1
During the forward pass of an MLP, what is computed first?
Correct AnswerB. Output of hidden layer
☑ MCQ PRACTICE
Machine Learning Unit 2
Practice objective questions for quick revision and examination preparation. Try answering each question before revealing the answer.
Machine Learning — Unit 2 MCQs
Artificial Neural Networks, MLP, RBF Networks, Curse of Dimensionality and SVM
2
Backpropagation primarily optimizes:
Correct AnswerC. Weights
3
Which function is typically used in the output layer for binary classification?
Correct AnswerD. Sigmoid
4
Which of the following is a common problem when MLPs are too complex?
Correct AnswerB. Overfitting
5
The equivalent of a biological neuron in ANN is:
Correct AnswerC. Node
6
A Multi-Layer Perceptron (MLP) can learn:
Correct AnswerC. Non-linear decision boundaries
7
The main idea of backpropagation is to:
Correct AnswerB. Propagate error backward to update weights
8
The output of an RBF neuron depends mainly on:
Correct AnswerA. Distance from the center
9
The curse of dimensionality mainly affects:
Correct AnswerB. High-dimensional data
10
The main objective of SVM is to:
Correct AnswerC. Maximize margin between classes
11
In SVM, support vectors are:
Correct AnswerC. Points closest to decision boundary
12
MLP stands for:
Correct AnswerB. Multi-Layer Perceptron
13
Number of hidden layers in MLP can be:
Correct AnswerB. One or more
14
Forward propagation moves from:
Correct AnswerB. Input to output
15
Backpropagation is mainly used for:
Correct AnswerB. Weight update
16
Backpropagation uses which rule?
Correct AnswerB. Delta rule
17
Activation function commonly used in MLP is:
Correct AnswerB. Sigmoid
18
RBF network hidden neurons typically use:
Correct AnswerC. Gaussian
19
RBF networks are mainly used for:
Correct AnswerB. Classification
20
Interpolation is used when data is:
Correct AnswerA. Missing
21
Basis functions convert input into:
Correct AnswerC. Feature space
22
SVM is mainly used for:
Correct AnswerC. Both classification and regression
23
SVM creates an __________ boundary:
Correct AnswerC. Optimal
24
Kernel trick avoids:
Correct AnswerB. Explicit feature mapping
25
Common SVM kernel is:
Correct AnswerD. All of the above
26
Gradient descent is mainly used in:
Correct AnswerB. Backpropagation
27
Distance metrics lose meaning in:
Correct AnswerB. High-dimensional space
Fill in the Blanks
Machine Learning — Unit 2
1
Δw = _________________________
Answerη (t – o) x
2
Margin is distance between hyperplane and _____________________.
Answersupport vectors
3
Primary building block of ANN is ____________________.
Answerneuron
4
Axon carries _________________ signals.
Answeroutput
5
Activation function decides whether neuron __________________.
Answerfires
6
XOR problem is solved using _____________________.
AnswerMLP
7
RBF stands for ________________________.
AnswerRadial Basis Function
8
Most common RBF activation is ________________________.
AnswerGaussian
9
SVM maximizes the _____________________.
Answermargin
10
Perceptron updates its ___________________.
Answerweights
11
MLP contains one or more _________________ layers.
Answerhidden
12
The last layer is called ________________ layer.
Answeroutput
13
Forward propagation computes the _________________.
Answeroutput
14
Backpropagation calculates the _____________________.
Answererror
15
Backpropagation is based on _____________________.
Answergradient descent
16
Error flows toward the __________________ layer.
Answerinput
17
MLP uses ________________ learning.
Answersupervised
18
Activation functions introduce ________________________.
Answernon-linearity
19
RBF networks use __________________ functions.
Answerradial
20
Curse of dimensionality occurs in ________________ dimensional space.
Answerhigh
21
Kernel helps SVM handle ________________ data.
Answernon-linear