StackedML
Practice
Labs
Questions
Models
Pricing
Sign in
Questions
/
Deep Learning
/
Regularization & Stability
/
Dropout
← Previous
Next →
239.
Dropout as Regularization
easy
Why does dropout act as a regularizer?
A
It prevents neurons from co-adapting by ensuring no neuron can rely on specific other neurons always being present
B
It adds a penalty term to the loss function proportional to the fraction of neurons that were active each step
C
It constrains weight magnitudes by zeroing activations whose corresponding weights exceed a threshold
D
It reduces the number of trainable parameters by permanently removing low-importance neurons during training in the network architecture
Sign in to verify your answer
← Back to Questions