IA pour l’Ingénieur 3DNI ISITCOM 2019-2020

Explorer Award Exam

Question: 1
What is a process to evaluate a model by dividing the data set once or several times
in training and testing called?
Select an answer
A. Cross-validation (CV)
B. K-fold cross validation
C. Leave one out cross validation (LOO-CV):
D. A/B testing
Question: 2
Pseudo-labeling is an example for a technique used in semi-supervised learning.
Select an answer
A. true
B. false
Question: 3
What occurs when a machine learning model can fit the training set perfectly and
fails with unseen future data?
Select an answer
A. Perfect fitting
B. Overfitting
C. Underfitting
D. Close fitting
Question: 4
In unsupervised learning, what is discovering the inherent groupings in data?
Select an answer
A. Clustering
B. Association
C. PCA
D. Regression
Question : 5
The connection between the Input layer, the Hidden layer and the Output layer
depicted by arrows in the graphic below, is commonly known as what?

1/12
IA pour l’Ingénieur 3DNI ISITCOM 2019-2020

Select an answer
A. Synapse or weight
B. Bias
C. Sigmoid operation
D. Neuron
Question: 6
In Watson Assistant, which three are System Entities?
(Please select the THREE that apply)
Select an answer
A. @sys-date
B. @sys-number
C. @email pattern entity: \b[a-zA-Z0-9.!#$%&'*+/=?^_`{|}~-]+@[A-Z0-9.-]+\.[A-
Z]{2,}\b
D. @sys-email
E. @sys-currency
Question: 7
In Watson Assistant, how is the standard dialog flow processed?
Select an answer
A. Top to bottom, left to right.
B. Bottom up, right to left.
C. Both top to bottom, left to right and Bottom up, right to left.
D. None of the above.
Question: 8
Which component is optional when developing a question and answer (FAQs)
chatbot?
Select an answer
A. Intent Classification
B. Entities identification
C. Dialog construction
D. Chatbot UI
Question: 9
In Watson Assistant, dialog can respond with a pause then a text response.
Select an answer
A. true
B. false
Question: 10
In Watson Assistant, entities represent the verbs, while intents represent the nouns in
the user's responses.
Select an answer
A. true
B. false
Question: 11
What could data samples that are too large or too small result in?
Select an answer
A. Skewed results
B. Sampling noise

2/12
IA pour l’Ingénieur 3DNI ISITCOM 2019-2020

Question: 12
Data pre-processing activities considers what data is available, what data is missing,
and what data can be removed.
Select an answer
A. true
B. false
Question: 13
CRISP-DM includes which six phases?
Select an answer
A. Assess situation, Data description, Data cleaning, generate test design,
determine next step, and Produce final report.
B. Business understanding, Data understanding, Data preparation, Modeling,
Evaluation and Delivery/Deployment
C. Determine business objectives, explore data, construct data, Assess model,
Review process, and Conduct final review.
D. Produce plan, Collect data, Select data, Select technique, Evaluate results, and
produce final presentation.
Question: 14
What are three common examples of selection bias?
(Please select the THREE that apply)
Select an answer
A. Under coverage
B. Voluntary response bias
C. Over coverage
D. Nonresponse bias
Question: 15
Which data challenge does NOT require reprocessing techniques?
Select an answer
A. Inconsistent values
B. Missing values
C. Data description
D. Duplicate values
Question: 16
Non-response survey bias occurs when some members of the population are
inadequately represented in the sample.
Select an answer
A. true
B. false
Question: 17
Chatbots cannot be both Reactive and Proactive at the same time.
Select an answer
A. true
B. false

3/12
IA pour l’Ingénieur 3DNI ISITCOM 2019-2020

Question: 18
What is the degree to which the chatbot engages and guides the user known as?
Select an answer
A. Purpose
B. Proactivity
C. Tone
D. Viewpoint
Question: 19
What is defining the chatbot's job description known as?
Select an answer
A. Purpose
B. Proactivity
C. Tone
D. Viewpoint
Question: 20
Which quiz TV show, did Watson win in 2011, where it was able to understand and
answer trivial questions and won against human contestants?
Select an answer
A. Eliza
B. Jeopardy!
C. Mastermind
D. The Chase
Question: 21
In Watson Knowledge Studio, you can create one workspace that contains one rule-
based model and one machine learning model.
Select an answer
A. true
B. false
Question: 22
Which three features are included in Watson Knowledge Studio?
(Please select the THREE that apply)
Select an answer
A. Text annotation
B. Text to speech conversion
C. Models Training
D. Models Evaluation
E. Data cleansing
Question: 23
With Watson Knowledge Studio we no longer need human annotators.
Select an answer
A. true
B. false

4/12
IA pour l’Ingénieur 3DNI ISITCOM 2019-2020

Question: 24
What is a training set in Watson Knowledge Studio?
Select an answer
A. A set of annotated documents that is set aside and used to evaluate the system
periodically.
B. A set of annotated documents that is used to test the trained model.
C. A set of documents that was annotated to train the model.
D. A set of documents that was annotated to validate the model.
Question: 25
Watson Studio Local supports locally similar features, such as Watson Studio Cloud,
in addition to Zeppelin notebooks.
Select an answer
A. true
B. false
Question: 26
Markdown cells in Jupyter Notebook allows Data Scientists to run Python or R codes
and see the resulting visualizations (like pie charts).
Select an answer
A. true
B. false
Question: 27
Watson Studio communities do not provide access to articles, tutorials, and data sets
that can be used.
Select an answer
A. true
B. false
Question: 28
Which two scenarios are applicable when training a custom classifier in Watson
Visual Recognition service?
(Please select the TWO that apply)
Select an answer
A. Provide three .zip positive example files and no negative ones.
B. Provide one .zip positive example file and no negative ones.
C. Provide no positive example files and three negative examples files.
D. Provide one .zip positive example file and one negative.
E. Provide no positive example files and one .zip negative example file.
Question: 29
You need to classify images for specific car brand, but it is not supported in Watson
Visual Recognition (default) classifier. What can you do in this case?
Select an answer
A. Train the General classifier with new classes introduced.
B. The General classifier applies adaptive learning and learns the new classes on
its own
C. A separate custom classifier has to be created that only classifies the image
classes it is trained on.
D. A separate library has to be used for these classifications.

5/12
IA pour l’Ingénieur 3DNI ISITCOM 2019-2020

Question: 30
Which three capabilities are officially supported by IBM Watson Visual Recognition
service?
(Please select the THREE that apply)
Select an answer
A. Train the General (default) classifier on business domain.
B. Custom Classifiers creation based on business needs.
C. Food model usage for enhanced identification of food items.
D. Tumor identification model to identify tumors in patient images.
E. Built-in models to use directly without training.
Question: 31
What is a data set of hand-written digits that act as a bench mark for training and
testing algorithms for computer vision?
Select an answer
A. MNIST
B. Iris
C. SpaceNet
D. Enron Email
Question: 32
What is adding computer generated images into user view known as?
Select an answer
A. Virtual reality
B. Augmented reality
C. Object illusion
D. Reality Image addition
E. Sampling bias
F. All of the above
Question: 33
LeNet 5 neural network proposed by Yann LeCun etal. was able to reach which
accuracy on MNIST dataset? (Choose the most accurate answer.)
Select an answer
A. 99%
B. 85%
C. 75%
D. 65%
Question: 34
The conventional computer vision pipeline consists of these operations in the
following order:
Select an answer
A. Image acquisition -> Segmentation -> Pre-Processing -> Feature Extraction &
Selection -> High level Processing
B. Image acquisition -> Pre-Processing -> Segmentation -> Feature Extraction &
Selection -> High level Processing
C. Pre-Processing -> Feature Extraction & Selection -> Segmentation -> Image
acquisition -> High level Processing
D. Image acquisition -> Segmentation -> High level Processing -> Feature
Extraction & Selection -> Pre-Processing
6/12
IA pour l’Ingénieur 3DNI ISITCOM 2019-2020

Question: 35
You performed feature extraction on an image and ended up with a huge set of
features, where some are redundant and irrelevant. What is the best course of action?
Select an answer
A. Perform Pre-processing on the image.
B. Perform Segmentation on the image.
C. Perform Feature Selection on the available features.
D. Start Visual Recognition Model Training.
Question: 36
In image representation, one of the most well known color models is RGB.
Select an answer
A. true
B. false
Question: 37
What does feature selection in a computer vision pipeline NOT help with?
Select an answer
A. Reducing the features dimensionality.
B. Avoiding overfitting when training the model with the features.
C. Resizing the image as a form of normalization.
D. Creating simpler models with less training time.
Question: 38
Hyponymy is when one word represents a _________ of another.
Select an answer
A. subclass
B. superclass
C. coclass
D. hypernym
Question: 39
Parsing medical documents and journals to gain new insights is what task?
Select an answer
A. an entities and relation identification
B. a classification
C. a translation
D. a bio analysis
Question: 40
Which two statements are correct regarding the definition of Homophones and
Homographs?
(Please select the TWO that apply)
Select an answer
A. Homophones are words that have the same written form but have unrelated
meanings.
B. Homographs are words that have similar sound when spoken but have
different meanings.
C. Homographs are words that have the same written form but have unrelated
meanings.
D. Homophones are words that have similar sound when spoken but have
different meanings or spellings.

7/12
IA pour l’Ingénieur 3DNI ISITCOM 2019-2020

Question: 41
What is one of the most popular lexical databases for English Language?
Select an answer
A. NLTK
B. OpenNLP
C. WordNet
D. LexNLP
Question: 42
Which two statements are correct regarding the definition of Hyponymy and
Polysemy? (Select two.)
Select an answer
A. A word is a polysem of another word if it represents a subclass of the other
word.
B. Hyponymy are words that have the same written form and related meaning.
C. A word is a hyponym of another word if it represents a subclass of the other
word.
D. Polysemy are words that have the same written form and related meaning.
Question: 43
IBM Watson Knowledge Studio models can be deployed to which three components?
(Please select the THREE that apply)
Select an answer
A. IBM Watson Explorer
B. Watson Discovery
C. Natural Language Classifier
D. Natural Language Understanding
E. Watson Assistant
Question: 44
IBM Language Translator service can be used on a few defined domains and
languages only.
Select an answer
A. true
B. false
Question: 45
Which two IBM Watson services CANNOT be trained by user?
(Please select the TWO that apply)
Select an answer
A. Personality Insights
B. Watson Computer Vision
C. Tone Analyzer
D. Natural Language Classifier

8/12
IA pour l’Ingénieur 3DNI ISITCOM 2019-2020

Question: 46
DeepQA is the name given to the question-answering (QA) system architecture,
which was developed for Watson to play the Jeopardy! game. Which statement
describes how Watson understood the text that was ingested?
Select an answer
A. Watson accessed to 200 million pages of structured and unstructured content
that filled 4 TB of disk storage, including the full text of Wikipedia.
B. Watson relied on the Encyclopedia Britannica to know what each of the words
mean.
C. Watson listened to humans having conversations over many hours of training.
D. Watson DeepQA used APIs to infer the meaning of each sentence.
Question: 47
Which IBM Watson service can be trained by a user?
Select an answer
A. Watson Assistant
B. Tone Analyzer
C. Personality insights
D. All of the above
Question: 48
Which two are typical use cases of integrating Watson Assistant with Natural
Language Understanding?
(Please select the TWO that apply)
Select an answer
A. Detecting a location in the user input.
B. Translating the user input text to another language.
C. Detecting the date and time in the user input.
D. Converting Text to Speech.
Question: 49
Which provides a set of libraries used for instantiating Watson services wrappers for
several programming languages?
Select an answer
A. Node.js
B. Watson SDKs
C. API Key
D. Context
Question: 50
Which service can be integrated with Watson Assistant to retrieve a response from
FAQ documents?
Select an answer
A. Natural Language Classifier
B. Natural Language Understanding
C. Watson Discovery
D. Watson Knowledge Catalog

9/12
IA pour l’Ingénieur 3DNI ISITCOM 2019-2020

Question: 51
In Watson Assistant integration, which three Watson services can be used in the pre-
processing layer?
(Please select the THREE that apply)
Select an answer
A. Watson Assistant
B. Natural Language Understanding
C. Tone Analyzer
D. Speech to Text
E. Watson Discovery
Question: 52
NLC training documents should contain at least how many columns?
Select an answer
A. 2
B. 3
C. 4
D. 1
Question: 53
You are testing a document retrieval algorithm for a search engine, you entered a
certain search word and it retrieved some documents. True Negative represents
irrelevant documents that were retrieved by the algorithm.
Select an answer
A. true
B. false
Question: 54
In Information retrieval, what represents relevant documents that an algorithm did
NOT retrieve?
Select an answer
A. True Positive
B. False Positive
C. True Negative
D. False Negative
Question: 55
Which three are methods to measure the performance of information retrieval
algorithm?
(Please select the THREE that apply)
Select an answer
A. Precision
B. Recall
C. T-Score
D. Accuracy
E. Homogeneity
Question: 56
The Watson Discovery News collection is continuously updated automatically.
Select an answer
A. true
B. false

10/12
IA pour l’Ingénieur 3DNI ISITCOM 2019-2020

Question: 57
What is the process of converting "ordinarily" to "ordinary" and "ly" an example of?
Select an answer
A. Tokenization
B. Stemming
C. Morphological parsing
D. Part of speech tagging
Question: 58
What is the process of breaking a sentence into a group of words?
Select an answer
A. Sentence segmentation
B. Tokenization
C. Part of speech (POS) tagging
D. Parsing (syntax)
Question: 59
What is the terminology used to describe marking a word in text as corresponding to
a particular part of speech?
Select an answer
A. Speech recognition
B. Token tagging
C. POS tagging
D. Speech marking
Question: 60
Porter is the most popular stemming algorithm for languages like English, French
and German.
Select an answer
A. true
B. false

11/12
 IA pour l’Ingénieur 3DNI ISITCOM 2019-2020

Nom………………………………………. Prénom…………………………………………
CIN…………………………….


Cocher la (ou les) case (s) relative(s) à la (ou les) bonne(s) réponse(s).

A B C D E A B C D E
Question 1 Question 31
Question 2 Question 32
Question 3 Question 33
Question 4 Question 34
Question 5 Question 35
Question 6 Question 36
Question 7 Question 37
Question 8 Question 38
Question 9 Question 39
Question 10 Question 40
Question 11 Question 41
Question 12 Question 42
Question 13 Question 43
Question 14 Question 44
Question 15 Question 45
Question 16 Question 46
Question 17 Question 47
Question 18 Question 48
Question 19 Question 49
Question 20 Question 50
Question 21 Question 51
Question 22 Question 52
Question 23 Question 53
Question 24 Question 54
Question 25 Question 55
Question 26 Question 56
Question 27 Question 57
Question 28 Question 58
Question 29 Question 59
Question 30 Question 60

12/12