Create an animated lesson
On requestThis article is free to read. Creating a lesson uses your daily or stored credits; review the language and settings before starting.
How does a neural network learn?
Quick answer
A neural network learns by adjusting the strengths of connections between its artificial 'neurons'. It makes a guess, measures how wrong it is, and nudges those connections to do better — repeating over huge amounts of data until its predictions improve.
Step by step
- It's layers of connected artificial 'neurons'.
- It makes predictions, then measures the error.
- It adjusts connection strengths to reduce error.
- Repeating over lots of data is how it learns.
People also ask
How does a neural network learn?
By predicting, measuring its error, and adjusting connection weights to reduce that error over many examples.
What is training a neural network?
Feeding it many examples so it gradually tunes its internal weights to make accurate predictions.
What is backpropagation?
The method that calculates how to adjust each weight to reduce the network's error.
Part of the guide: How AI Works →