AI0-001 AI Concepts and Foundations Practice Question
Which TWO of the following are techniques used for reducing overfitting in neural networks? (Choose two.)
⚠ Common exam trap
CompTIA often tests the distinction between regularization techniques and other training strategies, so the trap here is that candidates may confuse boosting (an ensemble method) with regularization, or assume that increasing model complexity (more layers) or learning rate can help reduce overfitting when they actually do the opposite.
Answer choices
Why each option matters
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
Dropout
Dropout (A) is correct because it randomly deactivates a fraction of neurons during each training iteration, which prevents the network from relying on specific neurons and forces it to learn more robust, generalizable features, thereby reducing overfitting. L2 regularization (C) is correct because it adds a penalty term proportional to the squared magnitude of the weights to the loss function, discouraging large weights and constraining model complexity to improve generalization. Boosting (B) is an ensemble meta-algorithm that combines weak learners to reduce bias, not a neural-network overfitting-reduction technique. Increasing the learning rate (D) typically causes unstable training or divergence rather than reducing overfitting. Increasing the number of hidden layers (E) raises model capacity and usually worsens overfitting rather than mitigating it.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✓
Dropout
Why this is correct
Dropout randomly deactivates units each training pass, preventing neurons from co-adapting to noise and forcing redundant representations. This regularising effect reduces overfitting, satisfying the question's requirement for a technique that improves generalisation rather than training accuracy.
- ✗
Boosting
Why it's wrong here
Boosting sequentially reweights misclassified training samples to fit them harder, which reduces bias in weak learners and can increase variance, so it does not reduce overfitting in neural networks. It is tempting because boosting is a legitimate ensemble method, correct when the goal is lowering bias in underfit models.
- ✓
L2 regularization
Why this is correct
L2 regularization adds a penalty proportional to squared weights to the loss, shrinking parameters toward zero and limiting model complexity. This constrains the network's capacity to memorise training noise, directly reducing overfitting as the question requires.
- ✗
Increasing the learning rate
Why it's wrong here
Increasing the learning rate alters the size of weight updates, causing unstable convergence and often sharper minima that generalise worse, so it does not reduce overfitting. It is tempting because the rate is a tunable hyperparameter, and lowering it is a genuine stabilisation technique, but raising it is not.
- ✗
Increasing the number of hidden layers
Why it's wrong here
Adding hidden layers increases model capacity, letting the network fit training data more closely and widening the train-validation gap. It is tempting because extra depth helps when a model underfits and cannot represent the underlying function, which is the opposite of the overfitting scenario here.
About these practice questions
This AI0-001 question is part of Courseiva's 962-question bank — original exam-style content with full explanations and wrong-answer analysis, never real exam questions or exam dumps. Learn why practice questions differ from exam dumps →
JA
Written by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
This AI0-001 practice question is part of Courseiva's free CompTIA certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the AI0-001 exam.