A lock-chamber ridership network posts a tiny training error and a large validation error. The first proposed fix is more epochs on the same architecture. What should they do instead?
Select an answer to reveal the explanation.
Short Explanation
Tiny train error and a fat validation gap is overfitting. Regularize with dropout, L1 or L2, weight decay, or fewer trees. More epochs grow the gap, Comprehend is not a penalty, and raising capacity is the underfitting fix.
Full Explanation
A tiny training error with a large validation error is overfitting. Regularize (dropout, L1/L2, weight decay, fewer trees) rather than adding epochs. Comprehend is not a regularizer, and raising capacity is the underfitting fix.