MLA-C01 Question 180
Select 3A data scientist is training a deep learning model on a dataset that exhibits overfitting due to the large number of parameters in the network. Which of the following regularization techniques can be applied to mitigate overfitting and improve the model's generalization ability?
- A
Apply dropout to randomly deactivate a subset of neurons during training.
- B
Use L1 regularization to enforce sparsity in the model weights.
- C
Increase the learning rate to speed up convergence.
- D
Use L2 regularization to penalize large weights in the model.
- E
Add more layers to the neural network to capture complex patterns.
Show answer and explanation
Correct answers: A, B, D
Explanation
Overfitting occurs when a model performs well on the training data but poorly on unseen data. Regularization techniques like dropout, L1, and L2 regularization are effective methods to control overfitting by either reducing reliance on specific neurons (dropout) or penalizing large or non-sparse weights (L1 and L2 regularization). These methods enhance the model's ability to generalize to new data.
- A. Correct.
Correct: Dropout randomly deactivates a subset of neurons during training, which helps prevent overfitting by reducing the reliance on specific neurons and promoting generalization.
- B. Correct.
Correct: L1 regularization adds a penalty proportional to the absolute values of the model weights, encouraging sparsity and reducing overfitting.
- C. Incorrect.
Incorrect: Increasing the learning rate may lead to instability in training and does not directly address overfitting. It is not a regularization technique.
- D. Correct.
Correct: L2 regularization adds a penalty proportional to the square of the model weights, which discourages large weights and helps mitigate overfitting.
- E. Incorrect.
Incorrect: Adding more layers can increase the capacity of the network, potentially exacerbating overfitting rather than mitigating it.