Optimization · Level 2 of 5
RMSProp
An optimizer that rescales gradients using a moving average of squared gradients.
Its adaptive denominator moderates coordinates with persistently large gradients.
Listen to the definition and example
Audio transcript
RMSProp. An optimizer that rescales gradients using a moving average of squared gradients. Its adaptive denominator moderates coordinates with persistently large gradients. For example: A trainer divides a gradient by a running root-mean-square estimate.
Explore this concept
Why it matters
This helps you understand what parameter updates are trying to improve and when they can fail.
Quick recall question
Try answering before looking back at the definition.