test1.2

Question 1

1. What is the primary function of a perceptron?
   A) Solve non-linearly separable problems
   B) Perform classification of linearly separable data
   C) Generate new data points
   D) Reduce the number of inputs in a dataset

Answer 1

B

Question 2

2. What is the main limitation of a single-layer perceptron?
   A) It can only process numerical inputs
   B) It cannot solve non-linearly separable problems
   C) It is computationally expensive
   D) It requires multiple hidden layers

Answer 2

B

Question 3

3. What does CNN stand for?
   A) Computational Neural Network
   B) Convolutional Neural Network
   C) Complex Node Network
   D) Conditional Neural Network

Answer 3

B

Question 4

4. Which type of learning does a perceptron use?
   A) Supervised Learning
   B) Unsupervised Learning
   C) Reinforcement Learning
   D) Semi-supervised Learning

Answer 4

A

Question 5

5. What is the primary function of the activation function in a neural network?
   A) Adjust the learning rate
   B) Transform input values into a desired range
   C) Store training data
   D) Improve computational efficiency

Answer 5

B

Question 6

6. In the perceptron model, what is the role of the bias term?
   A) To scale the inputs
   B) To prevent overfitting
   C) To shift the decision boundary
   D) To normalize the data

Answer 6

C

Question 7

7. What is the learning rule used in perceptron training?
   A) Backpropagation
   B) Hebbian Learning
   C) Perceptron Learning Rule
   D) Gradient Descent

Answer 7

C

Question 8

8. What type of problem does the “punches-kicks” dataset represent?
   A) Regression
   B) Classification
   C) Clustering
   D) Reinforcement Learning

Answer 8

B

Question 9

9. What is the primary goal of supervised learning?
   A) Discover hidden structures in data
   B) Predict outcomes based on labeled examples
   C) Reduce the number of features in a dataset
   D) Improve computational speed

Answer 9

B

Question 10

10. What happens when the perceptron makes a classification mistake?
    A) It removes the misclassified data point
    B) It adjusts the weights using the perceptron learning rule
    C) It increases the learning rate
    D) It decreases the bias term

Answer 10

B

Question 11

11. What is the role of weights in a neural network?
    A) They store the training data
    B) They determine the influence of each input
    C) They eliminate outliers
    D) They control the batch size

Answer 11

B

Question 12

12. What does the perceptron activation function output?
    A) A probability between 0 and 1
    B) A binary value (0 or 1)
    C) A continuous range of values
    D) A normalized value

Answer 12

B

Question 13

13. What is the error calculation formula for a perceptron?
    A)
    𝐸
    =
    (
    𝑇
    −
    𝑦
    )
    E=(T−y)
    B)
    𝐸
    =
    1
    2
    (
    𝑇
    −
    𝑦
    )
    2
    E=
    2
    1
    ​
    (T−y)
    2

C)
𝐸
=
∑
𝑤
𝑖
𝑥
𝑖
E=∑w
i
​
x
i
​

D)
𝐸
=
𝑤
1
𝑥
1
+
𝑤
2
𝑥
2
E=w
1
​
x
1
​
+w
2
​
x
2
​

Answer 13

A

Question 14

14. What does Adaline stand for?
    A) Adaptive Linear Neuron
    B) Advanced Learning Network
    C) Artificial Deep Learning Algorithm
    D) Adaptive Layer Neural Entity

Answer 14

A

Question 15

15. Which learning method does Adaline use?
    A) Perceptron Rule
    B) Gradient Descent
    C) Reinforcement Learning
    D) Clustering

Answer 15

B

Question 16

16. What is the advantage of the sigmoid activation function over the step function?
    A) It is computationally cheaper
    B) It allows continuous weight updates
    C) It eliminates the need for training
    D) It only works with deep networks

Answer 16

B

Question 17

17. What type of function does Adaline use for weight updates?
    A) Discrete step function
    B) Mean squared error
    C) Softmax function
    D) Logarithmic function

Answer 17

B

Question 18

18. What is the purpose of gradient descent in Adaline?
    A) To update weights in the direction of decreasing error
    B) To normalize input data
    C) To classify non-linearly separable data
    D) To eliminate redundant neurons

Answer 18

A

Question 19

19. Which function does the perceptron struggle to classify?
    A) AND
    B) OR
    C) XOR
    D) NOT

Answer 19

C

Question 20

20. What is the purpose of backpropagation?
    A) To adjust weights in a multi-layer network
    B) To find linearly separable solutions
    C) To eliminate outliers in a dataset
    D) To create a single-layer perceptron

Answer 20

A

Question 21

21. How many hidden layers are needed in a Multi-Layer Perceptron (MLP) to solve XOR?
    A) 0
    B) 1
    C) 2
    D) 3

Answer 21

B

Question 22

22. What is the role of the learning rate (
    𝛼
    α) in weight updates?
    A) Controls the step size for weight updates
    B) Determines the number of neurons in the hidden layer
    C) Fixes the number of training iterations
    D) Normalizes input data

Answer 22

A

Question 23

23. What is the main characteristic of deep neural networks?
    A) They contain multiple hidden layers
    B) They require no training
    C) They only work with image data
    D) They use rule-based learning

Answer 23

A

Question 24

24. What is the main purpose of hidden layers in a neural network?
    A) To make training faster
    B) To allow the network to learn complex patterns
    C) To store training examples
    D) To increase memory usage

Answer 24

B

Question 25

25. Which technique prevents a neural network from overfitting?
    A) Using more hidden layers
    B) Decreasing the learning rate
    C) Regularization (e.g., dropout)
    D) Increasing the bias

Answer 25

C

Question 26

26. What is the output range of the sigmoid activation function?
    A) -1 to 1
    B) 0 to 1
    C) -∞ to +∞
    D) 0 to 10

Answer 26

B

Question 27

27. What is the advantage of ReLU over sigmoid?
    A) It is faster to compute
    B) It provides better accuracy for small datasets
    C) It does not require weight updates
    D) It eliminates bias terms

Answer 27

A

Question 28

28. What is the primary function of a convolutional neural network (CNN)?
    A) Text processing
    B) Feature extraction for images
    C) Solving XOR problems
    D) Performing reinforcement learning

Answer 28

B

Question 29

29. What is the advantage of having more than one hidden layer in a neural network?
    A) It reduces computational cost
    B) It allows the network to learn more complex patterns
    C) It removes the need for activation functions
    D) It eliminates the error function

Answer 29

B

Question 30

30. Which of the following is NOT a supervised learning algorithm?
    A) Perceptron
    B) k-Means Clustering
    C) Support Vector Machine (SVM)
    D) Adaline

Answer 30

B