Glossary · Phase 4 · Deep learning
Inverted dropout
The usual way dropout works: scale up the surviving values during training, so nothing needs changing when the model is used.
Try it now
Press the button to see the next batch's thinned network. The survivors get scaled up so the total signal stays the same on average.
Learn it properly
Dropout · Randomly switch off neurons to prevent overfitting. About 25 minutes.
Other terms from the same lesson
- dropout: Randomly switching off some neurons during training so the network can't lean on any single one.
Never trained an AI?
Teach a small one to write, right in your browser. It takes about 20 seconds and needs no account.