CompTIA SecAI+ v1 Practice Exam
Questions And Answers 100% Verified
Instant Pdf Download 2026
Total Questions: 100
Passing Score: 80% (80 correct answers)
Section 1: AI/ML Fundamentals (Questions 1–20)
1. Which type of machine learning uses labeled training data to teach a
model to predict outcomes?
a) Unsupervised learning
b) Reinforcement learning
c) Supervised learning
d) Transfer learning
Answer: c) Supervised learning
2. In a neural network, what is the primary function of the activation
function?
a) To normalize input data
b) To introduce non-linearity
,c) To reduce overfitting
d) To backpropagate errors
Answer: b) To introduce non-linearity
3. Which algorithm is best suited for anomaly detection in security logs?
a) Linear regression
b) K-means clustering
c) Decision tree
d) Isolation Forest
Answer: d) Isolation Forest
4. What does "feature engineering" refer to in ML pipelines?
a) Automatically selecting the best algorithm
b) Transforming raw data into meaningful inputs for a model
c) Encrypting features for privacy
d) Removing outliers from test data
Answer: b) Transforming raw data into meaningful inputs for a model
5. Which metric is most appropriate for an imbalanced classification
problem (e.g., fraud detection)?
a) Accuracy
b) Precision-Recall AUC
c) Mean squared error
d) R-squared
Answer: b) Precision-Recall AUC
,6. What is the primary advantage of using a convolutional neural
network (CNN) for image-based security tasks?
a) Ability to process sequential data
b) Parameter sharing and spatial hierarchy learning
c) Built-in explainability
d) Lower memory usage than logistic regression
Answer: b) Parameter sharing and spatial hierarchy learning
7. Which technique helps prevent overfitting in a neural network?
a) Increasing learning rate
b) Adding more layers
c) Dropout regularization
d) Removing validation set
Answer: c) Dropout regularization
8. What is the role of a loss function in model training?
a) Measures model performance against ground truth
b) Augments the training dataset
c) Encodes categorical variables
d) Prevents gradient explosion
Answer: a) Measures model performance against ground truth
9. Which of the following is an unsupervised learning method?
a) Random Forest
b) DBSCAN
, c) XGBoost
d) Logistic regression
Answer: b) DBSCAN
10. What does "gradient descent" do?
a) Optimizes hyperparameters
b) Minimizes the loss function by updating weights
c) Splits data into training and test sets
d) Visualizes decision boundaries
Answer: b) Minimizes the loss function by updating weights
11. Which term describes when a model learns the training data too well,
including noise?
a) Underfitting
b) Overfitting
c) High bias
d) Data drift
Answer: b) Overfitting
12. In reinforcement learning, what is the exploration-exploitation trade-
off?
a) Balancing testing and training sets
b) Choosing between trying new actions vs using known rewarding
actions
Questions And Answers 100% Verified
Instant Pdf Download 2026
Total Questions: 100
Passing Score: 80% (80 correct answers)
Section 1: AI/ML Fundamentals (Questions 1–20)
1. Which type of machine learning uses labeled training data to teach a
model to predict outcomes?
a) Unsupervised learning
b) Reinforcement learning
c) Supervised learning
d) Transfer learning
Answer: c) Supervised learning
2. In a neural network, what is the primary function of the activation
function?
a) To normalize input data
b) To introduce non-linearity
,c) To reduce overfitting
d) To backpropagate errors
Answer: b) To introduce non-linearity
3. Which algorithm is best suited for anomaly detection in security logs?
a) Linear regression
b) K-means clustering
c) Decision tree
d) Isolation Forest
Answer: d) Isolation Forest
4. What does "feature engineering" refer to in ML pipelines?
a) Automatically selecting the best algorithm
b) Transforming raw data into meaningful inputs for a model
c) Encrypting features for privacy
d) Removing outliers from test data
Answer: b) Transforming raw data into meaningful inputs for a model
5. Which metric is most appropriate for an imbalanced classification
problem (e.g., fraud detection)?
a) Accuracy
b) Precision-Recall AUC
c) Mean squared error
d) R-squared
Answer: b) Precision-Recall AUC
,6. What is the primary advantage of using a convolutional neural
network (CNN) for image-based security tasks?
a) Ability to process sequential data
b) Parameter sharing and spatial hierarchy learning
c) Built-in explainability
d) Lower memory usage than logistic regression
Answer: b) Parameter sharing and spatial hierarchy learning
7. Which technique helps prevent overfitting in a neural network?
a) Increasing learning rate
b) Adding more layers
c) Dropout regularization
d) Removing validation set
Answer: c) Dropout regularization
8. What is the role of a loss function in model training?
a) Measures model performance against ground truth
b) Augments the training dataset
c) Encodes categorical variables
d) Prevents gradient explosion
Answer: a) Measures model performance against ground truth
9. Which of the following is an unsupervised learning method?
a) Random Forest
b) DBSCAN
, c) XGBoost
d) Logistic regression
Answer: b) DBSCAN
10. What does "gradient descent" do?
a) Optimizes hyperparameters
b) Minimizes the loss function by updating weights
c) Splits data into training and test sets
d) Visualizes decision boundaries
Answer: b) Minimizes the loss function by updating weights
11. Which term describes when a model learns the training data too well,
including noise?
a) Underfitting
b) Overfitting
c) High bias
d) Data drift
Answer: b) Overfitting
12. In reinforcement learning, what is the exploration-exploitation trade-
off?
a) Balancing testing and training sets
b) Choosing between trying new actions vs using known rewarding
actions
