What is the main advantage of using the Adam optimizer compared to plain SGD?

It always requires less memory than SGD.

It eliminates the need for a loss function.

It adapts the learning rate for each parameter based on estimates of first and second moments of gradients.

It guarantees convergence to the global minimum.

A neural network is designed to approximate a non-linear function. The network has an input layer, one hidden layer with 10 neurons, and an output layer. The hidden layer uses a linear activation function, and the output layer uses a linear activation function. What is the effect of this architecture?

The network is equivalent to a single linear model

The network can approximate any non-linear function

The network will overfit the training data

The network will have vanishing gradients

An engineer is training a neural network with a small dataset. The model overfits quickly, and the team wants to use early stopping to select the best model. Which validation strategy is most appropriate to avoid biased early stopping decisions?

Use k-fold cross-validation on the entire dataset and stop when average training loss is lowest

Use the training loss to decide when to stop

Use the test set repeatedly to monitor training

Use a separate validation set held out from training

GIAC Machine Learning Engineer

Domain 3Objective 3

Neural Networks GMLE Practice Questions (Page 7)

Part of the Advanced Machine Learning and Neural Networks domain, which makes up ~30% of our current practice bank. GIAC (SANS) does not publish an official question count, but from its 180-minute exam (~70–120 total, ~21–36 in this domain), expect 7–12 from this objective — we provide 39 practice questions to prepare you well beyond it. (estimate)

39questions here

8free pages

9concepts

Questions 31–35

31foundation · easy
What is the main advantage of using the Adam optimizer compared to plain SGD?
Select an answer first
32foundation · easy
Which mathematical rule is the foundation of backpropagation?
Select an answer first
33application · medium
A neural network is designed to approximate a non-linear function. The network has an input layer, one hidden layer with 10 neurons, and an output layer. The hidden layer uses a linear activation function, and the output layer uses a linear activation function. What is the effect of this architecture?
Select an answer first
34application · medium
An engineer is training a neural network with a small dataset. The model overfits quickly, and the team wants to use early stopping to select the best model. Which validation strategy is most appropriate to avoid biased early stopping decisions?
Select an answer first
35foundation · easy
Which regularization technique randomly sets a fraction of neurons to zero during training to prevent co-adaptation?
Select an answer first

Finished these 5 questions?

Review the revealed explanations, or continue through the curriculum.

Next 5 questions →Back to GMLE

Free Basic Practice is a study aid with revealable answers — not a scored exam. Examers.io is independent and not affiliated with or endorsed by GIAC (SANS). “GMLE” is a trademark of its owner, used for identification only.

Neural Networks GMLE Practice Questions (Page 7)

Questions 31–35

What is the main advantage of using the Adam optimizer compared to plain SGD?

Which mathematical rule is the foundation of backpropagation?

An engineer is training a neural network with a small dataset. The model overfits quickly, and the team wants to use early stopping to select the best model. Which validation strategy is most appropriate to avoid biased early stopping decisions?

Which regularization technique randomly sets a fraction of neurons to zero during training to prevent co-adaptation?

Keep studying — Advanced Machine Learning and Neural Networks