Font size
S
M
L
XL
WorksheetsAI900-02
Total questions: 40
Worksheet time: 20mins
Name
Class
Date
1.
Which azure service is best for detecting popular brand logos
a)
Computer Vision
b)
Object Detection
c)
Regression
d)
Azure anomaly detector
2.
Which are the common machine learning types
a)
regression machin learning
b)
classification machine learning
c)
clustering machine learning
d)
features and labels
3.
Machine Learning is when machines get better at a task without explicit programming
a)
"True"
b)
"False"
4.
This uses a nueral network inspired by the human brain
a)
deep learning
b)
Machine learning
c)
artifical inteligence
d)
Data scientist
5.
A data scientist has skills in which of the following
a)
math
b)
statistics
c)
predictive modeling
d)
machine learning
e)
human brains
6.
AI is software that does which of the following
a)
machine learning
b)
anomaly detection
c)
computer vision
d)
natural language processing
e)
conversational AI
7.
A logical grouping of units of data that are closely related or share the same data structure
a)
Dataset
b)
predictive modeling
c)
Object Detection
d)
Regression
8.
This is a data set that can be used against handwritten digits
a)
MNIST database
b)
Common Objects In Context (COCO) Dataset
c)
Data Labeling
d)
Ground Truth
9.
Contains common images using a JSON file to identify objects within an image
a)
Common Objects In Context (COCO) Dataset
b)
MNIST database
c)
Data Labeling
d)
Ground Truth
10.
This identifies raw data and adding context so that machine learning models can learn
a)
Data Labeling
b)
MNIST database
c)
Common Objects In Context (COCO) Dataset
d)
Ground Truth
11.
With Supervised Machine Learning, labeling is performed by
a)
a human
b)
a computer
12.
With unsupervised machine learning, labels are/are not human readable
a)
are not
b)
are
13.
A properly labeled dataset used as the objective standard to train and assess a model is called __
a)
Ground Truth
b)
Data Labeling
c)
MNIST database
d)
Common Objects In Context (COCO) Dataset
14.
The accuracy of your trained model will depend on the accuracy of __
a)
Ground Truth
b)
Data Labeling
c)
MNIST database
d)
Common Objects In Context (COCO) Dataset
15.
If you want a precise outcome, you should use
a)
Supervised Learning
b)
Unsupervised Learning
c)
Reinforcement Learning
d)
Semi-Supervised
e)
Q Learning
16.
If you want to recognize a structure or a pattern, you should use
a)
Unsupervised Learning
b)
Supervised Learning
c)
Reinforcement Learning
d)
Semi-Supervised
e)
Q Learning
17.
If you need to make a prediction, you should use
a)
Supervised Learning
b)
Unsupervised Learning
c)
Reinforcement Learning
d)
Semi-Supervised
e)
Q Learning
18.
Which of the following terms would make it easier to work with the data
a)
dimensionality reduction
b)
association
c)
clustering
d)
regression
e)
classification
19.
A game that can play against itself uses which
a)
Reinforcement Learning
b)
Supervised Learning
c)
Unsupervised Learning
20.
What are the machine learning steps
a)
1. get a problem
b)
2, create a rule
c)
3. apply the rule
d)
4. get feedback
e)
5. adjust the rule
21.
When machine learning applies a fule, it will seek feedback. Where does the feedback come from?
a)
Compare answer to the model
b)
compare answer to data label
c)
compare answer to COCO
d)
compare answer to associations
22.
when you give machine learning data that you already know the answer to, what it is called?
a)
training
b)
labeling
c)
predicting
d)
Multiclass Classification
23.
If you give your machine learning some labeled data with known output, it is an example of
a)
Supervised Learning
b)
Unsupervised Learning
c)
Reinforcement Learning
d)
Semi-Supervised
e)
Q Learning
24.
With machine learning, independent variables are __ and dependent variables are __
a)
input, output
b)
output, input
c)
input, input
d)
output, output
25.
With supervised learning, you first train the model. Then what do you do?
a)
Try it against test data
b)
Run against the data store
c)
Train the model harder
d)
Develop an algorithm
26.
Learning by trial and error best describes which
a)
Unsupervised Learning
b)
Supervised Learning
c)
Reinforcement Learning
d)
Semi-Supervised
e)
Q Learning
27.
Sorting data into several different groups is known as
a)
Multiclass Classification
b)
Training
c)
Predicting
d)
Labeling
28.
Massive amounts of data are most neccessary for which
a)
Unsupervised Learning
b)
Supervised Learning
c)
Reinforcement Learning
d)
Semi-Supervised
e)
Q Learning
29.
Which of the following are types of Unspervised Learning
a)
inductive
b)
transductive
c)
triductive
d)
lebeling
30.
Rewarding machine learning everytime it produces a correct answer is called
a)
Reinforcement Learning
b)
Supervised Learning
c)
Unsupervised Learning
d)
Semi-Supervised
e)
Q Learning
31.
Teaching machine learning by scoring its result is known as
a)
Q Learning
b)
Reinforcement Learning
c)
Supervised Learning
d)
Unsupervised Learning
e)
Semi-Supervised
32.
How is Induction and Transfuction different?
a)
Transduction uses more info and produces more specic results
b)
Transduction is more general or global than induction
c)
Induction is less risky than transduction
d)
induction requires more training or labeling than transduction
33.
Which of the following are categories of Supervised Learning
a)
Binary
b)
Multiclass
c)
Regression
d)
inductive
34.
You want to use binary classification to determine a yes or no value. What learning model must you use?
a)
Supervised Learning
b)
Unsupervised Learning
c)
Reinforcement Learning
d)
Semi-Supervised
e)
Q Learning
35.
You want your ML to look at food and classify what kind of food it is. Because there can be an unlimited nymber of possible classifications, which supervised learning would you use?
a)
Multiclass
b)
Regression
c)
Binary
36.
You have a parking garage with 5000 spots. You want ML to tell you how many open spaces are available. Which supervised learning would you use?
a)
Regression
b)
Multiclass
c)
Binary
37.
K Neariest Neighbor (k-NN) Multiclass learning is also known as
a)
Lazy learning
b)
Regression learning
c)
Complex learning
d)
CPU Intensive learning
38.
K Neariest Neighbor (k-NN) Multiclass works best with
a)
small amounts of data
b)
large amounts of data
39.
Accuracy with K Neariest Neighbor is often determined by
a)
how many characteristics shared with the neighbor
b)
the size of the data set
c)
reinforced learning
d)
inductive reasoning
40.
Which is an Unsupervised learning method
a)
k-Means Clustering
b)
K Neariest Neighbor
Reset
