental models and apply it to your world today.
The first one is on us. No signup, no card.