試題詳解

27. A deep learning-based image classification model performs extremely well on training data but performs poorly on unseen test images. Which of the following is the most effective approach to mitigate this issue?
(A) Increase the depth of the neural network to capture more complex patterns in training data
(B) Use dropout regularization to randomly deactivate neurons during training
(C) Train the model longer until the test accuracy starts improving
(D) Use a smaller training dataset to reduce overfitting
(E) Reduce the number of training epochs to prevent the model from converging completely