Technology
How does a neural network learn?
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.
See it in motion.
Watch a 2-minute animated lesson that shows exactly how a neural network works.
Step by step
- 1It's layers of connected artificial 'neurons'.
- 2It makes predictions, then measures the error.
- 3It adjusts connection strengths to reduce error.
- 4Repeating over lots of data is how it learns.
Frequently asked questions
- 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.