StackedML
Practice
Labs
Questions
Models
Pricing
Sign in
Questions
/
Deep Learning
/
Foundations
/
Neural networks
← Previous
Next →
210.
Depth vs Width in Neural Networks
medium
Why are deep networks (many layers) often preferred over wide shallow networks for complex tasks?
A
They can compose hierarchical features efficiently, representing complex functions with exponentially fewer parameters than shallow networks
B
They require less labeled data than shallow networks since depth enables better transfer of features across tasks in complex learning problems
C
They converge faster during training since deeper gradient pathways accelerate parameter updates throughout the entire network architecture
D
They are less prone to overfitting since depth acts as an implicit regularizer that constrains the hypothesis space
Sign in to verify your answer
← Back to Questions