StackedML
Practice
Labs
Questions
Models
Pricing
Sign in
Questions
/
Math Foundations
/
Calculus
/
Loss functions
← Previous
Next →
373.
Huber Loss Behavior
medium
Huber loss combines properties of MSE and MAE. How does it behave for small and large errors?
A
It behaves like MAE for all errors but uses squared differences for computational efficiency on GPU hardware
B
It behaves like MSE for all errors but clips the gradient at a fixed threshold to prevent exploding gradients in the optimization process
C
It behaves like MAE for small errors and like MSE for large errors, amplifying large deviations while ignoring small ones across different error magnitudes during training
D
It behaves like MSE for small errors and like MAE for large errors, providing robustness to outliers while maintaining smooth gradients near zero
Sign in to verify your answer
← Back to Questions