Deep Learning · Level 2 of 5
Dropout
A training regularizer that randomly masks activations or connections.
Random masking discourages some forms of co-adaptation and is usually disabled at evaluation.
Example
A hidden layer randomly zeros some activations during training.
Listen to the definition and example
Audio transcript
Dropout. A training regularizer that randomly masks activations or connections. Random masking discourages some forms of co-adaptation and is usually disabled at evaluation. For example: A hidden layer randomly zeros some activations during training.
Explore this concept
Why it matters
This helps explain why a neural model learns well, becomes unstable, or fails to generalize.
Start with
Quick recall question
Try answering before looking back at the definition.