QuestionQ18

VerifiedML Model Development

An ML engineer uses Amazon SageMaker AI to train a deep learning model. During testing, the model exhibits high variance. The ML engineer must adjust a hyperparameter to improve generalization without significantly increasing training time.

Which solution will MOST improve model generalization?

Explanation

High variance is characteristic of overfitting. Increasing L2 regularization penalizes large model weights, constraining model complexity and reducing overfitting, thereby improving performance on unseen data. AWS documentation identifies increasing regularization as an adjustment for overfitting and describes L2 regularization as reducing overall weight values.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!