wayground logo

Free Printable Worksheets

Font size

S
M
L
XL
Worksheets

Mega Quiz

Total questions: 85

Worksheet time: 39mins

Name
Class
Date
1.

Which can help to prevent exploding gradients?

a)

Standardization

b)

Batch normalization

c)

Lowering the learning rate

d)

Standardization

2.

___________ can help keep ReLU units from dying.

a)

Backpropagation

b)

Lowering the learning rate

c)

Batch normalization

d)

Standardization

e)

Updating the weights

3.

Check the standard components of neural network.

a)

A set of nodes, analogous to neurons, organized in layers.

b)

A set of weights representing the connections between each neural network layer and the layer beneath it. The layer beneath may be another neural network layer, or some other kind of layer.

c)

A set of biases, one for each node.

d)

An activation function that transforms the output of each node in a layer. Different layers may have different activation functions.

4.

“Convolutional Neural Networks can perform various types of transformation (rotations or scaling) in an input”. Is the statement correct True or False?

a)

True

b)

False

5.

In custom modelling, which cases are true?

a)

Easy to get started

b)

Requires ML expertise

c)

Data is accessible to provider

d)

Robust support, auto fine-tuning

e)

Expensive procedure

6.

Compute the following from the confusion matrix2. Sensitivity

a)

85.91

b)

87.51

c)

95.23

d)

85.79

e)

91.78

7.

In Automated ML, which cases are true?

a)

Robust Enterprise support

b)

Full control over parameter tuning

c)

Unlimited use-cases

d)

Difficult to extend

e)

Complete customization facility

8.

Which of the following option is true?

a)

Linear Regression errors values has to be normally distributed but in case of Logistic Regression it is not the case

b)

Logistic Regression errors values has to be normally distributed but in case of Linear Regression it is not the case

c)

Both Linear Regression and Logistic Regression error values have to be normally distributed

d)

Both Linear Regression and Logistic Regression error values have not to be normally distributed

9.

Standardisation / Normalization of features is required before training a Logistic Regression.

a)

True

b)

False

10.

An enormous data set with randomly sampled examples probably :

a)

does not contain redundant data

b)

tends not to carry much more predictive value than large batches

c)

has more predictive value than a medium size data set

11.

What happens if you enter a negative value for a feature cross?

a)

The model will not separate the classes

b)

Predictions will be completely wrong

c)

Both

d)

None

12.

An expensive robotic chicken crosses a very busy road a thousand times per day. An ML model evaluates traffic patterns and predicts when this chicken can safely cross the street with an accuracy of 99.99%.


Here accuracy value suggest that the ML model is doing a good job.

a)

True

b)

False

13.

Consider a classification model that separates email into two categories: "spam" or "not spam." If you raise the classification threshold, what will happen to precision?

a)

Probably increase.

b)

Definitely increase

c)

Probably decrease.

d)

Definitely decrease.

14.

Which of the following hyper parameter(s), when increased may cause random forest to over fit the data?

1. Number of Trees2. Depth of Tree3. Learning Rate

a)

Only 1

b)

Only 2

c)

Only 3

d)

Both 1 & 2

e)

Both 1 & 3

15.

Which of the following is an example of a deterministic algorithm?

a)

PCA

b)

K-Means

c)

Both

d)

None

16.

Minimizing loss+complexity is called empirical risk minimization

a)

True

b)

False

17.

In practice, machine learning models frequently cross continuous features.

a)

True

b)

False

18.

Imagine a linear model with 100 input features:


* 10 are highly informative.

* 90 are non-informative.Assume that all features have values between -1 and 1.


Which of the following statements are true?

a)

L2 regularization will encourage many of the non-informative weights to be nearly (but not exactly) 0.0.

b)

L2 regularization will encourage most of the non-informative weights to be exactly 0.0.

c)

L2 regularization may cause the model to learn a moderate weight for some non-informative features.

19.

Different cities in California have markedly different housing prices. Suppose you must create a model to predict housing prices. Which of the following sets of features or feature crosses could learn city-specific relationships between roomsPerPerson and housing price?

a)

One feature cross: [binned latitude X binned longitude X binned roomsPerPerson]

b)

Two feature crosses: [binned latitude X binned roomsPerPerson] and [binned longitude X binned roomsPerPerson]

c)

One feature cross: [latitude X longitude X roomsPerPerson]

d)

Three separate binned features: [binned latitude], [binned longitude], [binned roomsPerPerson]

20.

Check all the good features.

a)

house_type: victorian

b)

unique_house_id: 8SK982ZZ1242Z

c)

house_age_years: 27

d)

house_age: 851472000

e)

city_id: "br/sao_paulo"

21.

Which method is good for handling extreme outliers?

a)

Logarithmic scaling

b)

Binning

c)

Winsorizing

d)

Clipping

22.

A Pearson correlation between two variables is zero but, still their values can still be related to each other.

a)

True

b)

False

23.

Models cannot multiply strings by the learned weights, that's why we can not use feature engineering to convert strings to numeric values.

a)

True

b)

False

24.

If we incorporate index numbers (i.e. 1, 2, 3, 4) instead of string values or categorical data directly into our model, it won't impose any constraints that might be problematic.

a)

True

b)

False

25.

In case of pneumonia, how you can annotate?

a)

On a numerical scale

b)

0-n for their confidence that an image contains pneumonia symptoms or not.

c)

String value like Pneumonia / Not-pneumonia

d)

By color i.e. red for Pneumonia and Green for Not-pneumonia

26.

What you should include in case of uncertainty in an annotation?

a)

Serious case

b)

Unknown

c)

Healthy case

d)

Other

27.

State whether the statements are True or False.

Statement A: When the hypothesis space is richer, overfitting is more likely.

Statement B: When the feature space is larger, overfitting is more likely.

a)

False, False

b)

True, False

c)

True, True

d)

False, True

28.

You have generated data from a 3-degree polynomial with some noise. What do you expect of the model that was trained on this data using a 5-degree polynomial as function class?

a)

Low bias, high variance

b)

High bias, low variance

c)

Low bias, low variance

d)

High bias, low variance

29.

What will happen when eigenvalues are roughly equal in PCA?

a)

PCA will perform outstandingly

b)

PCA will perform badly

c)

Can’t Say

d)

None of above

30.

PCA works better if there is?


(i) A linear structure in the data

(ii) If the data lies on a curved surface and not on a flat surface

(iii) If variables are scaled in the same unit

a)

1 and 2

b)

2 and 3

c)

1 and 3

d)

1 ,2 and 3

31.

Select the option(s) that complete the following sentence correctly. Tf.keras is _______

a)

the TensorFlow variant of the open-source Keras API

b)

an implementation of the Keras API

c)

the Keras variant of the open-source TensorFlow API

d)

an end-to-end open-source platform for machine learning

e)

a high-level API for TensorFlow

32.

If you want to create a numpy array using random floating point number between a range, what you will use?

a)

np.arrange(lower_bound_floating_num. higher_bound_floating_num)

b)

np.random.uniform(ower_bound_floating_num. higher_bound_floating_num)

c)

np.random.uniform(ower_bound_floating_num. higher_bound_floating_num, array_shape)

d)

None

33.

If you want to find the number of rows and columns of dataset what you will use?

a)

dataset.size()

b)

dataset.shape

c)

dataset.shape()

d)

dataset.size

34.

How to populate an array with a sequence of numbers in numpy?

a)

np.arrange(5, 12)

b)

np.arange(5, 12)

c)

np.random.randint(5, 12)

d)

np.random.random(5, 12)

35.

NumPy uses a trick called _______ to virtually expand the smaller operand to dimensions compatible for linear algebra.

a)

list comprehension

b)

broadcasting

c)

matrix multiplication

d)

lambda function operation

36.

Underfitting occurs due to ______

a)

High bias

b)

High variance

c)

Low bias

d)

Low variance

37.

As a product manager, problems can come from ____

a)

Pressure from investor

b)

Upset users or customers

c)

Lost revenue

d)

Grumpy Engineers

e)

Struggling sales and customer teams

38.

When we talk about impact, we often want to think about the primary users.

a)

True

b)

False

39.

What are the steps to solve unwanted bias?

a)

Awareness

b)

Resource engagement

c)

Data management

d)

Iteration and learning

40.

Unbalanced selection of sorts data generate _____

a)

Selection bias

b)

Data bias

c)

Model bias

d)

Annotation bias

41.

Which method we should use in case of annotating very very small object in video annotating task?

a)

Computer vision

b)

Image detection algorithm

c)

Linear interpolation

d)

Convolutional neural network

42.

In supervised learning, a machine learning algorithm builds a model by examining many examples and attempting to find a model that minimizes loss; this process is called _______

a)

activation function

b)

cost function

c)

empirical risk minimization

d)

gradient descent

e)

lagloss function

43.

When performing gradient descent on a large data set, which of the following batch sizes will likely be more efficient?

a)

Batch gradient descent

b)

Stochastic gradient descent

c)

Mini-batch stochastic gradient descent

44.

Suppose an online shoe store wants to create a supervised ML model that will provide personalized shoe recommendations to users. That is, the model will recommend certain pairs of shoes to Marty and different pairs of shoes to Janet. The system will use past user behavior data to generate training data. Which of the following statements are true?

a)

"Shoes that a user adores" is a useful label.

b)

"The user clicked on the shoe's description" is a useful label.

c)

"Shoe size" is a useful feature.

d)

"Shoe beauty" is a useful feature.

45.

Best practices of A/B testing are.............

a)

Deciding on performance

b)

Deciding on test type based on your performance metric

c)

Choosing a maximum effect size you want to detect

d)

Determining the sample size

e)

Running the test until sample size is reached

46.

Relu helps to make model linear.

a)

True

b)

False

47.

Convolutional layer helps to reduce dimension of images.

a)

True

b)

False

48.

Pooling layer helps to reduce the dimension of input.

a)

True

b)

False

49.

Dense layer is used for classification in CNN.

a)

True

b)

False

50.

Stride in CNN mean the size of pixel shift over the input matrix.

a)

True

b)

False

51.

Checkout all the annotation tools for object detection.

a)

Appen

b)

LabelIMG

c)

Prodigy

d)

Datatrunks

e)

Hive

52.

Check all the feature extractors:

a)

Unet

b)

Vgg16

c)

Resnet

d)

inception

e)

yolo

53.

Which is the best for realtime object detection in case of speed?

a)

Yolo v4

b)

Yolo v3

c)

Faster RCNN

d)

SSD

e)

RCFN

54.

Which method is fast?

a)

Stemming

b)

Lemmatization

c)

Word2Vec

55.

What are the problems of RNN?

a)

Vanishing Gradient Problem

b)

Weight Initialization Problem

c)

Problem in Backpropagation due to activation function

d)

Exploding Gradient Problem

56.

What is the solution of the problems of RNN?

a)

Bidirectional RNN

b)

LSTM

c)

Word Embedding

d)

Stacked RNN

57.

Why dropout is used?

a)

To prevent underfitting

b)

To prevent overfitting

c)

To prevent exploding gradient descent

d)

To prevent vanishing gradient descent problem

58.

Suppose there is an image that contains humans. If you actually paint the pixels corresponding to humans in this image, then this is called semantic segmentation.

a)

True

b)

False

59.

SSD performs well in large objects.

a)

True

b)

False

60.

YOLO performs well on small objects.

a)

True

b)

False

61.

Which is the region based detector?

a)

RCNN

b)

YOLO

c)

SSD

62.

SSD uses fixed grid regression.

a)

True

b)

False

63.

Stemming can be used in _______

a)

Sentiment Classification

b)

Chatbots

c)

Spam Classifier

d)

Q/A App

64.

What are the better alternative solution of Bag Of Words?

a)

Lemmatization

b)

TF-IDF

c)

Word to Vec

65.

What are the disadvantages of TF-IDF?

a)

Semantic information is not stored

b)

Gives importance to uncommon workds

c)

Chance of overfitting

d)

Equal weight edge

66.

VGG16 is ______

a)

Feature extractor

b)

RNN model

c)

LSTM model

d)

Computer Vision algorithm

67.

RCNN has 3 models.

a)

True

b)

False

68.

UNET is used for ________

a)

Semantic segmentation of image

b)

Image classification

c)

Object detection

69.

___________ can help keep ReLU units from dying.

a)

Backpropagation

b)

Lowering the learning rate

c)

Batch normalization

d)

Standardization

e)

Updating the weights

70.

When the number of classes is large enough, we can use _______

a)

candidate sampling

b)

full softmax

71.

If we want to safe from underfitting, then regularization is the only way.

a)

True

b)

False

72.

The cost function of linear regression can be used in logistic regression.

a)

True

b)

False

73.

K-means clustering is semi-supervised machine learning system.

a)

True

b)

False

74.

Which can be used as dimensionality reduction algorithm in unsupervised learning?

a)

SMOTE

b)

PCA

c)

LDA

d)

Pooling layers

75.

Which is used to get optimal value in k-means clustering?

a)

PCA

b)

SMOTE

c)

RMSE

d)

Confusion Matrix

e)

Elbow Method

76.

Which will be better option for imbalanced dataset if the dataset is not huge?

a)

Undersampling

b)

Oversampling

77.

Which section backpropagation updates in CNN?

a)

Maxpooling layer

b)

Padding

c)

Feature extraction layer

d)

Kernel

78.

Which model or algorithms does not need scaling?

a)

CNN

b)

Logistic Regression

c)

Reinforcement learning

d)

Xgboost

79.

To pick up the high intensity from the feature map, which should be used in CNN?

a)

Padding

b)

Conv layer

c)

Pooling

d)

None of them

80.

Targetted marketing, Recommended Systems, and Customer Segmentation are applications in ...

a)

Unsupervised Learning: Clustering

b)

Supervised Learning: Classification

c)

Reinforcement Learning

d)

Unsupervised Learning: Regression

81.

Fraud Detection, Image Classification, Diagnostic, and Customer Retention are applications in ...

a)

Unsupervised Learning: Clustering

b)

Supervised Learning: Classification

c)

Reinforcement Learning

d)

Unsupervised Learning: Regression

82.

This picture shows a result of ...

a)

Supervised Learning: Classification

b)

Unsupervised Learning: Regression

c)

Unsupervised Learning: Prediction

d)

Supervised Learning: Regression

83.

This picture shows an application of ...

a)

Supervised Learning: Classification

b)

Unsupervised Learning: Clustering

c)

Unsupervised Learning: Prediction

d)

Supervised Learning: Regression

84.

Machine Learning has various function representation, which of the following is not function of symbolic?

a)

Decision Trees

b)

Rules in propotional Logic

c)

Hidden-Markov Models (HMM)

d)

Rules in first-order predicate logic

85.

Machine Learning has various search/ optimization algorithms, which of the following is not evolutionary computation?

a)

Perceptron

b)

Genetic Algorithm (GA)

c)

Neuro Evolution

d)

Genetic Programming (GP)