1 Ken Radh3touh Its Ok U Have Bilgacem For Further Assistance If You Dont Support Bilgacem Timchi T

1) ken radh3touh its ok u have bilgacem
for further assistance ,if you dont support

bilgacem timchi tnayek enti ou jarek ,
bilgacem is prophet , jesus is god , lftour
is 3ija
Label when the image remake detection service is called successfully sugestion What's the situation?
A. FALSE
B. uncertainty
C. TRUE
D. certainty
Correct Answer: A,B,C
According to Huawei Cloud EI Intelligent platform, which of the following solutions can be provided?
A. Intrusion recognition scheme
B. Vehicle identification scheme
C. Crowd statistics program
D. Policy query scheme based on knowledge graph
Correct Answer: A,B,C,D
HiAI Mobile computing platform provides rich upper-level functional services API, Can run efficiently on
mobile devices.
A. TRUE
B. FALSE
Correct Answer: A
twenty two.Among machine learning algorithms, which of the following is not unsupervised learning?
A. Xgboost
B. GMM
C. Clustering
D. Association rules
Correct Answer: A
What model is not a cyclic neural network?
A. GRU
B. GBOT
C. RNN
D. LSTM
Correct Answer: B
1) ken radh3touh its ok u have bilgacem for further assistance ,if you dont support bilgacem timchi tnayek enti ou
1
jarek , bilgacem is prophet , jesus is god , lftour is 3ija
ken radh3touh its ok
TensorFlow It is an end-to-end open source platform for machine learning and deep learning.
A. FALSE
B. TRUE
Correct Answer: B
Huawei Cloud El builds enterprise intelligence services based on three-tier services.Which of the
following options does not belong to Layer 3 services?
A. Integration services
B. General domain services
C. Industry sector services
D. Basic platform services

Correct Answer: A
What are the common clustering algorithms?
A. Hierarchical clustering
B. Kmeans
C. Density clustering
D. Spectral clustering
In the environment preparation for invoking HUAWEI CLOUD services, which of the following does not
require confirmation of installation?
A. setuptools
B. Jieba
C. websocket-client
D. requests
Correct Answer: B
According to the development process of the robot, it is usually divided into three generations,
respectively are: (Multiple Choice)
A. Robot with sensation
B. Teaching Reproduction Robot
C. Robots that will think
D. Intelligent robot
Correct Answer: A,B,D
Loss function and model function are the same thing.

A. FALSE
B. TRUE
Correct Answer: A
Which of the following is not a way for the TensorFlow program to read data?
A. Preload data
2
B. Read from the file
C. Feeding data
D. Write a file format reader

Correct Answer: D
The optimizer is an important part of training neural networks. The purpose of using the optimizer does
not include which of the following:
A. Avoid local extremes

B. Reduce the difficulty of manual parameter setting
C. Speed up algorithm convergence

D. Avoid overfitting
Correct Answer: D
Self-encoder is an unsupervised learning algorithm

A. False
B. True
Correct Answer: B
What research areas does artificial intelligence have? (Multiple choice)
A. Speech recognition
B. Natural language processing
C. Machine learning
D. Computer Vision
Which of the following steps in deep learning are automatically completed by the model?
A. Feature selection
B. Feature extraction
C. Model training
D. Analysis and positioning task
Correct Answer: A,B
The commonly used functions for mathematical operations in Python are basically in lhe math module
and the cmath module.
A. True
B. False
Correct Answer: A
Deep learning neural network training requires a lot of matrix calculations. Generally, we need to use
hardware to enable the computer to have parallel computing capabilities. The following hardware
devices can provide parallel computing capabilities.:
A. RAM
B. Motherboard
C. GPU
D. CPU
3
Correct Answer: C
The activation function plays an important role in the neural network model learning and understanding
of very complex problems. The following statement about the activation function is correct.
A. Most of the activation functions are nonlinear functions, and a few are linear functions
B. The activation function is partly a nonlinear function, partly a linear function

C. Activation functions are non-linear functions
D. Activation functions are linear functions
Correct Answer: C
Which of the following are AI Application areas?
A. Smart City
B. Wisdom education
C. Smart medical
D. Smart home
AI Chips are divided into business applications and can be divided into?
A. GPU
B. training
C. Model building
D. reasoning
Correct Answer: B,D
In deep learning tasks, when encountering data imbalance problems, which of the following methods
can we use to solve the problem?
A. Synthetic sampling
B. Random oversampling
C. batch deletion
D. Random undersampling
The continue statement in the Python language is used to jump out of this loop and then continue to
execute the next round of statements.
A. True
B. False
Correct Answer: A
The number of rows and columns of the three matrices A , B and C is 3 rows and 2 columns, 2 rows. 3
columns. 3 rows and 3 columns. which of the following operations is meaningful?
A. BC
B. AB-BC
C. AC
D. A+B
4
Correct Answer: A
Facing industry research and full scenarios AI The huge gap between applications, MindSpore Bridging
the application gap to help inclusiveness AI Of technological innovation does not include which of the
following?
A. New programming language

B. New programming paradigm
C. New ways of collaboration
D. New execution mode
Correct Answer: A
During application development, which of the following operations is not a typical network definition?
A. Network definition
B. Knowledge distillation
C. Network execution
D. Weight initialization
Correct Answer: B
A scalar k is multiplied by matrix A equal to k and each of the numbers in A is multiplied.

A. False
B. True
Correct Answer: B
In the classic convolutional neural network model,Softmax What hidden layer does the function follow?
A. Convolutional layer
B. Fully connected layer
C. Pooling layer
D. All of the above
Correct Answer: B
Bagging In integrated learning, the relationship between each base learner is?
A. Additive relationship
B. The latter model must be built on top of the previous model
C. relationship
D. Independent
Correct Answer: C
Which of the following environments does not support to install the T11nsorFlow?
A. Docker
B. Linux
C. Mac OS
D. OpenStack
Correct Answer: D
HUAWEI CLOUD EI There are many types of agents according to different industries and application
scenarios. Currently, Huawei Cloud EI The agent has?
5
A. Park Agent
B. Traffic Agent
C. Industrial Agent
D. Car intelligence
AI Chip is also called AI Accelerators are specially designed to handle a large number of computing
tasks in artificial intelligence applications functional module.
A. TRUE
B. FALSE
Correct Answer: A
In the process of training the neural network, our goal is to keep the loss function reduced. Which of the
following methods do we usually use to minimize the loss function?
A. Regularization
B. Gradient descent
C. Dropout
D. Cross-validation
Correct Answer: B
HiAI What are the advantages of mobile computing platforms?
A. Quick source code
B. plentiful API
C. Complete documentation
D. Diverse tool chain
In order for a machine to be intelligent, it must be knowledgeable. Therefore, there is a research field in
artificial intelligence. which mainly studies how computers automatically acquire knowledge and skills
to achieve self-impr011emenL What is the branch of this research called?
A. Natural language processing
B. Neural Network
C. Machine learning
D. Expert system
Correct Answer: C
According to the definition of information entropy what is the bit entropy of throwing a uniform coin?
A. 0
B. 0.5
C. 1
D. -1
Correct Answer: C
enter 32*32 Image with size 5*5 The step size of the convolution kernel is 1 Convolution calculation,
output image Size is:
6
A. 28*23
B. 29*29
C. 23*23
D. 28*28
Correct Answer: D
The following belongs to TensorFlow2.0 The characteristic is?
A. Introduce Keras interface
B. Support more platforms and more languages
C. Continue to be compatible Tensorflowl.x Related modules
D. Support static image mechanism
Voice recognition refers to the recognition of audio data as text data.
A. TRUE
B. FALSE
Correct Answer: A
The following code was used when compiling the
model:model.compile(optimizer='Adam,loss='categorical.crossentropy',metrics=
[tf.keras.metrics.accuracy]), currently using evaluate When the method evaluates the model, which of
the following indicators will be output?
A. categorical accuracy
B. categorical_ 1oss
C. loss
D. accuracy
Correct Answer: C,D
The Python language does not allow embedding another loop body in a loop body
A. True
B. False
Correct Answer: B
The time interval in the Python language is a Hoating-point fraction in seconds
A. False
B. True
Correct Answer: B
Where is the main foothold of symbolism?

A. The foothold is behavior control, adaptive and evolutionary computing.
B. The foothold is perception and action.
C. The foothold is reasoning, symbolic reasoning and machine reasoning.
D. The foothold is in neuron networks and deep learning.
Correct Answer: C
7
PyTorch All existing versions support Python2 with Python3?
A. FALSE
B. TRUE
Correct Answer: A
TensorFlow only supports Python and R language
A. True
B. False
Correct Answer: B
HUAWEI CLOUD EI Enable more use of corporate boundaries AI And big data services to accelerate
business development and benefit society. HUAWEI CLOUD EI The service can serve the enterprise in
the following aspects?
A. Industry wisdom
B. Industry data
C. algorithm
D. Computing power
There are a lot of data generated during the training of the neural network.What mechanism does
TensorFlow use to avoid excessive input data?
A. placeholder
B. fetch
C. feed
D. Client
Correct Answer: A
Which of the following description of the number of rows and columns of the matrix is correct?
A. The number of rows has no relationship with the number of columns.
B. The number of rows is less than the number of columns

C. The number of rows 1s equal to the number of columns
D. The number of rows 1s greater than the number of columns
Correct Answer: A
Atlas 800 AI There are multiple models of servers, of which the one based on the Kunpeng processor
platform is?
A. Atlas 800 model: 9000
B. Atlas 800 model: 3000
C. Atlas 800 model: 3010
Correct Answer: B
The constituent elements of the Python identifier include: numbers, letters, and underscores.
A. True
B. False
Correct Answer: A
8
Which of the following description 1s wrong about the image recognition cloud service?
A. The service can provide object and scene labels and can also provide more conceptual labels.
B. The service can also identify the sea, city, port and other scenes.
C. The service can perform semantic analysis on the content in the image to extract labels with specific abstract
meanings
D. The service identifies specific objects such as skyscrapers, buildings, and ships.
Correct Answer: A
Huawei's chip support HUAWEI HiAI Which module of?
A. HiAI Engine
B. HiAI Framework
C. HiAI Foundation
D. HiAI Service
Correct Answer: C
The word recognition in the speech recognition service refers to the synchronous recognition of short
speech. Upload the entire audio at once, and the recognition result will be returned in the response.
A. FALSE
B. TRUE
Correct Answer: B
Information theory is a branch of applied mathematics. The main research is to quantify how much
information a signal contains. Who is the first to propose information theory?
A. Shannon
B. Minsky
C. Turing
D. Simon
Correct Answer: A
If the deep learning neural network has the problem of gradient disappearance or gradient explosion,
our common solution is.
A. Regularization
B. use Relu Activation function
C. Gradient shear
D. Random undersampling
MSIR It is a simple, efficient and flexible graph-based function IR, A functional semantic type that can be
represented Have?
A. Interrupt
B. Free variable
C. Higher order function
D. Recursion
Correct Answer: B,C,D
9
Use with Atlas300 (3000) What conditions need to be checked when the accelerator card server
compiles and runs the program?
A. It has been installed Cmake Compilation tool
B. It has been installed CUDA Software package
C. Complete environment configuration
D. carry out Atlas Driver Installation
Correct Answer: A,C,D
Machine Learning Service is a data mining and analytics platform service that helps users quickly
discover data patterns and build predictive models through machine learning techniques and deploy
them as predictive analytics solutions
A. True
B. False
Correct Answer: A
Which of the following products is related to artificial intelligence? (Multiple Choice)
A. Voice input
B. Huawei Mate mobile phone
C. Self-driving
D. Alpha Go
Which of the following statements about universal form recognition services are correct?
A. colums Represents the column information occupied by the text block, the number is from 0 Start, list
form
B. words Representative text block recognition result
C. rows Represents the line information occupied by the text block, the number is from 0 Start, list form
D. The incoming image data needs to go through base64 coding
In a neural network, knowing the weight and deviations of each neuron is the most important step. If you
know the exact weights and deviations of neurons in some way, you can approximate any function What
is the best way to achieve this?
A. The above is not correct
B. Random assignment, pray that they are correct
C. Assign an initial value to iteratively update weight by checking the difference between the best value and the
initial
D. Search for a combmat1on of weight and deviation until the best value 1s obtained
Correct Answer: C
Training error will reduce the accuracy of the model and produce under-fitting. How to improve the
model fit? (Multiple choice)
A. Feature Engineering
B. Increase the amount of data
10
C. Reduce regularization parameters
D. Add features
Which of the following conditions is not a condition that n-fold bernoulli trials needs to meet?
A. Each test was repeated under the same conditions.
B. Each test itself is obeying normal distribution
C. There are only two possible outcomes for each trial, i.e. event A occurs and event A does not occur.
D. The results of each trial are independent of each other.

Correct Answer: B
In the deep learning network, the backpropagation algorithm is used to find the optimal parameters.
What rules are used in the backpropagation algorithm to obtain the derivation layer by layer?
A. Chain rule
B. Cumulative rule
C. Rule of Normalization
D. Law of reciprocity
Correct Answer: A
Which of the following options are <Artificial Intelligence Comprehensive Experiment> May be used in
calling Huawei cloud services?
A. region
B. sk
C. project_id
D. ak
The loss function reflects the error between the target output and actual output of the neural network.
The commonly used loss function in deep learning is:
A. Log loss function
B. Exponential loss function

C. Hinge Loss function
D. Mean square loss function
Correct Answer: D
Root The user cannot install Atlas300 (3000) Install the accelerator card on the server DDK (Device
Development Kit), Can only be installed by ordinary users.
A. FALSE
B. TRUE
Correct Answer: B
In the face search service, if we want to delete a certain face set, we can use this
code:firs_cliet.get_v2().get_face_set_service().delete_face_set("**"), among them "**" is to fill in the
actual face set name.
A. FALSE
11
B. TRUE
Correct Answer: B
Numerical calculation refers to the method and process of effectively using the digital computer to solve
the approximate problem of mathematical problems, and the discipline consisting of related
theories.Which of the following processes are involved in solving actual problems with a computer?
(Multiple Choice)
A. Programm1ng
B. Know the actual problem
C. Computer calculation results
D. Mathematical model
E. Numerical calculation method
Correct Answer: A,B,C,D,E
Which is not a Python operator?
A. Logical operator
B. Comparison operator
C. Reasoning operator
D. arithmetic operator
Correct Answer: C
Which of the following options is not a tensor attribute?
A. name (Logo)
B. type (Type)
C. node (Node)
D. shape (Dimensionality)
Correct Answer: C
What are the common types of dirty data?
A. Duplicate value
B. Malformed value
C. Logically wrong value
D. Missing value
Which of the following statements about supervised learning is correct?

A. Supervised learning is a rule-based algorithm
B. Supervised learning cannot use cross-validation for training

C. Supervised learning can be trained without labels
D. Decision tree is a supervised learning

Correct Answer: D
Artificial intelligence is a new technical science that studies and develops theories, methods and
application systems for simulating, extending and extending human intelligence. It is one of the core
research areas of machine learning.
12
A. True
B. False
Correct Answer: B
The for loop statement in the Python language can iterate through the items in any sequence.
A. True
B. False
Correct Answer: A
Question 76/129
HUAWEI HiAI What are the supported models?
A. glory V20
B. Mate 20
C. P30
D. iPhone10
What are the steps that are not part of the operation of a Python file object?
A. close
B. read
C. delete
D. write
E. open
Correct Answer: C
The three stages of artificial intelligence include computational intelligence, () , Cognitive intelligence.
A. Strong artificial intelligence

B. Weak artificial intelligence
C. Behavioral intelligence
D. Perceptual intelligence
Correct Answer: D
There are many types of neural networks in deep learning. The following neural network information is
one-way propagation:
A. LSTM
B. Convolutional Neural Network

C. Recurrent neural network
D. GRU
Correct Answer: B
Tensorflow supports multi-TPU cluster computing

A. True
B. False
13
Correct Answer: A
In machine learning, what input the model needs to train itself and predict the unknown?
A. Training algorithm
B. Neural Networks
C. Manual procedure
D. historical data
Correct Answer: D
Which of the following activation functions are prone to vanishing gradient problems?
A. ReLU
B. Softplus
C. Sigmoid
D. Tanh
Correct Answer: C,D
Regarding backpropagation, the following statement is wrong?
A. Back propagation refers to the back propagation of errors through the network
B. Backpropagation passes through the activation function
C. Backpropagation can only be used in feedforward neural networks

D. Backpropagation can be combined with gradient descent algorithm to update network weights
Correct Answer: C
In neural networks, which of the following methods are used to update the parameters when training the
network to minimize the loss function?
A. Forward propagation algorithm

B. Convolution calculation
C. Pooling calculation
D. Backpropagation algorithm
Correct Answer: D
Which description is wrong about the hyperparameter?

A. Hyperparameters cannot be modified
B. The value of the hyperparameter is not learned by the algorithm itself.

C. 1-typerparameters are parameters that set values before the algorithm begins learning.
D. Most machine learning algorithms have hyperparameters.

Correct Answer: A
TensorFlow It is one of the most popular deep learning frameworks.
A. FALSE
B. TRUE
Correct Answer: B
HUAWEI HiAI Which tool does the platform support to integrate with?
A. Jupyter Notebook
14
B. MyEclipse
C. Android Studio
D. Spider
Correct Answer: C
Which of the following is not an application of image recognition services?
A. Scene analysis
B. Target Detection
C. Speech synthesis
D. Smart photo album
Correct Answer: C
Code model.fit(mnist.train.inmage,mnist.train.labels.epochs=5)in of epochs Parameter representative?
A. The entire training set will be divided into 6 Share
B. The entire test set will be tested 5 Times

C. The entire training set will be trained 5 Times
D. The entire training set will be divided into 5 Share

Correct Answer: C
The training error will continue to decrease as the model complexity increases.
A. FALSE
B. TRUE
Correct Answer: B
Functions are well-organized, non-reusable code segments used to implement a single, or associated
Function.
A. True
B. False
Correct Answer: B
When training the network of God, we often encounter many problems. For the problem of vanishing
gradient, we can choose to make. Which of the following functions to alleviate the problem?
A. Sigmoid function
B. Relu function
C. tanh function
D. Softsign function
Correct Answer: B
Image label service returned tag There can be more than one.
A. FALSE
B. TRUE
Correct Answer: B
Which of the following does not belong to long-term memory LSTM (long Short*Term Memory) network
architecture?
A. Memory door
15
B. Forget the door
C. Output door
D. Input door
Correct Answer: A
What are the implementation modes ofTensorflow? (Multiple Choice)
A. Reverse mode
B. Stand-alone mode
C. Forward mode
D. D1stnbuted mode
Correct Answer: B,D
In the deep learning neural network, the perceptron is the simplest neural network. The correct
statement about its structure is:
A. Its network uses Sigmoid Activation function
B. Its network uses Relu Activation function
C. There are only two hidden layers

D. Only one hidden layer
Correct Answer: D
What is in the category of artificial intelligence? (Multiple choice)
A. Perception
B. None of the above
C. Action
D. Cognition
The following is the correct difference between machine learning algorithms and traditional rule-based
methods?
A. The mapping relationship of the model in machine learning is automatically learned
B. Traditional rule-based methods use explicit programming to solve problems
C. The mapping relationship of the model in the machine learning institute must be implicit
D. Traditional rule-based methods, in which the rules can be explicitly clarified manually
Traditional machine learning and deep learning are the core technologies of artificial intelligence. There
is a slight difference in the engineering process. The following steps. What you don't need to do in deep
learning is:
A. Model evaluation
B. Data cleaning
C. Feature engineering
D. Model building
Correct Answer: C
16
Which of the following does the convolutional neural network structure not include?
A. Fully connected layer
B. Convolutional layer
C. Pooling layer
D. Loop layer
Correct Answer: D
Among the following properties TensorFlow2.0 Does not support creation tensor. The method is?
A. fill
B. constant
C. create
D. zeros
Correct Answer: C
Global gradient descent algorithm, stochastic gradient descent algorithm and batch gradient descent
algorithm are all gradient descent algorithms. The following is wrong about its advantages and
disadvantages.
A. Batch gradient algorithm can solve the local minimum problem
B. The global gradient algorithm can find the minimum value of the loss function
C. The convergence process of the global gradient algorithm is time-consuming
D. Stochastic gradient algorithm can find the minimum value of the loss function
Correct Answer: D
If a model has a large deviation on the test set and a small variance, it means that the model?
A. Just fit
B. May be overfitting may be underfitting
C. Underfitting
D. Overfitting
Correct Answer: C
Only matrix A and matrix B have the same number of rows and columns. A and B can be added.
A. True
B. False
Correct Answer: A
There are many commercial applications for machine learning services.What are the main business
scenarios covered? (Multiple Choice)
A. Retailer grouping
B. Telecom customer retention
C. Predictive maintenance
D. Flnancial product recommendation
The Python language can use the "#" at the beginning of a single line of code for code comments .
A. False
17
B. True
Correct Answer: B
Generative confrontation networks are currently widely useD. The following scenarios can use this
network?
A. Information retrieval
B. Data enhancement
C. Image generation
D. Semantic segmentation
L1 with L2 Regularization is a method commonly used in traditional machine learning to reduce

generalization errors. The following is about the two. The right way is:
A. L1 with L2 Regularization can be used for feature selection

B. L1 Regularization can do feature selection
C. L2 Regularization can do feature selection
D. L1 with L2 Regularization cannot be used for feature selection

Correct Answer: B
Cut off 2019 At the end of the year,TensorFlow The released versions are?
A. 3
B. 2
C. 1
D. 0.1
Regularization is an important and effective technique for reducing generalization errors in traditional
machine learning. The following techniques are regular. The technology is:
A. Momentum optimizer
B. L1 Regularization
C. Dropout
D. L2 Regularization
Atlas 200 DK Yiteng 310 The main application of the chip is?
A. Training model
B. Build model
C. Model reasoning
Correct Answer: C
Which of the following are the characteristics of the Python language? {Multiple choice)
A. Object-oriented
B. Explanatory
C. Dynamic data type
18
D. Process oriented
What are the advantages of Python? (Multiple choice)
A. High-level language
B. Rich library
C. Simple
D. Free
Which of the following is not an artificial intelligence school?
A. Behaviorism
B. StatJst1caltsm
C. Symbolism
D. Connectionism
Correct Answer: B
The Python language can use multiple statements on the same line, separated by commas ".".
A. False
B. True
Correct Answer: A
TensorF1ow2.0 Can be used to check if it is tensor The methods are?
A. device
B. is_tensor
C. dtype
D. isinstance
Correct Answer: B,D
The naive Bayes algorithm does not require independent and identical distribution among sample
features.
A. TRUE
B. FALSE
Correct Answer: B
Which of the following descriptions about the Recurrent Neural Network (RNN) is correct?
A. Unlike convolutional neural networks. parameters of RNN cannot be shared

B. Can be used to process sequence data.
C. The units above the hidden layer are not associated With each other.
D. Cannot process variable length sequence data.

Correct Answer: B
TensorFlow2.0 not support tensor in GPU with CPU Transfer between each other.
A. FALSE
B. TRUE
19
Correct Answer: A
Which of the following can improve the computational efficiency of the neural network model? (Multjple
choice)
A. Large-scale distributed cluster
B. FPGA
C. GPU
D. TPU
The following about KNN Algorithm k The value description is correct?
A. K The larger the value, the smoother the segmentation surface of the classification
B. K The larger the value, the easier the model is to overfit
C. can k Value is set to 0
D. K Value is a hyperparameter
Correct Answer: A,D
Principal Component Analysis (PCA) is a statistical method. A set of variables that may be related to
each other is transformed into a set of linearly related variables by orthogonal transformation The
converted set of variables is called the principal component
A. True
B. False
Correct Answer: B
Which is the correct description of the python creation function? (Multiple choice}
A. The parameters need to be placed in parentheses
B. The function created starts with the def keyword followed by the function name and parentheses.
C. Return the result with return and the function ends.
D. The function content starts with a colon and needs lo be indented (Right Answers}
ModelArts Service and (). The combination of services can easily deploy the model to "end"?
A. ECS
B. HiLens
C. OBS
D. OCR
Correct Answer: B
When dealing with actual problems, when should machine learning be used in the following situations?
A. The rules are very complicated or cannot be described, such as face recognition and voice recognition
B. The data distribution itself changes over time, requiring continuous re-adaptation of the program, such as
predicting the trend of merchandise sales
C. Task rules will change over time, such as defect detection on the production line
D. The complexity of the rules is low, and the problem is small
20
An e-commerce company has always had problems with official website stalls, poor utilization of
network resources, and difficulty in operation and maintenance. Then which of the following can the
company use EI service?
A. Industrial Agent
B. Park Agent
C. Traffic Agent
D. Network agent
Correct Answer: D
tf.keras.datasets Can view keras The built-in data set.

A. TRUE
B. FALSE
Correct Answer: A
About Python correct tuple, which is correct description? (Multiple choice)
A. When a tuple contains only one element you need to add a comma after the element
B. The difference between a list and a tuple is that the elements of the tuple cannot be
C. The list uses square brackets and the tuple uses parentheses
D. Tuple creation is very simple just add elements in parentheses and separate them wrth commas
QUESTION NO: 1
Use with Atlas300 (3000) What conditions need to be checked when the accelerator card server compiles and
runs the program?
Hide answers/explanation
QUESTION NO: 2
What are the numeric types of Python? (Multiple Choice)
QUESTION NO: 3
The Python list can be identified by '[]", and the default index of the first element from left to right is 1.
Correct Answer: B
QUESTION NO: 4
The Python dictionary is widely identified by "{}". and the internal data consists of the key and its corresponding
value
Correct Answer: A
QUESTION NO: 5
The test error will keep getting smaller as the complexity of the model increases.
Correct Answer: A
QUESTION NO: 6
Traffic agent can achieve () Hour full time/Global traffic awareness?
Correct Answer: A
21
QUESTION NO: 7
Tensor Yes MindSpore The storage component of the data.
Correct Answer: B
QUESTION NO: 8
Correct Answer: D
QUESTION NO: 9
There are many types of neural networks in deep learning. The following neural network information is one-way
propagation:
Correct Answer: A
QUESTION NO: 10
Which of the following is not a module in the Tensorflow library?
Correct Answer: A
QUESTION NO: 11
The word recognition in the speech recognition service refers to the synchronous recognition of short speech.
Upload the entire audio at
once, and the recognition result will be returned in the response.
Correct Answer: B
QUESTION NO: 12
The following evaluation indicators belonging to the regression algorithm are?
Correct Answer: A
QUESTION NO: 13
It is fine to pass in only one image when calling the face comparison service.
Correct Answer: A
QUESTION NO: 14
In Huawei Cloud El Enterprise Intelligence, which basic platform services are included? (Multiple Choice)
QUESTION NO: 15
Correct Answer: A
QUESTION NO: 16
Correct Answer: C
QUESTION NO: 17
About the image content review service returned when the call is successful suggestion Field, the correct
statement is?
QUESTION NO: 18
The following is the correct difference between machine learning algorithms and traditional rule-based methods?
22
QUESTION NO: 19
Traditional machine learning and deep learning are the core technologies of artificial intelligence. There is a
slight difference in the
engineering process. The following steps. What you don't need to do in deep learning is:
Correct Answer: D
QUESTION NO: 20
In deep learning, if the neural network has a large number of layers, the problem of gradient disappearance is
more likely to occur. Strictly
speaking, in which of the following links does the sample disappearance problem occur?
Correct Answer: B
QUESTION NO: 21
In TensorFlow, data is represented in the form of tensors and calculation charts.
Correct Answer: B
QUESTION NO: 22
Correct Answer: B
QUESTION NO: 6
Traffic agent can achieve () Hour full time/Global traffic awareness?
Correct Answer: A
QUESTION NO: 7
Tensor Yes MindSpore The storage component of the data.
Correct Answer: A
QUESTION NO: 8
Correct Answer: A
QUESTION NO: 9
There are many types of neural networks in deep learning. The following neural network information is one-way
propagation:
Correct Answer: B
QUESTION NO: 10
Which of the following is not a module in the Tensorflow library?
Correct Answer: C
QUESTION NO: 11
The word recognition in the speech recognition service refers to the synchronous recognition of short speech.
Upload the entire audio at
once, and the recognition result will be returned in the response.
Correct Answer: A
QUESTION NO: 12
The following evaluation indicators belonging to the regression algorithm are?
23
Correct Answer: C
QUESTION NO: 13
It is fine to pass in only one image when calling the face comparison service.
Correct Answer: B
QUESTION NO: 14
In Huawei Cloud El Enterprise Intelligence, which basic platform services are included? (Multiple Choice)
QUESTION NO: 15
Correct Answer: B
QUESTION NO: 16
Correct Answer: B
QUESTION NO: 17
About the image content review service returned when the call is successful suggestion Field, the correct
statement is?
QUESTION NO: 18
The following is the correct difference between machine learning algorithms and traditional rule-based methods?
QUESTION NO: 19
Traditional machine learning and deep learning are the core technologies of artificial intelligence. There is a
slight difference in the
engineering process. The following steps. What you don't need to do in deep learning is:
Correct Answer: B
QUESTION NO: 20
In deep learning, if the neural network has a large number of layers, the problem of gradient disappearance is
more likely to occur. Strictly
speaking, in which of the following links does the sample disappearance problem occur?
Correct Answer: A
QUESTION NO: 21
In TensorFlow, data is represented in the form of tensors and calculation charts.
Correct Answer: B
QUESTION NO: 22
Correct Answer: D
QUESTION NO: 23
HiAI What are the advantages of mobile computing platforms?
24
QUESTION NO: 24
Python tuples are identified by "()'' and internal elements are separated by ":".
Correct Answer: B
QUESTION NO: 25
Which of the following schools does the neural network study belong to?
Correct Answer: B
QUESTION NO: 26
TensorFlow only supports Python and R language
Correct Answer: B
QUESTION NO: 27
Huawei AI The whole scenario includes public cloud, private cloud, various edge computing, IoT industry
terminals and consumer. The
deployment environment of terminal, edge, and cloud for fee-based terminals.
Correct Answer: A
QUESTION NO: 28
Feedforward neural network is a simple neural network, each neuron is arranged hierarchically, it is currently
one of the most widely used
and fastest growing artificial neural networks. The following statement about feedforward neural networks is
correct:
Correct Answer: D
QUESTION NO: 29
What are the common clustering algorithms?
QUESTION NO: 30
The determinant of square matrix A is a scalar.
Correct Answer: A
QUESTION NO: 31
The optimizer is an important part of training neural networks. The purpose of using the optimizer does not
include which of the following:
Correct Answer: C
QUESTION NO: 32
Which of the following are the elements of artificial intelligence? (Multiple choice)
QUESTION NO: 33
Among the following properties TensorFlow2.0 Does not support creation tensor. The method is?
Correct Answer: C
QUESTION NO: 34
TensorFlow It is one of the most popular deep learning frameworks.
25
Correct Answer: A
QUESTION NO: 35
TensorFlow The data type used is?
Correct Answer: C
QUESTION NO: 36
Which of the following are the cloud services provided by Huawei Cloud El Visual Cognition? (Multiple choice)
QUESTION NO: 37
Which of the following descriptions about Python program are correct? (Multiple choice)
QUESTION NO: 38
GAN is a deep learning model and one of the most promising methods for unsupervised learning in complex
distribution in recent years.
Correct Answer: A
QUESTION NO: 39
TensorFlow2.0 middle Keras The three main advantages of the interface are user-friendly, modular and
Combination, easy to expand.
Correct Answer: A
QUESTION NO: 40
Artificial intelligence is currently suitable for scenarios with I' known environment, clear objectives, and
predictable action . Deep learning in
the fields of image recognition, speech recognition, translation, etc , artificial intelligence basically has human
recognition ability. even
surpassing human beings Based on these capabilities. it has been applied to many scenes such as medical
care and public safety However,
there is still a lack of reasoning and cognition
Correct Answer: A
QUESTION NO: 41
Deep learning is different from machine learning and there are no unsupervised algorithms
Correct Answer: B
QUESTION NO: 42
Which of the following statements about universal form recognition services are correct?
QUESTION NO: 43
When the universal text recognition service is successfully called, the returned text recognition result is stored in
which of the following
fields?
Correct Answer: B
QUESTION NO: 44
In neural networks, weights are often shareD. Which of the following neural networks will share weights?
26
Correct Answer: B,D
QUESTION NO: 45
Grid search is a method of parameter adjustment.
Correct Answer: B
QUESTION NO: 46
Which of the following is HUAWEI HiAI Foundation Function of the module?
Correct Answer: C
QUESTION NO: 47
What people say about neural networks is wrong?
Correct Answer: A
QUESTION NO: 48
What is the incorrect relationship between neural networks and deep learning?
Correct Answer: C
QUESTION NO: 49
TensorFlow2.0 The methods that can be used for tensor merging are?
Correct Answer: B
QUESTION NO: 50
HUAWEI HiAI Empower APP What value?
QUESTION NO: 51
Which of the following is not MindSpore common Operation?
Correct Answer: A
QUESTION NO: 52
Atlas accelerate AI What processor is used for inference?
Correct Answer: B
QUESTION NO: 53
The number of rows and columns of the three matrices A , B and C is 3 rows and 2 columns, 2 rows. 3 columns.
3 rows and 3 columns. which
of the following operations is meaningful?
Correct Answer: B
QUESTION NO: 54
What is wrong about the image content review service?
Correct Answer: C
QUESTION NO: 55
L1 with L2 Regularization is a method commonly used in traditional machine learning to reduce generalization
errors. The following is about
the two. The right way is:
Correct Answer: A
27
QUESTION NO: 56
GPU Good at computationally intensive and easy to parallel programs.
Correct Answer: A
QUESTION NO: 57
Which of the following are solutions for the Huawei Cloud El industry scenario? (Multiple Choice)
QUESTION NO: 58
AI Chips are divided into business applications and can be divided into?
Correct Answer: A,D
QUESTION NO: 59
Deep learning neural network training requires a lot of matrix calculations. Generally, we need to use hardware
to enable the computer to
have parallel computing capabilities. The following hardware devices can provide parallel computing
capabilities.:
Correct Answer: C
QUESTION NO: 60
Which of the following items are included in the results returned when the face search service is successfully
called?
QUESTION NO: 61
CPU Mainly through increasing instructions, increasing the number of cores, etc. to improve AI performance.
Correct Answer: A
QUESTION NO: 62
View Atlas300 (3000) Which command should be used if the accelerator card driver is installed successfully?
Correct Answer: A
QUESTION NO: 63
Python authors deliberately design very restrictive grammars that make bad programming habits (such as the
next line of if statement not
indented to the right) cannot be compiled
Correct Answer: A
QUESTION NO: 64
Loss function and model function are the same thing.
Correct Answer: B
QUESTION NO: 65
Which of the following options are <Artificial Intelligence Comprehensive Experiment> May be used in calling
Huawei cloud services?
QUESTION NO: 66
What are the application scenarios for the break statement in the Python language? (Multiple Choice)
28
QUESTION NO: 67
The Python dictionary is identified by "{}", and the internal data consists of the key and its corresponding value.
Correct Answer: A
QUESTION NO: 68
The number of hidden layers of the deep learning neural network has a certain impact on the performance of the
network. The following
statement about its impact is correct.:
Correct Answer: B
QUESTION NO: 69
Huawei firmly believes that the value of Al is ultimately reflected in solving the actual business problems of the
enterprise. Therefore, Huawei
is firmly committed to the development of Al technology, focusing on domain solutions, focusing on basic
technologies and enabling
platforms, and working with industry practice leaders to develop industry solutions
Correct Answer: A
QUESTION NO: 70
TensorFlow in keras The built-in optimizer of the module has?
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
Huawe1 Certified Network Assoclate-Artficial Intelligence
0 Sign
1. Whal is the subject of artificial intelligence?

0 A. Mathematics and Physiology
0 B Psychology and Physiology
0 C Linguistics
0 D Comprehensive mterd1sc1pl1nary and marginal d1sc1phnes (Right Answers)
Answer : D
Problem List Check The Sign Ouesllons Show Correct Answers
X
Question 1 of 171 Time lef\1:41:37
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
2. Funclions are well-organized, non-reusable code segments used to implemenl a single, or associated
lunclion.
0 A. True
0 B False (Right Answers)
Answer: B
29
Problem Lisi Check The Sign Ouesllons Show Correct Answers
X
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
3. Whal is wrong description or backpropagation?

0 A. The learning process or the backpropagation algorithm consists of a forward propagation process and a
back-propagation process.
0 B The backpropagation algorithm is a learning algorithm suitable for multi-layer neural networks, which is
based on tl\e gradient descent method
0 C The backpropagation phase sends training inputs to the network to obtain an stimuli response (Right
Answers)
Question 3 of 171 Time lef\1:37.13
Noce
0 D. The backpropagation algorithm is mainly repeated by two loops (excitation propagation, weight update)
until the response of the network to the input reaches the predetermined target range.
Answer : C
X
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
4. In May 1997, the famous "Human-Machine Wars" final computer defeated Kasparov, the world chess king,
with a total score of J.5 to 2 5. Is this computer called?
0 A Dark blue (Right Answers>
0 B Dari< green
0 C Ponder
0 D. Blue sky
Answer : A
X
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
5. Whal are the service categories included in Huawei Cloud El Enterprise Intelligence? (Multiple Choice)
0 A El 'lisual cognihon (Right Answers)
0 B El online games
30
0 C El speech semantics (Righi Answers)
0 D El industry scene (Righi Answers)
Answer : ACD
X
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
6. As the following, what are the Python language design philosophy? (Multiple Choice)
0 A Beautiful (Righi Answers)
0 B Expensive
0 C Expic1t (Right Answers)
0 D Simple (Righi Answers)
Answer : ACD
X
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
Huawe1 Certified Network Assoclate-Artficial Intelligence Question 7 of 171 Time Lef\1:33:1
0 Sign
7. The history of robots is not long. In 1959. the United States, Engelberg and Devol made the world's first
generation of industrial robots, and the history of robots really began.
According to the development process of the robot, it is usually divided into three generations, respectively
are: (Multipe Choice)
0 A Teaching Reproducuon Robot (Righi Answers)
0 B Robot with sensation (Right Answers)
0 C. Robots that will think
0 D Intelligent robol (Right Answers)
Answer: ABO
Noc~
X
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
8 There are a lot of data generated during the training of the neural network. What mechanism does
0 A. Client
31
0 B feed
0 C placeholder (Right Answers)
0 D. fetch
Answer : C
X
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
8. Python tuples are identifi ed by "O'' and internal elements are separated by ":".
0 A. True
Answer: B
X
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
9. What are the numeric types of Python? (Multiple Choice)

0 A int ~nteger type) (Right Answers)
0 B long (long integer type) (Right Answers)
0 C Hoat (floating point type) (Right Answers)
0 D complex (complex number type) (Right Answers)
Answer : ABCD
Problem List Check The Sign Quesllons Show Correct Answers
X
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
10. What is the English abbfe'Jiation for /'J?

0 A. Automatic Intelligence
0 B Amfical Intelligence (Right Answers)
0 C Automatic Information
0 D. Artlfical Information
Answer : B
32
X
Question 11 of 171 Time lef\1:30·34
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
12 There are many commercial applications for machine learning seJVices Whal are the main business
scenarios covered? (Multiple Choice}
0 A Flnanc1al product recommendation (Right Answers)
0 B Pred1ct1V0 maintenance !Right Answers)
0 C Telecom customer retention (Right Answers)
0 D Retailer grouping (Right Answers)
Answer : ABCD
X
Question 12 of 171 Time Lef\1:30:9
Noc~
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-31 1-ENU
0 Sign
11. What is wrong description of the Python module?

0 A The Python module is a Python file that ends with .P'f and contains Python object definitions and Python
statements.
0 B The Python module allows you to logically organize your Python code snippets.
0 C Python modules can define lunchons classes and vanables but the module does not contain executable
code (Right Answers)
0 D. Assigning relevant code to a module can make your code better and easier to understand.
Answer : C
X
Question 13 of 171 Time lef\1.29:50
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
12. Which of the following options is not the- session mode used by Tensorflow?
0 A Explicitly call the session to generate function
0 B Explicitly call the session to close function
0 C Through the Python context manager
0 D MultJple POST quenes !Right Answers)
33
Answer : D
X
Question 14 of 171 Time lef\1.29·31
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
Noce
13. Principal Component Analysis (PCA) is a statistical method. A set of variables that may be related to each
other is transformed into a set of rinearly related variables by orthogonal transformation
The converted set of variables is called the principal component
0 A. True
Answer: B
X
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
14. What of the following does belong to convolutional neural network (CNN)? (Multiple Choice)
0 A VGGNet (Right Answers)
0 B ResNel (Right Answers)
0 C AlexNet (Right Answers)
0 D GoogleNet (Right Answers)
Answer : ABCD
X
Question 16 of171 Time lef\1.28.19
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
15. What is wrong description of the normal distribution?

0 A. In natural phenomena and social phenomena, many random variables obey or approximate a normal
distribution.
0 B The normal distribution lakes the maximum value at lhe mean
0 C The larger the standard deviation the steeper the normal d1stribuhon curve (Right Answers)
0 D. The larger the standard deviation, the slower the normal distribution curve.
Answer : C
34
X
Noc~
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
16. Which of the following options is not a reason for traditional machine learning algorithms to promote the
development of deep learning?
0 A Dimensional disaster
0 B local invariance and smooth regularization
0 C Manifold learning
0 D Feature Engineering !Right Answers)
Answer : D
X
Question 18 of171 Time lef\1.27:41
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
17. The commonly used functions for mathematical operations in Python are basically in lhe math module and
the cmath module.
0 A True (Ri ght Answers)
0 B False
Answer: A
X
Question 19 of 171 Time lef\1.27;25
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
18. What does not belong to supervised learning?

0 A. Logistic regression
0 B Support vector machine
0 C Decision tree
0 D Pnnc1pal component analysis (Right Answers)
Answer : D
X
35
Question 20 of171 Time lef\1.27.11
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
21 The Python list can be identified by · O ·, and the default index of the first element from left to right is 1.
0 A True
Answer: B
X
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
22 What are the core concepts in TensorFlow? (Multiple Choice)
0 A Grtdding
0 B Calculation Chart (Right Answers)
0 C Doi Product
0 D Tensor (Right Answers)
Answer : BO
X
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
19. Which of the following is not included in the recurrent neural network usage scenario?
0 A. Machine translation
0 B Speech recognition
0 C Image style migration (Right Answers)
0 D. Text generation
Answer : C
X
Noce
+) AnyDesk O 473939326
~ • 473939326
36
X+
Huawei H13-311-ENU
0 Sign
Question 24 of 171 Time lef\1.24.28
Noce
20. Numerical calculation refers to the method and process of effectively using the digital computer to solve the
approximate problem of mathematical problems, and the discipline consisting of related
theories.
Which of the following processes are inwl\/8d in solving actual problems with a computer? (Multiple Choice)
0 A Know the actual problem (Right Answers)
0 B Mathematical model (Right Answers)
0 C Numerical calculat,on method (Right Answers)
0 D Programm1ng (Right Answers )
0 E. Computer calculation results (Right Answers>
Answer : ABCDE
X
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
21. Which of the following is not the difference between Python 2 and Python 3?
0 A. print
0 B Unicode
0 C import (Right Answers)
0 D. xrange
Answer : C
X
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
22. Deep learning algorithms can be divided into supervised learning and unsupervised learning.
0 A True (Right Answers)
0 B False
Answer: A
X
Noce
+) AnyDesk O 473939326
~ • 473939326
37
X+
Huawei H13-311-ENU
Huawe1 Certified Network Assoclate-Artficial Intelligence Question 27 of 171 Time Lef\1:23:9
0 Sign
23. The matrix.produced by the exchange of matrix A rows and columns is called the transpose of A. What are
the correct properties of matrix transposition? (Multiple Choice)
0 A (A T ) T = A (Right Answers)
0 B (A+ 8) T = AT +BT (Right Answers>
0 C (!EA) T = !EAT (Right Answers)
0 D. (AB} T = A T +BT
Answer : ABC
Noc~
X
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
24. What quotation marks can the Python language use? (Multiple Choice)
0 A Single quotes (Right Answers)
0 B Double quotes (Right Answers)
0 C Three quotes (Right Answers)
0 D. Four quotes
Answer : ABC
X
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
25. Which of the following data types does Tensorflow not support?
0 A int8
0 B ftoat32
0 C double64 (Righi Answers)
0 D. bool
Answer : C
X
Question 29 of 171 Time lef\1.22·36
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
38
0 Sign
26. In TensorFlow, data is represented in the form of tensors and calculation charts.
0 A True
Answer: B
X
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
31 Python is a fully object-oriented language. Which of the following options belong to the Python object?
(Multiple Choice)
0 A Function (Righi Answers)
0 B Module (Rig ht An swim;)
0 C Number (Right Answers )
0 D Character stnng (Righi Answers)
Answer : ABCD
X
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
Iii
0 Sign
32 Which of the following options belong to Tensorflow concept? (Multiple Choice)
0 A Tensor (Right Answers>
0 B Vanables (Right Answers)
0 C Placeholder (Right Answers )
0 D Operation (Right Answers)
0 E Conversation (Right Answers )
Answer : ABCDE
X
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
39
0 Sign
27. The following are the description of the distribution function. distribution law. and density function of random
variables. which is incorrect?
0 A. Distribution law can only describe the value rule of discrete random variables
0 B The density function can only descnbe the value rule of continuous random vanables
0 C Discrete random vanables have no d1slribulion function (Right Answers)
0 D. Distribution function describes the value rule of random variables
Answer : C
X
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
Iii
0 Sign
28. Which of the following conditions do the randomized trials need to meet? (Multiple Choice)
0 A IL can be repeated under the same conditions. (Right Answers)
0 B There may be more than 011e possible outcome for each tnal. and all possible outcomes of the tnal can
be clanfied m advance (Right Answers)
0 C All results of lhe lest cannot be clarified in advance
0 D It Is not possible to determine which result will appear before conducting a test. (Right Answers)
Answer : ABD
X
Noc~
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
Noce
29. In a neural network, knowing the weight and deviations of each neuron is the most important step. If you
know the exact weights and deviations of neurons in some way, you can approximate any
function
What is the best way lo achieve this?
0 A- Random assignment, pray that they are correct
0 B Search for a combmat1on of weight and deviation until the best value 1s obtained
0 C Assign an initial value to iterallVely update weight by checking the difference between the best value
and the initial (Right Answers)
0 D The abolle is not correct
Answer: C
40
X
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
30. What are the regularizations in deep learning? (Multiple Choice)

0 A L1 norm. L2 norm (Right Answers>
0 B Data set enhancement (Right Answers)
0 C Integration method (Right Answ ers)
0 D Dropout (Right Answers>
Answer : ABCD
X
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
31. Which of the following about the dictionary in Python is correct? (Multiple Choice)
0 A Each key and its corresponding value need to be separated by ": (Right Answers)
0 B Separate the different key-value with • • (Right Answers)
0 C The entire dictionary is included in the • {} • (Right Answers)
0 D The keys of the d1cl1onary are unique and the data type 1s uniform (Right Answers)
Answer : ABCD
X
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
32. The determinant of square matrix A is a scalar.

0 B False
Answer: A
X
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
41
Huawei H13-31 1-ENU
0 Sign
33. Which of the following about the gradient descent is incorrect?

0 A. Random gradient descent is a commonly used one in gradient descent.
0 B Gradient descent includes random gradient descent and batch gradient descent.
0 C The gradient descent algorithm 1s fast and reliable [Right Answers}
0 D. Random gradient descent is one of the commonly used optimization algorithms in deep learning
algonlhms.
Answer : C
X
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
34. Which of the following conditions is not a condition that n-fold bernoulli trials needs to meet?
0 A. Each test was repeated under the same conditions.
0 B There are only two possible outcomes for each trial, Le. event A occurs and event A does not occur.
0 C Each test itself is obeying normal distribution (Righi Answers)
0 D. The results of each tnal are independent of each other.
Answer : C
X
Question 40 of 171 Time Lef\1:10.7
Noc~
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
41 The Python language can use muttiple statements on the same line, separated by commas ".".
0 A. True
Answer: B
X
Noc~
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-31 1-ENU
0 Sign
42
42 Which of the following files is the graph st11.1cture ofTensorFlow saved in?
0 A. checkpoint file
0 B index file
0 C .cpkt file
0 D .meta file (Right Answers)
Answer : D
X
Noc~
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
Noc~
35. Machine Learning Service is a data mining and analytics platform service that helps users quickly discover
data patterns and build predictive models through machine learning techniques and
deploy them as predictive analytics solutions
0 A True (Right Answers>
0 B False
Answer: A
X
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
36. Which of the following about the description of expectations and variances is incorrect?
0 A. Expectation reflects the average level of random variable values
0 B The variance reftects the degree of deviation between the random variable and its mathematical
expectation
0 C Expectation and variance are both numerical characteristics of random variables
0 D The greater the expectation the smaller the variance (Right Answers)
Answer : D
X
Noc~
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
43
37. The Python dictionary is widely identified by " {}·. and the internal data consists of the key and its
corresponding value.
0 B False
Answer: A
X
Noc~
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
38. Which of the following options is not a tensor attribute?

0 A. name (Logo)
0 B shape (Dimensionality)
0 C type (Type)
0 D node (Node) (Rlght Answers)
Answer : D
X
Noc~
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
39. Which of the following products is related to artificial intelligence? (Multiple Choice)
0 A AlphaGo (Right Answers)
0 B Self-dmmg (Right Answers)
0 C Voice input (Right Answers>
0 D Huawe, Mate mobile phone (Right Answers)
Answer : ABCD
X
Question 47 of 171 Time Lefl1:8:2
Noc~
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
40. Python regular expressions are a spedal sequence of characters lhal makes it easy lo check if a string
matches a pattern
44
0 B False
Answer: A
X
Noc~
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
41. What operating systems does Python support? (Multiple Choice)

0 A DOS (Right Answers)
0 B Lmux (Righi Answers)
0 C Mac OSX (Right Answers}
0 D Windows (Righi Answers )
Answer : ABCD
X
Noc~
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
42. Which of the following about the description of the number of rows and columns for the determinant is
correct?
0 A. The number of rows 1s greater than the number of columns
0 B The number of rows 1s equal lo the number of columns (Right Answers>
0 C The number of rows is less than the number of columns
0 D. The number of rows has no relationship with the number of columns.
Answer : B
X
Noc~
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
51 Deep learning makes it easy to derive simple mathematical functions from a large amount of high-
dimensional data to describe complex decision interfaces.
0 A. True
Answer: B
45
X
Noc~
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
52 Which of the following does not belong to automatic hyperparameter optimization algorithm?
0 A. Grid search
0 B Random gradient descent (Right Answers)
0 C Random search
0 D. Model-based hyperparameter optimizat,on
Answer : B
X
Question 52 of 171 Time left0.59;51
Noc~
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
43. What are the characteristics of Python code? (Multiple Choice)

0 A leg1bil1ty (Right Answers)
0 B Simpl1c1ty (Right Answers)
0 C Rapidity
0 D Scalab1hty (Right Answers)
Answer : ABD
X
Question 53 of 171 Time left0.58.15
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
44. What are the commonly used activation functions? (Multiple Choice)
0 A s1gmo1d (Right Answers)
0 B tanh (Right Answers)
0 C relu (Right Answers)
0 D. danish
Answer : ABC
• HCIAAI
·..:,
46
Problem List Check The Sign Quesllons
X
Question 54 of 171 Time left0.57:52
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
0 A s1gmo1d (Right Answers)
0 B tanh (Right Answers)
0 C relu (Right Answers)
0 D. danish
Answer : ABC
• HCIAAI
·..:,
Problem List Check The Sign Quesllons
X
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
46. What is the incorrect relationship between neural networks and deep learning?
0 A. The concept of deep learning sterns from the study or artificial neural networks.
0 B A neural network algorithm with multiple hidden layers is a deep learning algorithm
0 C Single layer neural networks are also a type of deep learning (Right Answers)
0 D. Convolutional neural networks are a type of deep learning.
Answer : C
X
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
47. Which of the following environments does not support to install the T11nsorFlow?
0 A. Linux
0 B Mac OS
0 C Docker
0 D OpenStack (Right Answers)
Answer : D
47
X
Question 57 of 171 Time left0.56·30
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
48. Deep learning is a branch of machine learning

0 B False
Answer: A
X
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
49. The trace operation returns the sum of the diagonal elements of the matrix. Therefore, the trace of matrix A
and its transposed matrix are equal
0 B False
Answer: A
X
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
50. GAN is a deep learning model and one of the most promising methods for unsupervised learning in complex
distribution in recent years.
0 B False
Answer: A
X
Noce
+) AnyDesk O 473939326
~ • 473939326
48
X+
Huawei H13-311-ENU
0 Sign
51. Which of the following options is not central to linear algebra?

0 A Probability theory (Right Answers)
0 B l inear transformation
0 C Matrix theory
0 D. Vector space
Answer : A
X
Noce
+) AnyDesk O 473939326
~ • 473939326
X+
Huawei H13-311-ENU
0 Sign
62 A scalar k is muhiplied by matrix A equal to k and each of the numbers in A is multipITed.
0 B False
Answer: A
X
Noc~
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
Huawe1 Certified Network Associate-Artficial Intelligence
0 Sign
52. What are the conditions for m row n column matrix A and p row q column matrix B to be multiplied?
D A. m=p. n=q
0 B n=p (Right Answers)
0 C m=n
0 D. p=q
Answer: B
Problem ust Check The Sign Ouesllons Show Correct Answers
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
64 Which ofthe following are the characteristics ofTensorFlow? (Multiple choice)
0 A Open source (Right Answers)
0 8 Convenience (Right Answers )
49
0 C Mature (Right Answers)
0 D Flexible (Right Answers)
Answer: ABCD
Problem usl Check The Sign Ouesllons Show Correct Answers
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
53. Which is the correct description of the python creation function? (Multiple choice}
0 A The function created starts with the def keyword followed by the function name and parentheses. (Ri ght
Answers)
0 8 The parameters need to be placed in parentheses (Right Answers)
0 C The function content starts with a colon and needs lo be indented (Right Answers}
0 D Return the result with return and the function ends. (Right Answers)
Answer: ABCD
X
Question 65 of 171 Time left0.51;28
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
66 About Python correct tuple, which is correct description? (Multiple choice)
0 A The difference between a list and a tuple is that the elements of the tuple cannot be modified fRight
Answers)
0 8 The list uses square brackets and the tuple uses parentheses (Right Answers)
0 C Tuple creation is very simple just add elements in parentheses and separate them wrth commas (Right
Answers)
0 D When a tuple contains only one element you need to add a comma after the element i£ (Right
Answers>
Answer: ABCD
Problem ust Check The Sign Questions Show Correct Answers
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
67 What research areas does artificial intelligence have? (Multiple choice)
0 A Natural language processing (Right Answers)
0 8 Computer Vision (Right Answers)
0 C Machine learning (Right Answers )
0 D Speech recognition (Right Answers)
Answer: ABCD
50
X
Question 67 of 171 Time Left0:51:0
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
54. Tensorflow supports multi-TPU cluster computing

0 B False
Answer: A
Problem list Check The Sign Ouesllons Show Correct Answers
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
55. As shown below, which are the following matrix A characteristic value? (Multiple choice)
0 A 2 (Right Answers)
0 B -2
O C. 4
0 D 4 (Right Answers)
Answer: AD
X
Question 69 of 171 Time left0.50·31
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
56. Only matrix A and matrix B have the same number of rows and columns. A and B can be added.
0 B False
Answer: A
X
Question 70 of 171 Time Left0:50: 1
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-31 1-ENU
0 Sign
71 The continue statement in the Python language is used to jump out of this loop and then continue to
0 B False
Answer: A
51
X
Question 71 of 171 Time left0:49:48
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
72 Which of the following are the cloud services provided by Huawei Cloud El Visual Cognition? (Multiple
choice)
0 A Text recogmlton (Right Answers)
0 8 Face recognilton (Right Answers)
0 C Image recogmlton (Right Answers )
0 D Content detection {Right Answers)
0 E Image processing (Right Answers>
Answer: ABC.DE
Problem ust Check The Sign Ouesttons Show Correct Answers
X
Question 72 of 171 Time left0:49·31
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
57. X. Y are random variables, C is a constant, the nature of the difference in the following options which is
wrong? (Multiple choice)
0 A. D(C)=O
0 8 D(X+ Y)=D(X)+D(Y) (Right Answers)
0 C D(CX)=C'C'D()()
0 D D(XY)=D(X)D{Y) (Right Answers)
Answer: BO
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
74 Vector is a number.
D A. True
0 8 False (Right Answers)
Answer: 8
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
52
0 Sign
58. About Bayesian formula- P(WlX}=P(XlW)*P(WyP(X) What is the correct description?

0 A. P(WIX) is a prior probability
0 8 P(XIW) 1s a cond1t1onal probabihty (Right Answers)
0 C P(W) is the posterior probability
0 D. P(X) is the posterior probability
Answer: 8
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
76 Gyclic neural networks are more suitable for deafing with image recognition p<oblems.
D A. True
Answer: 8
X
Question 76 of 171 Time left0:48.14
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
77 Tensorflow Operations and Computation Graph are not - run in the Session
D A. True
Answer: 8
Problem ust Check The Sign Ouesl.lons Show Correct Answers
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
59. What is in the category of artificial intelligence? (Multiple choice)

0 A Action (Right Answers)
0 8 PerceptJon (Right Answers)
0 C Cognition (Right Answers)
0 D. None of the above
Answer: ABC
X
+) AnyDesk O 473939326 X +
53
GJi • 473939326
Huawei H13-311-ENU
0 Sign
60. The Python language does not allow embedding another loop body in a loop body
D A. True
Answer: 8
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
61. What are the mainstream deep learning open source tools? (Multiple choice)
0 A TensorFlow (Right Answers>
0 8 Caffe (Right Answers>
0 C Torch (Right Answers)
0 D Theano (Right Answers>
0 E Scikit..Jeam
Answer: ABCD
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
81 The for loop statement in the Python language can iterate through the items in any sequence.
0 B False
Answer: A
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
82 Corwolutional neural networks are more suitable for image recognition problems than cyclic neural
networks.
0 B False
Answer: A
X
54
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
62. Huawei Cloud El builds enterprise intelligence seivices based on three,tier seivices. Which of the following
options does not belong to Layer 3 seivices?
0 A. Basic platform seivices
0 B General domain sell/ices
0 C Industry sector selVices
0 D Integration seMces (Ri ght Answers>
Answer: D
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
84 Which ofthe following is true about unsupeNised learning?
0 A UnsupeMsed algonthm only p,ocesses ·teatures· and does not process ·tags· (Right Answers>
0 B Dimensionality reduction algorithm is not unsupel'lised learning
0 C K-means algorithm and SVM algorithm belong lo unsupervised learning
Answer: A
Problem usl Check The Sign Questions Show Correct Answers
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
63. Which ofthe following options is not the Tensorflow build process?
0 A. Building a calculation chart
0 B Input tensor
0 C Generate a session
0 D Update weights I Right Answers)
Answer: D
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-31 1-ENU
0 Sign
86 What model is not a cyclic neural network?
55
0 A. RNN
D B LSTM
0 C GBOT (Right Answers )
0 D. GRU
Answer: C
Problem ust Check The Sign Ouestlons Show Correct Answers
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-31 1-ENU
0 Sign
87 Under the large-scale relationship analysis scenario, which of the following options does not belong to
the three high demands of massive relationship processing?
0 A. Efficient relationship discovery data for massive data
0 8 Efficient shanng of massiVe data (Right Answers)
0 C Efficient storage and access requirements for massive amounts of data
0 D. High scalability and high availability requirements for relational analysis platforms
Answer: 8
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
64. What are the commonly used loss functions? (Multiple choice)
0 A Mean variance (Ri ght Answers)
0 8 Sigmoid cross entropy (Right Answers}
0 C Soft max cross enl10py (Right Answers)
0 D Sparse cross entropy (Right Answers)
0 E Weighted S1gmo1d Cross Entropy (Right Answers)
Answer: ABCDE
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
65. What are the steps that are not part of the operation of a Python file object?
0 A. open
0 8 delete (Righi Answers)
0 C read
0 D. write
0 E close
Answer : 8
56
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
66. Tensorflow is the second generation of artificial intelligence learning system developed by Google based on
( ).
0 A DistBelief (Right Answers>
0 B PaleyFunction
0 C ConvexOne
0 D. Infinity
Answer: A
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
91 The Python language can use the "#" at the beginning of a single line of code for code comments .
0 B False
Answer: A
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
92 According to the definition of information entropy what is the bit entropy of throwing a uniform coin?
O A.O
0 B 0.5
0 C 1 (Right Answers)
0 D. -1
Answer: C
Problem List Check The Sign Questions Show Correct Answers
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
Huawe1 Certified Network Associate-Artficial Intelligence Question 93 of 171 Time left0:43:41
0 Sign
57
67. Artificial intelligence is currently suitable for scenarios with I'known environment, clear objectiws, and
p,edictable actionsj±. Deep learning in the fields of image recognition, speech recognition,
translation, etc , artificial intelligence basically has human recognition ability. even surpassing human beings
Based on these capabilities. it has been applied to many scenes such as medical care
and public safety However, there is still a lack of reasoning and cognition
0 A True {Right Answers)
0 B False
Answer : A
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
Huawe1 Certified Network Associate-Artficial Intelligence Question 94 of 171 Time Left0:43:2
0 Sign
94 Python authors deliberately design very restrictive grammars that make bad programming habits (such
as the next line of if statement not indented to the right) cannot be compiled
0 B False
Answer: A
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
68. The timestamps in the Python language- are represented by how long (in seconds) elapsed from midnight
(epoch) on January 1, 1970
0 B False
Answer: A
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
96 Corwolutional neural networks are more suitable for dealing with speech recognition problems
D A. True
Answer: 8
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
58
0 Sign
97 What are the results returned by the if conditional statements in the Python language? (Multiple choice)
0 A. TRUE (Right Answers)
0 B. O
Oc FALSE
O D. null
Answer: TC
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
69. Which of the following description of the validation set is wrong?

0 A The verification set can coincide with the test set. (Right Answers>
0 B The test set can coincide with the training sel
0 C The subset used to pick hyperparamelers is called a validation set
0 D. Typically 80% of the training data 1s used for training and 20% 1s used for verification.
Answer: A
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
70. Which of the following can optimization problem be classified according to the constraints? (Multiple choice)
0 A equality constraints (Right Answers)
0 8 Inequality conslfamts (Right Answers)
0 C Unconstrained conditions (Righi Answers)
0 D. Semi-constraint conditions
Answer: ABC
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
71. Which of the following are the characteristics of the Python language? {Multiple choice)
0 A Explanatory (Right Answers>
0 B Process oriented
0 C Object-oriented !Right Answers)
0 D Dynamic data type (Right Answers)
Answer: ACD
59
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
101 Which of the following are the elements of artificial intelligence? (Multiple choice)
0 A Algorithm (Right Answers)
0 8 Computing power (Right Answers)
0 C Dal a (Rig ht Answers)
0 D Scene (Right Answers)
Answer: ABCD
Problem list Check The Sign Questions Show Correct Answers
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
102 Which of the following is not a specific technology of artificial intelligence?
0 A. Knowledge map
0 B Machine translation
0 C Riemann geometry (Righi Answers)
0 D. Semantic understanding
Answer: C
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
72. Which of the following statements about O\/llrflow and underflow in numerical calculations is correct?
(Multiple choice)
0 A Underflow occurs when the number close to zero Is rounded to zero (Right Answers)
0 B O;erflow occurs when the number close to zero 1s rounded to zero
0 C Overflow occurs when a large number of levels are approximated as 1P or - 1P (Right Answen )
0 D. Underflow occurs when a large number of levels are approximated as iP or - 1
P
Answer: AC
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
60
0 Sign
104 Which of the following description about Python regular expression is not correct?
0 A. Python has added the re regular expression module since version 1.5.
0 B re regular expression module gives the Python language all the regular expression functionality.
0 C re regular exp<essions are powerful tools for working with strings.
0 D re regular expressions can only process stnng data not numeric data (Right Answers)
Answer: D
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
73. In which year was the first introduction of "artificial intelligence"?

D A. 1946
0 B 1960
0 C 1916
0 D 1956 (Righi Answers)
Answer: D
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
74. Which of the following description about the relationship between the Poisson distribution and the binomial
distribution is correct?
0 A. The mathematical model of the Poisson distribution and the binomial distribution are both Lagrangian
models.
0 8 Poisson d1stnbutJon 1s an approx1mallon of the bmom1al d1stnbubon when n 1s very large and p 1s
small (Right Answer.;)
0 C Poisson distribution has no relationship with binomial distribution
0 D. Poisson distribution can replace binomial distribution
Answer: B
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-31 1-ENU
0 Sign
107 Which of the following is not an artificial intemgence school?
0 A. Symbolism
0 8 StatJst1caltsm (Right Answers)
61
0 C Behaviorism
0 D. Connectionism
Answer: B
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
75. Which of the following does the convolutional neural network structure not include?
0 A. Convolutional layer
0 B Pooling layer
0 C Loop layer (Righi Answers)
0 D. Fully connected layer
Answer: C
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
76. Which of the following functions can numerically stabilize overflow and underflow?
0 A. Softminus function
0 B Softplus function
0 C Soft max function (Righi Answers)
0 D. Softmin func!Jon
Answer: C
Problem ust Check The Sign Ouesuons Show Correct Answers
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
77. What are the advantages of Python? (Multiple choice)

0 A Simple (Righi Answers)
0 8 Free (Right Answers)
0 C High-level langvage (Right Answers)
0 D Rich library (Right Answers)
Answer: ABCD
X
+) AnyDesk O 473939326 X +
GJi • 473939326
62
Huawei H13-311-ENU
0 Sign
111 Python script execution mode includes interactive mode and script mode
0 B False
Answer: A
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
112 Which of the following features does Python support? (Multiple choice)
0 A Inheritance (Right Answers )
0 8 Heavy load (Right Answers>
0 C Denved (Right Answers)
0 D Multiple inhentance (Right Answers)
Answer: ABCD
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
78. Which of the following description of the number of rows and columns of the matrix is correct?
0 A. The number of rows 1s greater than the number of columns
0 B The number of rows 1s equal to the number of columns
0 C The number of rows is less than the number of columns
0 D The number of rows has no relat,onsh,p with the number of columns. (Right Answers)
Answer: D
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
114 Which of the following is the computing model ofTensorFlow?
0 A Calculauon chart (Right Answers )
0 B Tensor
0 C Conversation
0 D. Vanable
Answer: A
63
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
79. Which of the following can improve the computational efficiency of the neural network model? (Multjple
choice)
0 A TPU (Right Answers)
0 8 GPU (Right Answers)
0 C FPGA (Right Answers)
0 D Large-scale distributed cluster (Right Answers>
Answer: ABCD
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
80. Which of the following statements about Python are correct? (Multiple choice)
0 A Invented m 1989 by the Dutch Guido van Rossum the first public release was issued in 1991. (Right
Answers)
0 8 Python 1s purely free software and the source code follows the GPL (GNU General Public License)
protocol (RightAnswers)
0 C Python syntax is simple and clean One of the features is to force blank characters to be used as
statement 1ndentat1on [Right Answers)
0 D Python 1s often nicknamed glue language which can easily connect various modules made 1n other
languages (Right Answers>
Answer: ABCD
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
117 Which of the following does not belong to long,term memory LSTM (long Short·Term Memory) network
architecture?
0 A Memory door (Right Answers)
0 B Forget the door
0 C Input door
0 D. Output door
Answer: A
X
64
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
118 In order for a machine to be intelligent, it must be knowledgeable. Therefore, there is a research field in
artificial intelligence. which mainly studies how computers automatically acquire knowledge
and skills to achieve self-impr011emenL What is the branch of this research called?
0 A. Expert system
0 B Machine learning (Right Answers>
0 C. Neural Network
0 D. Natural language processing
Answer: B
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
81. Which of the following descriptions about Python program are correct? (Multiple choice)
0 A The program consists of modules (Right Answers)
0 8 Module contains statements (Right Answers)
0 C statement contains an expression (Right Answers)
0 D Expressions create and process objects (Right Answers)
Answer: ABCD
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
82. Huawei Machine learning SeNice MLS is a one-stop platform that supports the entire process of data
analysis. Which of the following is not a feature of MLS?
0 A. A rich libfary of machine learning algorithms.
0 B machine learning program is intuitive and easy to use.
0 C Distributed and scalable big data computing engine.
0 D Support for the R language but does not support the Python language (Right Answers)
Answer: D
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
121 Training error will reduce the accuracy of the model and produce under-fitting. How to improve the
65
0 A lnc,ease the amount of data (Right Answers)
0 8 Feature Engineenng (Right Answers)
0 C Reduce regularization parameters
0 D Add features (Right Answers)
Answer: ABD
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
122 Which of the following is not a way for the TensorFlow program to read data?
0 A. Preload data
0 B Feeding data
0 C Read from the file
0 D Write a file format reader (Right Answers)
Answer: D
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
83. Self-encoder is an unsupervised learning algorithm

0 B False
Answer: A
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
124 The number or matrix is stored in the tensor
D A. True
Answer: 8
Problem ust Check The Sign Ouestmns Show Correct Answers
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
66
0 Sign
84. Which of the following descriptions about the Recurrent Neural Network (RNN) is correct?
0 A Can be used to process sequence data. (Right Answers)
0 B Cannot process variable length sequence data.
0 C Unlike convolutional neural networks. parameters of RNN cannot be shared
0 D. The units above the hidden layer are not associated With each other.
Answer: A
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
Huawe1 Certified Network Associate-Artficial Intelligence Question 126 of 171 Time left0:31:36
0 Sign
85. Artificial intelligence is a new technical science that studies and develops theories, methods and application
systems for simulating, extending and extending human intelligence. It is one of the
core research areas of machine learning.
0 A. True
0 B False (Righi Answers)
Answer: B
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
127 The TensorFlow framework does not support Windows systems.
D A. True
Answer: 8
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-31 1-ENU
0 Sign
86. Which of the following descriptions of the depth feedforward network is correct?
0 A Deep feedforward network 1s a kind of neural network. (Right Answers)
0 B The deep feedforward network has only one hidden layer.
0 C The unit on lhe hidden layer of the deep feedfoiward network will have countless.
0 D. Deep feedfoiward network 1s used to deal with linear problems.
Answer: A
Problem ust Check The Sign Ouestlons Show Correct Answers
X
67
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
87. The time interval in the Python language is a Hoating-point fraction in seconds
0 B False
Answer: A
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
88. Which of the following is not part of the deep learning development framework?
0 A. CNTK
0 B Keras
0 C SAFA (Right Answers )
0 D. MXNet
Answer: C
X
Question 130 of 171 Time Left0:30.7
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
131 Which of the following are included in the application scenarios ofTensorflow? (Multiple choice}
0 A Speech recognition (Right Answers)
0 8 Face recognitJon (Right Answers)
0 C Image style changes (Right Answers)
0 D Autopilot (Right Answers)
Answer: ABCD
Problem ust Check The Sign OuestJons Show Correct Answers
X
Question 131 of 171 Time left0;29:34
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
132 Which of the following neural network structures will share weights? (Muttiple choice)
0 A Convolutional neural network (Right Answers)
0 8 Recurrent neural netwoflc (Rlghl Answers)
0 C Fully connected neural network
0 D. All of the above
68
Answer: AB
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
Huawe1 Certified Network Associate-Artficial Intelligence Question 133 of 171 Time left0;28;29
0 Sign
89. The correlation coefficient. also known as the linear correlation coefficient, is used lo measure the linear
relationship between two variables, which is a real number greater than zero
D A. True
Answer: 8
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-31 1-ENU
0 Sign
134 X. Y are random variables. C is a constant. which of the following description about the nature of the
mathematical expectation is incorrect?
0 A. E(C) = CA. E(C) = C
0 B E(X+Y) = E(X)+E{Y)
0 C E(CX) = CE(X)
0 D E(XY) = E(X)E{Y) (Right Answers>
Answer: D
Problem ust Check The Sign Quest.Jons Show Correct Answers
X
Question 134 of 171 Time left0;28.15
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
90. Which of the following statement abo~t the matrix is incorrect?

0 A. The arbitraiy matrix is multiplied by the unit matrix and will not change.
0 B The transpose of the symmetric matrix A 1s also A itself.
0 C The transpose of the orthogonal matrix A is equal to the inverse of A.
0 D There 1s no inverse matnx for the diagonal matnx. (Right Answers)
Answer: D
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-31 1-ENU
0 Sign
69
91. Which is not a deep learning algorithm?
0 A. Setr-encoder
0 B Convolutional neural networks
0 C Recurrent neural networks
0 D Support vector machine (Right Answers)
Answer: D
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
137 TensorFlow is Google's first generation of proprietary machine learning systems
D A. True
Answer: 8
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
92. Which of the following schools does the neural network study belong to?
0 A. Symbolism
0 8 ConnectJonism (Righi Answers)
0 C Behaviorism
Answer: B
Problem ust Check The Sign OuestJons Show Correct Answers
X
Question 138 of 171 Time Left0:27. 7
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
93. Reducing the gap between the training error and the test error will result in over-fitting How to prevent over-
fitting? (Multiple Choice)
0 A Cross validation (Right Answers)
0 8 Integration method (Right Answers )
0 C Increase regulanzation !Right Answers )
0 D Feature Engmeenng !Right Answers)
Answer: ABCD
Problem list Check The Sign Questions Show Correct Answers
X
70
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
94. Which is wrong with the Python module time introduction method?
0 A. import time
0 B from time import •
0 C import time as t
0 D from llme (Right Answers)
Answer: D
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
141 In what year did Huawei officially provide services in the form of cloud services, and combined with
more partners to provide richer artificial intelligence practices?
D A. 2002
0 B 2013
0 C 2015
0 D 2017 (Right Answers)
Answer: D
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
142 Which description is wrong about the lambda function of the Python language?
0 A. lambda is just an expression, and the function body is much simpler than def.
0 B The body of a lambda can be an expression or a block of code
0 C The lambda function can access the parameters in the global namespace.
0 D The lambda function accepts only one parameter value (Right Answers)
Answer: D
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-31 1-ENU
Huawe1 Certified Network Associate-Artficial Intelligence Question 143 of 171 Time left0;25:11
0 Sign
95. The meaning of artificial intelligence was first proposed by a scientist in 1950, and at the same time a test
model of machine intelligence was proposed Who is this scientist?
71
D A. Minsky
0 B Zade
0 C Tunng (Right Answers )
0 D. Von Neumann
Answer: C
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-31 1-ENU
Huawe1 Certified Network Associate-Artficial Intelligence Question 144 of 171 Time left0;24:55
0 Sign
144 The number of rows and columns of the three matrices A , B. and C is 3 rows and 2 columns, 2 rows. 3
columns. 3 rows and 3 columns. which of the following operations is meaningful?
D A.AC
0 8 BC (Right Answers)
O c MB
0 D. AB--BC
Answer: 8
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
96. The constituent elements of the Python identifier include: numbers, letters, and underscores.
0 B False
Answer: A
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
97. Which description is wrong about the hyperparameter?

0 A. 1-typerparameters are parameters that set values before the algorithm begins learning.
0 B Most machine learning algorithms have hyperparameters.
0 C Hyperparameters cannot be modified (Right Answers}
0 D. The value of the hyperparameter is not learned by the algonthm itself.
Answer: C
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
72
0 Sign
14 7 Which of the following is not a module in the Tensorflow library?
0 A. tf.nn
0 B tf. layers
0 C tf boost (Right Answers)
0 D. tf.contrib
Answer: C
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
98. As shown in the figure below, what is the value of the determinant A?
123
A=456
789
0 A24
0 B. 18
0 C. -24
0 D O (Right Answers>
Answer: D
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
14 9. How many spaces does PEP 8 stipulate that Python's first fine indentation needs to indent?
D A.1
0B2
0 C 4 (Right Answers)
0 D. 8
Answer: C
X
Question 149 of 171 Time left0;23.12
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
99. What are the Python language data types? (Multiple Choice)
0 A numbers (Right Answers)
0 8 slnng (Right Answers)
73
0 C list (Right Answers)
0 D tuple (Righi Answers)
0 E d1ct1onary (Righi Answers)
Answer: ABCDE
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
151 What are the algorithms supported by Tensorflow? (Multiple Choice}
0 A GNN (Right Answers)
D B ZNN
0 C RNN (Rig ht Answers)
0 D. HUT
Answer: AC
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-31 1-ENU
0 Sign
152 Which is not a Python operator?
0 A. arithmetic operator
0 8 Reasoning operator (Right Answers)
0 C Logical operator
0 D. Companson operator
Answer: 8
Problem usl Check The Sign Questions Show Correct Answers
X
Question 152 of 171 Time Left 0: 11 :0
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
00. Which method is not supported in TensorFlow to define variables?

0 A. Random number
0 B Constant
0 C Calculated from the initial values of other variables
0 D null (Right Answers)
Answer: D
X
+) AnyDesk O 473939326 X +
74
GJi • 473939326
Huawei H13-311-ENU
0 Sign
154 What does factors thal promote lhe development of artificial intelligence not include?
0 A. Big data
0 B Computing ability
0 C Algorithm theory
0 D Blockcham (Right Answers )
Answer: D
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
01. Which of the following are solutions for the Huawei Cloud El industry scenario? (Multiple Choice)
0 A Intelligent Logistics (Right Answers)
0 8 Intelligent Water (Right Answers)
0 C lntell,genl transportation (Right Answers)
0 D Intelligent Finance (Right Answers)
0 E Intelligent manufactunng (Right Answers)
Answer : ABCDE
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
02. Vector group a1=(1, 1, 1), a2=(0,2.5), a3=(1.3,6), which of the following options is correct?
0 A Linear correlation (Right Answers>
0 B Linear independence
0 C aHa2+a3=0
0 D. 2a1+a2+a3=0
Answer: A
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
157 What are the scenarios or industries that are suitable for using Python? (Multiple Choice)
0 A Artlfic1al intelligence (Right Answers)
0 8 web development (Right Answers)
75
0 C Game development (Right Answers)
0 D Hardware development
Answer: ABC
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-31 1-ENU
0 Sign
03. Which descriptions are correct about python·s index? (Multiple Choice)
0 A Index from left to right defaults from O (Right Answers )
0 B Index from left to right defaults from 1
0 C Index from right lo left defaults from -1 (Right Answers )
0 D. Index from right to left defaults from 0
Answer: AC
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-31 1-ENU
0 Sign
04. Huawei firmly believes that the value of Al is ultimately reftected in solving the actual business problems of
the enterprise. Therefore, Huawei is firmly committed to the development of Al
technology, focusing on domain solutions, focusing on basic technologies and enabling platforms, and
working with industry l)(actice leaders to develop industry solutions
0 A. True (Right Answers)
0 B False
Answer: A
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
05. The Python dictionary is identified by ·o·. and the internal data consists of the key and its corresponding
value.
0 B False
Answer: A
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
76
0 Sign
161 In Huawei Cloud El Enterprise Intelligence, which basic platform seNices are included? (Multiple
Choice)
0 A Machine learning (Right Answers)
0 8 Deep learning (Right Answers)
0 C Graph engine (Right Answers)
0 D Batch processing (Right Answers)
Answer: ABCD
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
162 Linear algebra is a branch of algebra that deals with linear problems.
The linear relationship is that the relationship between exponential objects is expressed in ( ) times form?
0 A. 1 (Right Answers)
0B2
0 C. 3
0 D. 4
Answer: A
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
06. "print· in Python 3 must be used with ·o·. 0 A True (Right Answers)
0 B False
Answer: A
X
Question 16:l of 171 Time Left0:5:24
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
164 Deep learning is different from machine learning and there are no unsupervised algorithms
D A. True
Answer: 8
X
Question 164 of 171 Time Left0:5·3
77
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
07. What are the implementation modes ofTensorflow? (Multiple Choice)

0 A Stand-alone mode (Right Answers )
0 8 D1stnbuted mode (Right Answers)
0 C Reverse mode
0 D. Forward mode
Answer: AB
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
08. What are the application scenarios for the break statement in the Python language? (Multiple Choice}
0 A. Any Python statement
0 8 whde loop statement (Right Answers>
0 C for loop statement (Right Answers)
0 D Nested loop statement (Right Answers)
Answer: BCD
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
167 TensorFlow only supports Python and R language
D A. True
Answer: 8
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-31 1-ENU
0 Sign
09. What are the commonly used gradient descent optimization functions? (Multiple Choice)
0 A Random gradient descent (Ri_ghl Answers)
0 8 Adadelta (Right Answers)
0 C Adagrad (Right Answers}
0 D momentum (Right Answers)
78
0 E RMSProp (Right Answers>
Answer : ABCDE
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
10. Whal is not the optimization method in deep learning?

0 A. Random gradient descent
0 B Back propagation algorithm
0 C Principal component analysis (Right Answers )
0 D. Momentum
Answer: C
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
170 . Information theory is a branch of applied mathematics. The main research is to quantify how much
D A. Minsky
0 B Turing
0 C Simon
0 D Shannon (Right Answers)
Answer: D
X
+) AnyDesk O 473939326 X +
GJi • 473939326
Huawei H13-311-ENU
0 Sign
171 . Which of the following description 1s wrong about the image recognition cloud service?
0 A. The service identifies specific objects such as skyscrapers, buildings, and ships.
0 B The service can also identify the sea, city, port and other scenes.
0 C The service can perform semantic analysis on the content in the image to extract labels with specific
abstract meanings
0 D The seMce can prOVlde ob1ect and scene labels and can also prOV1de more conceptual labels. (Right
Answers>
Answer: D
X
79
The speed of grid search must be faster than random search.
Answer: False
The biggest difference between logic regression and perceptron is that logistic regression uses the
Sigmoid activation function at the output layer, whereas perceptron uses the sign function. In addition,
neither of the two models can resolve the XOR problem.
Answer: True
HiLens can help small and micro companies realize smart access control.
Answer: True
Polynomial regression cannot resolve the nonlinear problem.
Answer: False
Softmax can’t be used for multiple classification problems.
Answer: False
The Adam optimizer in TensorFlow 2.0 is the built-in optimizer of the Keras module.
Answer: True
The HiAI mobile computing platform architecture supports various mainstream front-end frameworks,
such as TensorFlow and Caffe.
Answer: True
To install TensorFlow of the GPU version, you need to install CUDA and cuDNN.
Answer: True
In a feedforward network, neurons in the same layer are connected to each other.
Answer: False
Backpropagation makes gradient descent more efficient by means of dynamic programming.
Answer: True
In a medical image recognition task, we usually are more concerned about the recall ratio than the
precision ratio when evaluating model performance.
Answer: True
After the image label service is called, only one corresponding label value is returned.
Answer: False
Similar to hidden layer, the quantity of neurons at both an input layer and an output layer of a neural
network is a hyperparameter that can be adjusted.
Answer: False
Cifar10 is a built-in data set of Keras.
Answer: True
The generalization error will continue to decrease as the complexity of the model increases.
Answer: False
MindSpore supports GPU acceleration.
Answer: True
Voice recognition refers to the recognition of text data as audio data.
Answer: False
PyTorch supports GPU acceleration.
Answer: True
In the KNN model, a small K value may cause model underfitting.
Answer: False
Malformed data is not dirty data.
Answer: False
Factory A produces 60% of an automotive company’s products, and factory B accounts for 40%. The
defective rate of factory A is 1%, and that of factory B is 3%. If a vehicle is defective, what is the
possibility that the vehicle is produced by factory A?
Answer: B. 1/3
Which of the following statements about dropout is incorrect?
Answer: D. dropout can’t be added during inference.
80
Which of the following reasons is the main reason why deep learning developed rapidly after 2010?
Answer: D. Improvement of computing power
Which of the following methods cannot resolve overfitting in a decision tree?
Answer: D. Use the integration algorithm to set up an AdaBoost model.
Which of the following network layers does not need to be included in a convolutional neural network?
Answer: C. Pooling layer
Which of the following is not a feature of the Atlas 500 AI edge stations?
Answer: C. High energy consumption
What determines the number of neurons in the input layer of the neural network?
Answer: Data dimension
What operation does “isinstance” represent in TensorFlow 2.0?
Answer: C. Determine whether it is Tensor
Which of the following schools does the expert system belong to?
Answer: A. Symbolism
Input_shape = (4, 28, 28, 1)
X = tf.random.normal(input_shape)
Y = tf.keras.layers.Conv2D(filters=2, kernel_size=3, strides=(1, 1), padding=’valid’, activation=’relu’,
Input_shape=input_shape[1:])(x)
print(y.shape)
What is the output of the preceding code?
Answer: B. (4,26,26,2)
Without Padding, use a 3X3 convolution kernel to convolve a 5X5 image with stride 1. Which of the
following options is the size of the output image?
Answer: A. 3X3
When was the Turing Test proposed?
Answer: A. 1950
Which of the following options is the architecture adopted by Huawei Atlas series chips?
Answer: D. Da Vinci
In the process of training the neural network, which of the following methods is used to update the
parameters so as to minimize the loss function?
Answer: C. Gradient descent
Which of the following options is not the feature of ModelArts?
Answer: C. Support automatic tuning
In machine learning, how many support vectors can the SVM algorithm have?
Answer: D. Several
import numpy as np
n = np.arange(24).reshape(2,-1,2,2)
Which of the following is the return result of n.shape after running the preceding code?
Answer: A. (2,3,2,2)
Which of the following options is the artificial intelligence development framework launched by
Huawei?
Answer: C. MindSpore
Which of the following options is a commonly used loss function for classification problems?
Answer: B. Cross Entropy Loss Function
Which of the following problems can’t be handled by the perceptron?
Answer: C. XOR
Which of the following models is not suitable for the scenario where the covariance between data
features is large?
Answer: C. Naive Bayes
Which of the following options is Huawei’s artificial intelligence computing platform?
Answer: D. Atlas
81
A company wants to use the HiAI platform to develop its own APP. Which of the following IDEs can be
used in conjunction with the HiAI platform?
Answer: B. Android Studio
Which of the following options is the classification of the AI chip business from the application?
Answer: C. Training and Inference
Which of the following options belong to the activation function of the neural network?
Answer:
A. Sigmoid
B. Softmax
C. Softplus
D. ReLu
Which of the following statements about LSTM are correct?
Answer:
A. The LSTM unit has three types of gates: input gate, forget gate, and output gate.
B. The input of one time sequence of LSTM is concatenated with the output of its last time sequence.
C. If peephole is used in LSTM, the vector of the output and memory unit of the previous time
sequence are concatenated with the input vector.
D. In LSTM, if the dimension of the input vector is 200 and the dimension of the hidden layer is 100,
and the peephole is not considered, the parameter quantity at this layer is calculated as follows: 4x
((100 + 200) X 100 + 100)
Which of the following options are the main advantages of the Keras interface in TensorFlow 2.0?
Answer: A,B,C
A. Convenient for users
B. Modular and combinable
C. Easy to expand
Which of the following options belong to the application scenarios of natural language processing?
Answer: B & D
B. Machine translation
D. Public opinion monitoring
Which of the following may cause overfitting?
Answer: B,C,D
B. The model capacity is large
C. The training set includes a large amount of dirty data
D. The number of training sets is small
Which of the following are common Python computer vision libraries?
Answer: C & D
C. matplotlib
D. cv2
Which of the following options belongs to Huawei HiAI platform content?
Answer: A, B, D
A. HiAI Service
B. HiAI Foundation
D. HiAI Engine
What are the elements of AI?
Answer:
A. Algorithms
B. Computing power
C. Data
D. Scenarios
Which of the following are major functions of Mind Expression in MindSpore?
Answer:
82
A. Operator-level automatic differentiation
B. Automatic parallelization
C. Automatic operator generation
D. Semi-automatic data labeling
Which of the following are the problems you may encounter when training a generative adversarial
network?
Answer:
A. An overview learning rate leads to model divergence
B. The discriminator is too well trained at the beginning of training. As a result, the generator
cannot be trained.
C. The mode collapse results in insufficiently diversified images.
D. The number of data sets is small, causing model overfitting
Which of the following are the major modules of the Ascend AI processor?
Answer:
A. Chip system control CPU
B. AI computing engine
C. Multi-level System-On-a-Chip cache or buffer
D. Digital vision pre-processing module
Which of the following options are Huawei products?
Answer:
A. Atlas 200DK
B. Ascend 310
C. Ascend 910
Which of the following are AI applications?
Answer:
A. Smart speaker
B. Smart customer service
C. Autonomous driving
D. Alpha Go
Which of the following statements about TensorFlow are correct?
Answer:
A. Keras is a high-level API of TensorFlow
B. TensorFlow 1.0 uses static graphs
C. TensorFlow 2.0 uses Eager Execution by default
Compared with the Sigmoid activation function, which of the following options are the advantages of
ReLU?
Answer:
A. Mitigating the problem of vanishing gradient
D. Save calculation
Which of the following are deep learning frameworks?
Answer:
A. TensorFlow
B. PyTorch
C. MXNet
D. MindSpore
ken radh3touh adi
KIJG T3WCNKV[$GVVG T5G TXKEG

KillTest
Q&A
=KULLKXLXKK[VJGZKYKX\OIKLUXUTK_KGX
83
NZZV ]]]QORRZKYZIUS
The safer , easier way to help you pass any IT exams.
1 / 29
Exam : H13-311
Title :
Version : V8.02
Huawei Certified Network
Associate-Artficial
Intelligence
2 / 29
1. What is the subject of artificial intelligence?

A. Mathematics and Physiology
B. Psychology and Physiology
C. Linguistics
D. Comprehensive interdisciplinary and marginal disciplines
Answer: D
2. Functions are well-organized, non-reusable code segments used to implement a single, or associated
function.
A. True
B. False
Answer: B
3. What is wrong description of backpropagation?

A. The learning process of the backpropagation algorithm consists of a forward propagation process and
a back-propagation process.
B. The backpropagation algorithm is a learning algorithm suitable for multi-layer neural networks, which is
based on the gradient descent method.
C. The backpropagation phase sends training inputs to the network to obtain an stimuli response.
D. The backpropagation algorithm is mainly repeated by two loops (excitation propagation, weight update)
until the response of the network to the input reaches the predetermined target range.
Answer: C
4. In May 1997, the famous "Human-Machine Wars" final computer defeated Kasparov, the world chess
king, with a total score of 3.5 to 2.5. Is this computer called?
A. Dark blue
B. Dark green
C. Ponder
D. Blue sky
Answer: A
5. What are the service categories included in Huawei Cloud EI Enterprise Intelligence? (Multiple Choice)
A. EI visual cognition
B. EI online games
C. EI speech semantics
D. EI industry scene
Answer: ACD
A. Beautiful
B. Expensive
C. Explicit
84
D. Simple
Answer: ACD
3 / 29
7. The history of robots is not long. In 1959, the United States, Engelberg and Devol made the world's first
generation of industrial robots, and the history of robots really began.
According to the development process of the robot, it is usually divided into three generations,
respectively are: (Multiple Choice)
A. Teaching Reproduction Robot
B. Robot with sensation
C. Robots that will think
D. Intelligent robot
Answer: ABD
8. There are a lot of data generated during the training of the neural network. What mechanism does
A. Client
B. feed
C. placeholder
D. fetch
Answer: C
9. Python tuples are identified by "()" and internal elements are separated by ":".
A. True
B. False
Answer: B

A. int (integer type)
B. long (long integer type)
C. float (floating point type)
D. complex (complex number type)
Answer: ABCD
11. What is the English abbreviation for AI?

A. Automatic Intelligence
B. Artifical Intelligence
C. Automatic Information
D. Artifical Information
Answer: B
12. There are many commercial applications for machine learning services. What are the main business
A. Financial product recommendation
B. Predictive maintenance
C. Telecom customer retention
D. Retailer grouping
4 / 29
Answer: ABCD

A. The Python module is a Python file that ends with .py and contains Python object definitions and
Python statements.
85
B. The Python module allows you to logically organize your Python code snippets.
C. Python modules can define functions, classes, and variables, but the module does not contain
executable code.
D. Assigning relevant code to a module can make your code better and easier to understand.
Answer: C
14. Which of the following options is not the session mode used by TensorFlow?
A. Explicitly call the session to generate function
B. Explicitly call the session to close function
C. Through the Python context manager
D. Multiple POST queries
Answer: D
15. Principal Component Analysis (PCA) is a statistical method. A set of variables that may be related to
each other is transformed into a set of linearly related variables by orthogonal transformation. The
converted set of variables is called the principal component.
A. True
B. False
Answer: B
A. VGGNet
B. ResNet
C. AlexNet
D. GoogleNet
Answer: ABCD

A. In natural phenomena and social phenomena, many random variables obey or approximate a normal
distribution.
B. The normal distribution takes the maximum value at the mean.
C. The larger the standard deviation, the steeper the normal distribution curve.
D. The larger the standard deviation, the slower the normal distribution curve.
Answer: C
18. Which of the following options is not a reason for traditional machine learning algorithms to promote
the development of deep learning?
A. Dimensional disaster
B. Local invariance and smooth regularization
5 / 29
C. Manifold learning
D. Feature Engineering
Answer: D
19. The commonly used functions for mathematical operations in Python are basically in the math module
and the cmath module.
A. True
B. False
Answer: A
20. What does not belong to supervised learning?

A. Logistic regression
B. Support vector machine
C. Decision tree
86
D. Principal component analysis
Answer: D
21. The Python list can be identified by " [] ", and the default index of the first element from left to right
is 1.
A. True
B. False
Answer: B
22. What are the core concepts in TensorFlow? (Multiple Choice)

A. Gridding
B. Calculation Chart
C. Dot Product
D. Tensor
Answer: BD
A. Machine translation
B. Speech recognition
C. Image style migration
D. Text generation
Answer: C
24. Numerical calculation refers to the method and process of effectively using the digital computer to
solve the approximate problem of mathematical problems, and the discipline consisting of related
theories.
Which of the following processes are involved in solving actual problems with a computer? (Multiple
Choice)
A. Know the actual problem
B. Mathematical model
6 / 29
C. Numerical calculation method
D. Programming
E. Computer calculation results
Answer: ABCDE
25. Which of the following is not the difference between Python 2 and Python 3?
A. print
B. Unicode
C. import
D. xrange
Answer: C
26. Deep learning algorithms can be divided into supervised learning and unsupervised learning.
A. True
B. False
Answer: A
27. The matrix produced by the exchange of matrix A rows and columns is called the transpose
of A. What are the correct properties of matrix transposition? (Multiple Choice)
A. (A T) T = A
B. (A + B) T = A T +B T
C. (λA) T = λA T
87
D. (AB) T = A T +B T
Answer: ABC
28. What quotation marks can the Python language use? (Multiple Choice)
A. Single quotes
B. Double quotes
C. Three quotes
D. Four quotes
Answer: ABC
29. Which of the following data types does TensorFlow not support?
A. int8
B. float32
C. double64
D. bool
Answer: C
30. In TensorFlow, data is represented in the form of tensors and calculation charts.
A. True
B. False
Answer: B
7 / 29
31. Python is a fully object-oriented language. Which of the following options belong to
the Python object? (Multiple Choice)
A. Function
B. Module
C. Number
D. Character string
Answer: ABCD
32. Which of the following options belong to TensorFlow concept? (Multiple Choice)
A. Tensor
B. Variables
C. Placeholder
D. Operation
E. Conversation
Answer: ABCDE
33. The following are the description of the distribution function, distribution law, and density function of
random variables, which is incorrect?
A. Distribution law can only describe the value rule of discrete random variables
B. The density function can only describe the value rule of continuous random variables
C. Discrete random variables have no distribution function
D. Distribution function describes the value rule of random variables
Answer: C
34. Which of the following conditions do the randomized trials need to meet? (Multiple Choice)
A. It can be repeated under the same conditions.
B. There may be more than one possible outcome for each trial, and all possible outcomes of the trial can
be clarified in advance.
C. All results of the test cannot be clarified in advance.
D. It is not possible to determine which result will appear before conducting a test.
Answer: ABD
88
35. In a neural network, knowing the weight and deviations of each neuron is the most important step. If
you know the exact weights and deviations of neurons in some way, you can approximate any function.
What is the best way to achieve this?
A. Random assignment, pray that they are correct
B. Search for a combination of weight and deviation until the best value is obtained
C. Assign an initial value to iteratively update weight by checking the difference between the
best value and the initial.
D. The above is not correct
Answer: C
36. What are the regularizations in deep learning? (Multiple Choice)

A. L1 norm, L2 norm
8 / 29
B. Data set enhancement
C. Integration method
D. Dropout
Answer: ABCD
37. Which of the following about the dictionary in Python is correct? (Multiple Choice)
A. Each key and its corresponding value need to be separated by ":"
B. Separate the different key-value with ","
C. The entire dictionary is included in the " { } "
D. The keys of the dictionary are unique and the data type is uniform
Answer: ABCD
38. The determinant of square matrix A is a scalar.

A. True
B. False
Answer: A
39. Which of the following about the gradient descent is incorrect?

A. Random gradient descent is a commonly used one in gradient descent.
B. Gradient descent includes random gradient descent and batch gradient descent.
C. The gradient descent algorithm is fast and reliable.
D. Random gradient descent is one of the commonly used optimization algorithms in deep learning
algorithms.
Answer: C
40. Which of the following conditions is not a condition that n-fold bernoulli trials needs to meet?
A. Each test was repeated under the same conditions.
B. There are only two possible outcomes for each trial, i.e. event A occurs and event A does not occur.
C. Each test itself is obeying normal distribution.
D. The results of each trial are independent of each other.
Answer: C
41. The Python language can use multiple statements on the same line, separated by commas ",".
A. True
B. False
Answer: B
42. Which of the following files is the graph structure of TensorFlow saved in?
A. checkpoint file
B. index file
C. .cpkt file
89
D. .meta file
Answer: D
9 / 29
43. Machine Learning Service is a data mining and analytics platform service that helps users quickly
discover data patterns and build predictive models through machine learning techniques and deploy them
as predictive analytics solutions.
A. True
B. False
Answer: A
44. Which of the following about the description of expectations and variances is incorrect?
A. Expectation reflects the average level of random variable values
B. The variance reflects the degree of deviation between the random variable and its mathematical
expectation
C. Expectation and variance are both numerical characteristics of random variables
D. The greater the expectation, the smaller the variance
Answer: D
45. The Python dictionary is widely identified by " { } ", and the internal data consists of the key and its
A. True
B. False
Answer: A
46. Which of the following options is not a tensor attribute?

A. name (Logo)
B. shape (Dimensionality)
C. type (Type)
D. node (Node)
Answer: D
47. Which of the following products is related to artificial intelligence? (Multiple Choice)
A. AlphaGo
B. Self-driving
C. Voice input
D. Huawei Mate mobile phone
Answer: ABCD
48. Python regular expressions are a special sequence of characters that makes it easy to check if a
string matches a pattern.
A. True
B. False
Answer: A
49. What operating systems does Python support? (Multiple Choice)

A. DOS
B. Linux
10 / 29
C. Mac OSX
D. Windows
Answer: ABCD
90
50. Which of the following about the description of the number of rows and columns for the determinant is
correct?
A. The number of rows is greater than the number of columns
B. The number of rows is equal to the number of columns
C. The number of rows is less than the number of columns
D. The number of rows has no relationship with the number of columns.
Answer: B
51. Deep learning makes it easy to derive simple mathematical functions from a large amount of
high-dimensional data to describe complex decision interfaces.
A. True
B. False
Answer: B
52. Which of the following does not belong to automatic hyperparameter optimization algorithm?
A. Grid search
B. Random gradient descent
C. Random search
D. Model-based hyperparameter optimization
Answer: B

A. legibility
B. Simplicity
C. Rapidity
D. Scalability
Answer: ABD
A. sigmoid
B. tanh
C. relu
D. danish
Answer: ABC
55. Which of the following statements about the positive definiteness of the matrix is correct? (Multiple
Choice)
A. A matrix in which all eigenvalues are positive is called positive definite.
B. A matrix in which all eigenvalues are non-negative are called semi-positive definite.
C. A matrix in which all eigenvalues are negative is called negative definite.
11 / 29
D. A matrix in which all eigenvalues are non-positive are called semi-negative definite.
Answer: ABCD
56. What is the incorrect relationship between neural networks and deep learning?
A. The concept of deep learning stems from the study of artificial neural networks.
B. A neural network algorithm with multiple hidden layers is a deep learning algorithm.
C. Single layer neural networks are also a type of deep learning.
D. Convolutional neural networks are a type of deep learning.
Answer: C
57. Which of the following environments does not support to install the TensorFlow?
A. Linux
B. Mac OS
91
C. Docker
D. OpenStack
Answer: D
58. Deep learning is a branch of machine learning.

A. True
B. False
Answer: A
59. The trace operation returns the sum of the diagonal elements of the matrix. Therefore, the trace
of matrix A and its transposed matrix are equal.
A. True
B. False
Answer: A
60. GAN is a deep learning model and one of the most promising methods for unsupervised learning in
complex distribution in recent years.
A. True
B. False
Answer: A

A. Probability theory
B. Linear transformation
C. Matrix theory
D. Vector space
Answer: A
62. A scalar k is multiplied by matrix A equal to k and each of the numbers in A is multiplied.
A. True
B. False
12 / 29
Answer: A
63. What are the conditions for m row n column matrix A and p row q column matrix B to be multiplied?
A. m=p, n=q
B. n=p
C. m=n
D. p=q
Answer: B
64. Which of the following are the characteristics of TensorFlow? (Multiple choice)
A. Open source
B. Convenience
C. Mature
D. Flexible
Answer: ABCD
65. Which is the correct description of the Python creation function? (Multiple choice)
A. The function created starts with the def keyword followed by the function name and parentheses.
B. The parameters need to be placed in parentheses.
C. The function content starts with a colon and needs to be indented.
D. Return the result with return and the function ends.
Answer: ABCD
92
66. About Python correct tuple, which is correct description? (Multiple choice)
A. The difference between a list and a tuple is that the elements of the tuple cannot be modified.
B. The list uses square brackets, and the tuple uses parentheses.
C. Tuple creation is very simple, just add elements in parentheses and separate them with commas.
D. When a tuple contains only one element, you need to add a comma after the element. 。
Answer: ABCD
67. What research areas does artificial intelligence have? (Multiple choice)
B. Computer Vision
C. Machine learning
D. Speech recognition
Answer: ABCD
68. TensorFlow supports multi-TPU cluster computing.

A. True
B. False
Answer: A
69. As shown below, which are the following matrix A characteristic value? (Multiple choice)
13 / 29
A. 2
B. -2
C. -4
D. 4
Answer: AD
70. Only matrix A and matrix B have the same number of rows and columns, A and B can be added.
A. True
B. False
Answer: A
71. The continue statement in the Python language is used to jump out of this loop and then continue to
A. True
B. False
Answer: A
72. Which of the following are the cloud services provided by Huawei Cloud EI Visual Cognition? (Multiple
choice)
A. Text recognition
B. Face recognition
C. Image recognition
D. Content detection
E. Image processing
Answer: ABCDE
73. X, Y are random variables, C is a constant, the nature of the difference in the following options, which
is wrong? (Multiple choice)
A. D(C)=0
B. D(X+Y)=D(X)+D(Y)
C. D(CX)=CCD(X)
D. D(XY)=D(X)D(Y)
Answer: BD
93
74. Vector is a number.
A. True
B. False
Answer: B
75. About Bayesian formula: P(W|X)=P(X|W)*P(W)/P(X) What is the correct description?

14 / 29
A. P(W|X) is a prior probability
B. P(X|W) is a conditional probability
C. P(W) is the posterior probability
D. P(X) is the posterior probability
Answer: B
76. Cyclic neural networks are more suitable for dealing with image recognition problems.
A. True
B. False
Answer: B
77. TensorFlow Operations and Computation Graph are not - run in the Session.
A. True
B. False
Answer: B
78. What is in the category of artificial intelligence? (Multiple choice)

A. Action
B. Perception
C. Cognition
D. None of the above
Answer: ABC
79. The Python language does not allow embedding another loop body in a loop body.
A. True
B. False
Answer: B
80. What are the mainstream deep learning open source tools? (Multiple choice)
A. TensorFlow
B. Caffe
C. Torch
D. Theano
E. Scikit-learn
Answer: ABCD
81. The for loop statement in the Python language can iterate through the items in any sequence.
A. True
B. False
Answer: A
82. Convolutional neural networks are more suitable for image recognition problems than cyclic neural
networks.
A. True
15 / 29
B. False
Answer: A
94
83. Huawei Cloud EI builds enterprise intelligence services based on three-tier services. Which of the
following options does not belong to Layer 3 services?
A. Basic platform services
B. General domain services
C. Industry sector services
D. Integration services
Answer: D
84. Which of the following is true about unsupervised learning?

A. Unsupervised algorithm only processes "features" and does not process "tags"
B. Dimensionality reduction algorithm is not unsupervised learning
C. K-means algorithm and SVM algorithm belong to unsupervised learning
Answer: A
85. Which of the following options is not the TensorFlow build process?
A. Building a calculation chart
B. Input tensor
C. Generate a session
D. Update weights
Answer: D
86. What model is not a cyclic neural network?

A. RNN
B. LSTM
C. GBDT
D. GRU
Answer: C
87. Under the large-scale relationship analysis scenario, which of the following options does not belong to
the three high demands of massive relationship processing?
A. Efficient relationship discovery data for massive data
B. Efficient sharing of massive data
C. Efficient storage and access requirements for massive amounts of data
D. High scalability and high availability requirements for relational analysis platforms
Answer: B
88. What are the commonly used loss functions? (Multiple choice)
A. Mean variance
B. Sigmoid cross entropy
C. Softmax cross entropy
16 / 29
D. Sparse cross entropy
E. Weighted Sigmoid Cross Entropy
Answer: ABCDE
89. What are the steps that are not part of the operation of a Python file object?
A. open
B. delete
C. read
D. write
E. close
Answer: B
95
90. TensorFlow is the second generation of artificial intelligence learning system developed by Google
based on ( ).
A. DistBelief
B. PaleyFunction
C. ConvexOne
D. Infinity
Answer: A
91. The Python language can use the "#" at the beginning of a single line of code for code comments.
A. True
B. False
Answer: A
92. According to the definition of information entropy, what is the bit entropy of throwing a uniform coin?
A. 0
B. 0.5
C. 1
D. -1
Answer: C
93. Artificial intelligence is currently suitable for scenarios with “known environment, clear objectives, and
predictable actions”. Deep learning in the fields of image recognition, speech recognition, translation, etc.,
artificial intelligence basically has human recognition ability, even surpassing human beings. Based on
these capabilities, it has been applied to many scenes, such as medical care and public safety. However,
there is still a lack of reasoning and cognition.
A. True
B. False
Answer: A
94. Python authors deliberately design very restrictive grammars that make bad programming habits
(such as the next line of if statement not indented to the right) cannot be compiled.
A. True
17 / 29
B. False
Answer: A
95. The timestamps in the Python language are represented by how long (in seconds) elapsed from
midnight (epoch) on January 1, 1970.
A. True
B. False
Answer: A
96. Convolutional neural networks are more suitable for dealing with speech recognition problems.
A. True
B. False
Answer: B
97. What are the results returned by the if conditional statements in the Python language? (Multiple
choice)
A. TRUE
B. 0
C. FALSE
D. null
Answer: AC
96
A. The verification set can coincide with the test set.
B. The test set can coincide with the training set.
C. The subset used to pick hyperparameters is called a validation set.
D. Typically 80% of the training data is used for training and 20% is used for verification.
Answer: A
99. Which of the following can optimization problem be classified according to the constraints? (Multiple
choice)
A. equality constraints
B. Inequality constraints
C. Unconstrained conditions
D. Semi-constraint conditions
Answer: ABC
00. Which of the following are the characteristics of the Python language? (Multiple choice)
A. Explanatory
B. Process oriented
C. Object-oriented
D. Dynamic data type
Answer: ACD
18 / 29
01. Which of the following are the elements of artificial intelligence? (Multiple choice)
A. Algorithm
B. Computing power
C. Data
D. Scene
Answer: ABCD
02. Which of the following is not a specific technology of artificial intelligence?

A. Knowledge map
C. Riemann geometry
D. Semantic understanding
Answer: C
03. Which of the following statements about overflow and underflow in numerical calculations is correct?
(Multiple choice)
A. Underflow occurs when the number close to zero is rounded to zero
B. Overflow occurs when the number close to zero is rounded to zero
C. Overflow occurs when a large number of levels are approximated as ∞ or -∞
D. Underflow occurs when a large number of levels are approximated as ∞ or -∞
Answer: AC
04. Which of the following description about Python regular expression is not correct?
A. Python has added the re regular expression module since version 1.5.
B. re regular expression module gives the Python language all the regular expression functionality.
C. re regular expressions are powerful tools for working with strings.
D. re regular expressions can only process string data, not numeric data.
Answer: D
05. In which year was the first introduction of "artificial intelligence"?

A. 1946
97
B. 1960
C. 1916
D. 1956
Answer: D
06. Which of the following description about the relationship between the Poisson distribution and the
binomial distribution is correct?
A. The mathematical model of the Poisson distribution and the binomial distribution are both Lagrangian
models.
B. Poisson distribution is an approximation of the binomial distribution when n is very large and p is small.
C. Poisson distribution has no relationship with binomial distribution
D. Poisson distribution can replace binomial distribution
Answer: B
19 / 29
07. Which of the following is not an artificial intelligence school?

A. Symbolism
B. Statisticalism
C. Behaviorism
D. Connectionism
Answer: B
08. Which of the following does the convolutional neural network structure not include?
A. Convolutional layer
B. Pooling layer
C. Loop layer
D. Fully connected layer
Answer: C
09. Which of the following functions can numerically stabilize overflow and underflow?
A. Softminus function
B. Softplus function
C. Softmax function
D. Softmin function
Answer: C
10. What are the advantages of Python? (Multiple choice)

A. Simple
B. Free
C. High-level language
D. Rich library
Answer: ABCD
111. Python script execution mode includes interactive mode and script mode.
A. True
B. False
Answer: A
12. Which of the following features does Python support? (Multiple choice)
A. Inheritance
B. Heavy load
C. Derived
D. Multiple inheritance
Answer: ABCD
98
13. Which of the following description of the number of rows and columns of the matrix is correct?
A. The number of rows is greater than the number of columns
B. The number of rows is equal to the number of columns
20 / 29
C. The number of rows is less than the number of columns
D. The number of rows has no relationship with the number of columns.
Answer: D
14. Which of the following is the computing model of TensorFlow?

A. Calculation chart
B. Tensor
C. Conversation
D. Variable
Answer: A
15. Which of the following can improve the computational efficiency of the neural network model?
(Multiple choice)
A. TPU
B. GPU
C. FPGA
D. Large-scale distributed cluster
Answer: ABCD
16. Which of the following statements about Python are correct? (Multiple choice)
A. Invented in 1989 by the Dutch Guido van Rossum, the first public release was issued in 1991.
B. Python is purely free software and the source code follows the GPL (GNU General Public License)
protocol.
C. Python syntax is simple and clean. One of the features is to force blank characters to be used as
statement indentation.
D. Python is often nicknamed glue language, which can easily connect various modules made in other
languages.
Answer: ABCD
17. Which of the following does not belong to long-term memory LSTM (Long Short-Term Memory)
network architecture?
A. Memory door
B. Forget the door
C. Input door
D. Output door
Answer: A
118.In order for a machine to be intelligent, it must be knowledgeable. Therefore, there is a research field
in artificial intelligence, which mainly studies how computers automatically acquire knowledge and skills to
achieve self-improvement. What is the branch of this research called?
A. Expert system
B. Machine learning
C. Neural Network
D. Natural language processing
21 / 29
Answer: B
18. Which of the following descriptions about Python program are correct? (Multiple choice)
A. The program consists of modules
99
B. Module contains statements
C. statement contains an expression
D. Expressions create and process objects
Answer: ABCD
19. Huawei Machine Learning Service MLS is a one-stop platform that supports the entire process of
data analysis. Which of the following is not a feature of MLS?
A. A rich library of machine learning algorithms.
B. machine learning program is intuitive and easy to use.
C. Distributed and scalable big data computing engine.
D. Support for the R language, but does not support the Python language.
Answer: D
20. Training error will reduce the accuracy of the model and produce under-fitting. How to improve the
A. Increase the amount of data
B. Feature Engineering
C. Reduce regularization parameters
D. Add features
Answer: ABD
21. Which of the following is not a way for the TensorFlow program to read data?
A. Preload data
B. Feeding data
C. Read from the file
D. Write a file format reader
Answer: D
22. Self-encoder is an unsupervised learning algorithm.

A. True
B. False
Answer: A
23. The number or matrix is stored in the tensor.

A. True
B. False
Answer: B
24. Which of the following descriptions about the Recurrent Neural Network (RNN) is correct?
A. Can be used to process sequence data.
22 / 29
B. Cannot process variable length sequence data.
C. Unlike convolutional neural networks, parameters of RNN cannot be shared.
D. The units above the hidden layer are not associated with each other.
Answer: A
25. Artificial intelligence is a new technical science that studies and develops theories, methods and
application systems for simulating, extending and extending human intelligence. It is one of the core
research areas of machine learning.
A. True
B. False
Answer: B
26. The TensorFlow framework does not support Windows systems.

A. True
1) ken radh3touh its ok u have bilgacem for further assistance ,if you dont support bilgacem timchi tnayek enti ou 100
B. False
Answer: B
27. Which of the following descriptions of the depth feedforward network is correct?
A. Deep feedforward network is a kind of neural network.
B. The deep feedforward network has only one hidden layer.
C. The unit on the hidden layer of the deep feedforward network will have countless.
D. Deep feedforward network is used to deal with linear problems.
Answer: A
28. The time interval in the Python language is a floating-point fraction in seconds.
A. True
B. False
Answer: A
29. Which of the following is not part of the deep learning development framework?
A. CNTK
B. Keras
C. BAFA
D. MXNet
Answer: C
30. Which of the following are included in the application scenarios of TensorFlow? (Multiple choice)
B. Face recognition
C. Image style changes
D. Autopilot
Answer: ABCD
31. Which of the following neural network structures will share weights? (Multiple choice)
23 / 29
A. Convolutional neural network
B. Recurrent neural network
C. Fully connected neural network
D. All of the above
Answer: AB
32. The correlation coefficient, also known as the linear correlation coefficient, is used to measure the
linear relationship between two variables, which is a real number greater than zero.
A. True
B. False
Answer: B
33. X, Y are random variables, C is a constant, which of the following description about the nature of the
mathematical expectation is incorrect?
A. E(C) = CA. E(C) = C
B. E(X+Y) = E(X)+E(Y)
C. E(CX) = CE(X)
D. E(XY) = E(X)E(Y)
Answer: D
34. Which of the following statement about the matrix is incorrect?

A. The arbitrary matrix is multiplied by the unit matrix and will not change.
B. The transpose of the symmetric matrix A is also A itself.
C. The transpose of the orthogonal matrix A is equal to the inverse of A.
D. There is no inverse matrix for the diagonal matrix.
Answer: D
35. Which is not a deep learning algorithm?

A. Self-encoder
B. Convolutional neural networks
C. Recurrent neural networks
D. Support vector machine
Answer: D
36. TensorFlow is Google's first generation of proprietary machine learning systems.

A. True
B. False
Answer: B
37. Which of the following schools does the neural network study belong to?
A. Symbolism
B. Connectionism
C. Behaviorism
24 / 29
Answer: B
38. Reducing the gap between the training error and the test error will result in over-fitting. How to
prevent over-fitting? (Multiple Choice)
A. Cross validation
B. Integration method
C. Increase regularization
D. Feature Engineering
Answer: ABCD
39. Which is wrong with the Python module time introduction method?
A. import time
B. from time import *
C. import time as t
D. from time
Answer: D
40. In what year did Huawei officially provide services in the form of cloud services, and combined with
more partners to provide richer artificial intelligence practices?
A. 2002
B. 2013
C. 2015
D. 2017
Answer: D
41. Which description is wrong about the lambda function of the Python language?
A. lambda is just an expression, and the function body is much simpler than def.
B. The body of a lambda can be an expression or a block of code.
C. The lambda function can access the parameters in the global namespace.
D. The lambda function accepts only one parameter value.
Answer: D
42. The meaning of artificial intelligence was first proposed by a scientist in 1950, and at the same time a
test model of machine intelligence was proposed. Who is this scientist?
A. Minsky
B. Zade
C. Turing
D. Von Neumann
Answer: C
43. The number of rows and columns of the three matrices A, B, and C is 3 rows and 2 columns, 2 rows,
3 columns, 3 rows and 3 columns, which of the following operations is meaningful?
A. AC
B. BC
25 / 29
C. A+B
D. AB-BC
Answer: B
44. The constituent elements of the Python identifier include: numbers, letters, and underscores.
A. True
B. False
Answer: A
45. Which description is wrong about the hyperparameter?

A. Hyperparameters are parameters that set values before the algorithm begins learning.
B. Most machine learning algorithms have hyperparameters.
C. Hyperparameters cannot be modified.
D. The value of the hyperparameter is not learned by the algorithm itself.
Answer: C
46. Which of the following is not a module in the TensorFlow library?

A. tf.nn
B. tf. layers
C. tf. boost
D. tf.contrib
Answer: C
47. As shown in the figure below, what is the value of the determinant A?
A. 24
B. 18
C. -24
D. 0
Answer: D
48. How many spaces does PEP 8 stipulate that Python's first line indentation needs to indent?
A. 1
B. 2
C. 4
D. 8
Answer: C
A. numbers
B. string
26 / 29
C. list
D. tuple
E. dictionary
Answer: ABCDE
50. What are the algorithms supported by TensorFlow? (Multiple Choice)

A. CNN
B. ZNN
C. RNN
D. HUT
Answer: AC
51. Which is not a Python operator?

A. arithmetic operator
B. Reasoning operator
C. Logical operator
D. Comparison operator
Answer: B
52. Which method is not supported in TensorFlow to define variables?

A. Random number
B. Constant
C. Calculated from the initial values of other variables
D. null
Answer: D
53. What does factors that promote the development of artificial intelligence not include?
A. Big data
B. Computing ability
C. Algorithm theory
D. Blockchain
Answer: D
54. Which of the following are solutions for the Huawei Cloud EI industry scenario? (Multiple Choice)
A. Intelligent Logistics
B. Intelligent Water
C. Intelligent transportation
D. Intelligent Finance
E. Intelligent manufacturing
Answer: ABCDE
55. Vector group a1=(1,1,1), a2=(0,2,5), a3=(1,3,6), which of the following options is correct?
A. Linear correlation
B. Linear independence
27 / 29
C. a1+a2+a3=0
D. 2a1+a2+a3=0
Answer: A
56. What are the scenarios or industries that are suitable for using Python? (Multiple Choice)
A. Artificial intelligence
B. web development
C. Game development
D. Hardware development
Answer: ABC
57. Which descriptions are correct about Python's index? (Multiple Choice)
A. Index from left to right defaults from 0
B. Index from left to right defaults from 1
C. Index from right to left defaults from -1
D. Index from right to left defaults from 0
Answer: AC
58. Huawei firmly believes that the value of AI is ultimately reflected in solving the actual business
problems of the enterprise. Therefore, Huawei is firmly committed to the development of AI technology,
focusing on domain solutions, focusing on basic technologies and enabling platforms, and working with
industry practice leaders to develop industry solutions.
A. True
B. False
Answer: A
59. The Python dictionary is identified by "{}", and the internal data consists of the key and its
A. True
B. False
Answer: A
60. In Huawei Cloud EI Enterprise Intelligence, which basic platform services are included? (Multiple
Choice)
A. Machine learning
B. Deep learning
C. Graph engine
D. Batch processing
Answer: ABCD
61. Linear algebra is a branch of algebra that deals with linear problems.
The linear relationship is that the relationship between exponential objects is expressed in ( ) times
form?
A. 1
28 / 29
B. 2
C. 3
D. 4
Answer: A
62. "print" in Python 3 must be used with "()".

A. True
B. False
Answer: A
63. Deep learning is different from machine learning and there are no unsupervised algorithms.
A. True
B. False
Answer: B
64. What are the implementation modes of TensorFlow? (Multiple Choice)

A. Stand-alone mode
B. Distributed mode
C. Reverse mode
D. Forward mode
Answer: AB
65. What are the application scenarios for the break statement in the Python language? (Multiple
Choice)
A. Any Python statement
B. while loop statement
C. for loop statement
D. Nested loop statement
Answer: BCD
66. TensorFlow only supports Python and R language

A. True
B. False
Answer: B
67. What are the commonly used gradient descent optimization functions? (Multiple Choice)
A. Random gradient descent
B. Adadelta
C. Adagrad
D. momentum
E. RMSProp
Answer: ABCDE
68. What is not the optimization method in deep learning?

29 / 29
A. Random gradient descent
B. Back propagation algorithm
C. Principal component analysis
D. Momentum
Answer: C
69. Information theory is a branch of applied mathematics. The main research is to quantify how much
A. Minsky
B. Turing
C. Simon
D. Shannon
Answer: D
70. Which of the following description is wrong about the image recognition cloud service?
A. The service identifies specific objects such as skyscrapers, buildings, and ships.
B. The service can also identify the sea, city, port and other scenes.
C. The service can perform semantic analysis on the content in the image to extract labels with specific
abstract meanings.
D. The service can provide object and scene labels, and can also provide more conceptual labels.
Answer: D
71. The tree model used at the bottom layer of the random forest and GBDT algorithms is only
the classification and registration tree (CART)
False
72. Distributed training is a feature of the TensorFlow framework.

True
73. During model training, a larger value of the epoch parameter indicates a better training
result.
False
74. Mini-batch gradient descent (MGBD) cannot balance the robustness of stochastic gradient
descent and the efficiency of batch gradient descent, and we may risk getting stuck at local
minima. It is not commonly used in practice.
False
75. MindSpore accelerates model convergence through automatic parallelism and second-order
optimization.
True
76. Air eliminates model differences between different backends through unified operator IR
definitions. You can coordinate tasks on all platforms (device, edge, and cloud) based on the
same model file.
True
77. Assuming the number of hyperparameters is the same, stochastic gradient descent (SGD)
combined with manual adjustment achieves a better effect than adaptative learning rates.
False
78. An AI processor is also referred to as an AI accelerator. It is a functional module used to

process a large quantity of computing tasks in an AI application. Ascend 910 is typical
product.
True
79. Federated learning can protect user privacy to some extent.

True
80. In the convolutional layer, the dropout ration is the ratio of features whose values are set to
81. Generally, the dropout ratio ranges from 0.2 to 0.5.

True
82. As the compute core of an Ascend AI Processor, the AI Core is responsible for complex
matrix computation.
True
83. A decision tree selects a label from features of the trining data to work as the node splitting
standard. Different label selection standards generate different decision tree algorithms.
False
84. In a convolutional neural network, the size of each kernel is not necessarily the same as that
of a pooling layer window, but their steps should be the same.
False
85. The ModelArts inference platform implements fast and efficient model inference and image
recognition.
True
86. A model constructed using machine learning algorithms can be a function that conforms to
actual data distribution.
False
87. Convolutional neural networks cannot be used to process text data.

False
88. If a 32 x 32 image (without padding) is input and a 5 x 5 kernel is used for convolution with a
step of 2, the size of the output image is 13 x 13.
False
89. One of the filter methods is using the Chi-Squared test to analyze the mapping between
features and labels to implement feature selection.
True
90. The softplus function has a continuous derivative and defines a smooth curved surface.
True
91. MindSpore supports Ascend processors and CPUs, but does not support GPUs.
False
92. Which of the following statements is true about rule-based approaches?

Rule-based approaches use explicit programming to solve problems.
93. In TensorFlow 2.0, x = tf.constant([1,2,3])y = tfibroadcast_to(x,[3,3])print(y)

Which of the following is the output for this code?
B. [[1,2,3], [1,2,3], [1,2,3]]
94. How does Al perform in the cognitive intelligence phase?

C. Machines begin to understand, think, and make decisions like humans.
95. Which of the following indicators cannot be used to evaluate a model?

d. Code complexity
96. concat_sample_1 = tfrandom.normal([4,100,100,3])concat_sample_2 =

tfrandom.normal([40,100,100,3))concated_sample_1 =
tf.concat([concat_sample_1,concat_sample_2],axis=0)print(concated_sample_1.s
hape) Which of the following is the output for this code?
B. (44,100,100,3)
97. An image recognition experiment uses 42,510 training images. There are more than 200
categories, each of which contains only 10 images. There are also categories with only 20 to
50 images. Which of the following data problems does this phenomenon fit into?
B. Data imbalance
98. "Knowledge representation is the unique method of representing knowledge using a

set of symbols in a structure that can be understood by computers." Which of the
following is true about this statement?
A. This statement is false. The method of representing knowledge is not unique.
99. Which of the following statements is true about optimizers?

c. The SGD and momentum optimizers use the same learning rate for each iteration.
00. Which of the following is the basic unit of data storage in PyTorch, MindSpore, and
TensorFlow?
b. tensor
01. Which of the following determines the upper limit of machine learning?
A Data
02. Which of the following HUAWEI CLOUD services is NOT a basic Al platform service?
C. Relational Database Service (RDS)
03. Which of the following statements is false about the ReLU function?
D. Compared with Sigmoid and tanh, the convergence of the ReLU function is slow.
04. Which of the following is not a hyperparameter for model training?

C. b in the one-dimensional linear regression model Y=wX+b
05. The core of an Al framework and one of the decisive factors of the programming paradigm is
the automatic differentiation used by the Al framework. Which of the following is used by
MindSpore?
C. Source to Source (2S)
06. Which of the following is NOT a main function of HUAWEI CLOUD GeoGenius?
D. Real-time traffic detection
07. Which of the following statements is true about neural networks?

C. LSTMis suitable for processing events with long interval and delay in the time sequence.
08. Which of the following commands is used to install TensorFlow?

D. pip install tensorflow
09. Compared with the random forest algorithm, which of the following is false about the GBDT
algorithm?
A. Training using the GBDT algorithm is faster than that of the random forest algorithm (under the
same dataset and tree quantity).
10. The accuracy of a trained model is high on a training dataset but low on a test dataset. In
this case, which of the following operations can be performed?
B. Perform data augmentation.
C. Add regular terms.
11. Which of the following format conversion operations are not performed by the storage
control unit of the DaVinci Architecture?
B. Video stream compression
C. Color space conversion
12. Which of the following statements are false about GANs?

a. The discriminator input is mainly noise and real sample data.
d. The input of the discriminator is mainly sample data provided by the generator.
13. Which of the following statements are true about gradient descent?
A. Stochastic gradient descent (SGD) randomly chooses samples for each training job, causing
unstability. As a result, the loss function fluctuates or even encounters reverse displacements during
the process of dropping to the minimum value.
C. Mini-batch gradient descent (MBGD) is a balance between BGD and SGD, and is the optimal
choice for alll datasets.
D. BGD uses all samples for each training job.
14. Which of the following are used for backpropagation?

B. Activation functions
D. Neuron weight
15. Which of the following statements are true about tensors?

a.The most basic data structure in TensorFlow is tensor.
b.A first-order tensor can be considered as a vector.
d. A tensor is a generalization based on the vector and matrix.
16. Which of the following statements are false about Keras?

B. Keras is a neural network development package used to build CNN sequential models. It cannot
be used to build other neural networks.
D. Keras can be used as the backend of TensorFlow.
17. Which of the following statements are true about decision trees?
B. Each non-leaf node in a decision tree represents a test on a feature attribute; each branch
represents the output of the feature attribute within a value range; and each leaf node holds a class
label.
D. Building a decision tree means selecting and measuring feature attributes and determining their
topology.
18. Which of the following cells are predefined in MindSpore?

a.MAELoss
d. WithGradCell
19. Which of the following are the advantages of the softsign activation function over the
sigmoid function?
B. Gradient saturation slows down.
C. The output is zero-centered.
20. Which of the following statements are true about common activation functions in deep
learning?
C. The tanh function is symmetric with respect to the origin, and the mean of its output is closer to 0.
D. During training of a deep neural network, the sigmoid, tanh, and softsign functions cannot
prevent the vanishing gradient problem.
21. Which of the following operations are involved when TensorFlow is used for model
training?
A. Define input and output nodes.
B. Train and save models.
D. Define learning parameters and operations.
22. Which of the following statements are true about linear regression?
A. The error of linear regression is affected by many independent factors. According to the central
limit theorem (CLT), the error follows normal distribution.
B. The loss function of linear regression can be obtained using the normal distribution function and
maximum likelihood estimation (MLE).
C. The result of multivariate linear regression analysis may be a hyperplane in a high-dimensional
space.
23. Which of the following are common ensemble learning algorithms in machine learning?
A. Random forest
B. GBDT
D. Adaboost
24. Which of the following statements are true about using OCR to train an invoice recognition
model with the support of a dedicated AI processor?
ABCD
25. Which of the following are model hyperparameters?

A. Learning rate, iteration count, and batch size in a training neural network
B. K in the k-nearest neighbors (KNN) algorithm
C. Number of trees in a random forest
26. The average value of tensors generated using tf.random.normal([4], 2, 1, tf.float32) is 0. (Fill
in the blank with a number.)
(1) 2
27. At a convolution layer, there are 256 5 x 5 convolution kernels, the size of an input feature
map is 32 x 32 x 5, the convolution step is 2, and zero padding is employed(padding size=l).
So, the size of the output feature map is * 0 * 0. (Fill in each blank with a number.)
(1) 15
(2) 15
(3) 256
28. In MindSpore, if you want to define a function for collecting information after each step is
complete, you need to inherit the 0 class.
(1) Callback;callback
29. In TensorFlow 2.0, if tf.keras.layers.RNN is used to process timing information and you want
to obtain the output status at each moment, set () to True.
(1)return_sequence
30. HUAWEI CLOUD 0 helps enterprises build invoice management systems and self-service
reimbursement systems. (Fill in the blank with the abbreviation of the service name.)
(1)OCR;ocr
31. HUAWEI HiAl built an open smart ecosystem based on the " device, chip" three-
layered open architecture. It enables developers to quickly utilize Huawei's powerful
A1 processing capabilities to provide a better smart application experience for users.
(1)Cloud;cloud;CLOUD
1, In a convolutional neural network, the size of each kernel is not necessarily the same as that of a pooling
layer window, but their steps should be the same.
True
False
Mfg, §W7! 4iftWIN: False
2, Mini-batch gradient descent (MBGD) cannot balance the robustness of stochastic gradient descent and the
efficiency of batch gradient descent, and we may risk getting stuck at local minima. It is not commonly used in
practice.
True
False
situ! IIEMM: False
3, Device, edge, cloud can respectively refer to smartphones, Huawei intelligent edge devices, and HUAWEI
CLOUD.
0 True
False Mb'
4 o 7! Edo: True
4, MindSpore accelerates model convergence through automatic parallelism and second-order optimization.
True
False
54M7! IrEinta,;W: True
An Atlas 900 Al cluster is suitable for oil exploration, smart city, and drug discovery scenarios.
0 True
0 False
W, MT! Witntr,M: True
The dropout technology can be used only during Al model training.
False
7, Creative work, such as writing, will not be easily replicated by Al.

0 True
0 Fa Ise
nitg, ttZ7! EifiNM: True
8, Filter methods for feature selection are restricted they do not tend to select redundant variables because they
consider the relationships between features.
True
False NM! lotus: False
9, Assuming a dataset contains the areas and prices of 21,613 housing units in a city, a classification model can
predict the prices of other housing units in the city.
True
False Mg7! 11BilegA: False
10, The tree model used at the bottom layer of the random forest and GBDT algorithms is only the classification
and regression tree (CART).
True
False F.Z17! OPSA: False
11, TensorFlow is Microsoft's second-generation open-source software library built for digital computing.
True
False SPA, ttgr! OfaIM: False
12, MindSpore accelerates the framework of the Generalized Operator Modelling of the Ocean (GOMO) model,
significantly improving performance.
True
False
Ntg7! Imo: True
13, MindSpore can be quickly deployed on the cloud and edge, improving resource utilization and privacy
protection, and helping developers focus on Al apps. Due to the size of the model, MindSpore cannot be
deployed on mobile phones.
True
False
ffia7! EWAN: False
14, Poisson distribution is generally regarded as the common distribution of errors in linear regression.
True
False affirh, kmlEtiffMN: False
Convolutional neural networks cannot be used to process text data.
® True
0 False
Nt47! Eliltag: False
Grid search randomly selects hyperparameter combinations. Therefore, it has a high efficiency and accuracy.
True
False SAM *,R17! mit€PAW: False
17, Although each industry and application faces unique challenges, they are all gradually adopting Al-based
workflows and solutions.
True
False Ign7! Eitrig: True
18, HUAWEI HiAI 3.0 supports the development of P50 smartphones.
LEiri.M: True
19, A decision tree selects a label from features of the training data to work as the node splitting standard.
Different label selection standards generate different decision tree algorithms.
True
False ffia7! EWA*: False
Distributed training is a feature of the TensorFlow framework.
0 True
0 False
MA-7! amm: True
Which one of the following actions can the function tf.squeeze be used for in TensorFlow 2.0?
A. Element-wise addition
B. Computation of the absolute value
C. Tensor concatenation
D. Dimensionality reduction
Mt17! ETIMM: D
22, Which of the following statements is true about classification and regression models in machine learning?
A. For classification problems, the output variables are discrete values with a limited quantity For regression
problems, the output variables are continuous values.
B. The cross-entropy loss function is required for both regression and classification problems.
C. There may be underfitting in both regression and classification problems. Classification models are more
prone to underfitting than regression models.
D. The logistic regression model can be used to predict housing unit prices.
NW! A
23, Dirty data refers to data with quality problems. Which of the following statements is false about the data
quality?
A. Noise: Data contains incorrect records or exceptions.
B. Incompleteness: Data lacks attributes or contains some missing values.
C. Unprocessed: Data for which feature engineering has not been performed.
D. Inconsistency: Data contains conflicting records.
fit47! lorkT: C
24, In TensorFlow 2.0, x = tf.constant([1,2,3])y = tfbroadcast_to(x, [3, 3])print(y)Which of the
following is the output for this code?
A. [[1,2,3]]
B. [[1,2,3], [1,2,3], [1,2,31]
C. [0,2,3,1,2,3.1,2,311
t D. [[1,1,1], [2,2,2],
[3,3.3]]
MI47! EOM*: B
25, Which of the following statements is false about machine learning?
A. Model generalization capability: the extent to which a learned model can be applied to new samples, also
called robustness.
0 B. Error: difference between the prediction of a learned model on a sample and the actual result of the sample.
Errors can be classified into training errors and generalization errors.
C. Underfitting: The training error is large.
D. Overfitting: The training error is large. MA7! D
26, Which of the following statements is true about neural networks?
0 A. The full name of LSTM is Less Short-Term Memory.
CI B. LSTM has nothing to do with RNNs.
CI C. LSTM is suitable for processing events with long interval and delay in the time sequence. D. Neural
networks can only propagate information forward.
MAT! ITEIR§V: C
27, concat_sample_1 = tfrandom.normal([4,100,100,3])&Vbr>concat_sample_2 =

tfrandom.normal([40,100,100,3])&ltibr>concated_sample_1 =
aconcat([concat_sample_tconcat_sample_2],axis=0)&ltdbr>print(concated_sample_l.shape)&lt/br>Which of the
A. (44,200,200,6) B. (44,100,100,3)
C. (40,100,100,6)
D. (44,100,100,6) 4347! ii AR: B
28, Which of the following is the correct shape of tensor [[0,1], [2,3], [4,5]]?
(2,3)
(3,2)
(1,6)
D.(611)
7! EN: B
29, Which of the following is a lightweight and high-performance service module that helps MindSpore
developers efficiently deploy online inference services in production environments?
A. MindIR
B. MindSpore Serving
C. MindArmour
D. MinclInsight
M147! B
30, Which of the following statements is false about decision trees?

A. Except the root node, all nodes in a decision tree are called leaf nodes.
B. The C4.5 algorithm uses the information gain ratio to select feature attributes.
C. The key step to building a decision tree involves splitting it based on feature attributes. ® D. The variance can
be regarded as a quantitative index of purity.
.4A7! iEriltkM: A
31, A vendor wants to provide an intelligent EMR system for a hospital. Which of the following technologies is
involved in the system?
B. Object detection
C. Expert systems and knowledge graph
D. Image generation and enhancement
%MT! MIPAV: A
32, Which of the following is the reasoning method of the production system that draws conclusions through a
rule library?
A. Forward
B. Backward
C. Bidirectional
D. Random
E4R7! IIEEMX: A
33, A generative adversarial network is like a game system. The generator produces fake samples, and the
discriminator tries to distinguish real data from the data created by the generator. What is the ideal result?
A. The generator is trained to fool the discriminator.
B. The discriminator cannot distinguish real data from the data created by the generator.
C. The discriminator distinguishes real data from the data created by the generator.
D. Develop a high-precision discriminator and a generator that cannot fool the discriminator. WRI7! B
34, Which of the following statements is false about the ReLU function?
A. The ReLU function is not differentiable at x = 0 and a derivative is forcibly defined at this point.
B. The ReLU function effectively alleviates the vanishing gradient problem.
C. The surface defined at the zero point of the ReLU function is not smooth enough in some regression
problems.
FriA7! ILEMP: D
35, Which of the following determines the upper limit of machine learning?
A. Data
B. Algorithms
C. Scenarios
D. Computing power
5M-7! IIEMM: A
36, Which of the following is a math operator in MindSpore?
A. Mul
B. ControlDepend
C. Callback
D. Select
Wit17! EiR§V: A
37, Which of the following statements is true about variance and bias?
A. A model with low bias but high variance has high robustness.
B. A model with high bias but low variance has high precision.
0 C. A model with low bias and variance has low precision.
D. A model with high bias and variance performs poorly and will not be used. 5Acti'7! ITERM: D
38, Which of the following statements is true about the low-latency design of CPUs?
A. The low clock frequency reduces the latency.
B. Complex logic control units increase the latency.
C. A CPU is equipped with few ALUs and caches. The caches merge access requests to the same data and
then access the DRAM, reducing the latency.
D. The powerful ALU can complete the computation in a short clock cycle. MT! D
39, Which of the following are common application scenarios of HUAWEI CLOUD OCR?
A. Photo recognition for cats and dogs
B. Driving license recognition
E./ C. VAT invoice recognition
D. Aircraft itinerary recognition
ME*17! EiitntM: BCD
40, Which of the following statements are true about ensemble learning?
A. Ensemble learning uses the decision tree as the bottom-layer algorithm.
B. The voting policy of ensemble learning can be the averaging method. i C. A batch of features are randomly
selected for the subtree training in a random forest.
D. Xgboost supports parallel training like the random forest.
.4417! .BC
41, Which of the following statements are true about grid search based on hyperparameter tuning?
171 A. Grid search exhaustively searches for all possible hyperparameter combinations to form a
hyperparameter value grid. 0 B. Grid search is expensive and time-consuming.
E4'
Grid search works well when there are relatively few hyperparameters.
Grid search suits neural networks well.
p,m7! aiRtm: ABC
42, Which of the following are part of MindSpore?

A. Mind Expression API layer
B. Mind Compiler's real-time graph-level compilation subsystem 1/ C. CUDA and third-party chips
D. MindData
mg7! ABD
43, Which of the following statements are false about the Gated Recurrent Unit (GRU)?
A. Unlike long short-term memory (LSTM), GRU merges the cell state and hidden state.
B. GRU is a variant of a Convolutional Neural Network (CNN).
C. GRU combines the forget and update gates into a single input gate.
D. GRU assembles the output of the previous time step with the input of the current time step. P, in 7 EfAik,E:
BC
44, Which of the following statements are false about common optimizers?
1=1 A. One of the advantages of Adagrad optimizers is that the parameter update operation does not end too
early.
B. When the momentum optimizer is used, parameters are updated by using the same learning rate, but
momentum coefficients keep changing with each iteration.
Ee C. Adam sets different learning rates for different parameters.
D. The RMSprop optimizer introduces an attenuation coefficient to ensure a certain proportion of gradient
attenuation in each round. Val7! EiAnt: AB
45, Which of the following are common ensemble learning algorithms in machine learning?
A. Random forest
B. GBDT
1? C. Polynomial regression
D. Adaboost
git47! ABD
1, Federated learning can protect user privacy to some extent.

True
False
MAVF, M531 ffifAlq?: True
2, MindSpore supports automatic graph splitting, which splits graphs based on the input and output data
dimensions of the operators and integrates data parallelism,
tensor parallelism, and model parallelism.
True
F a l s e IgATI TiThMt:
False
protection, and helping developers focus on Al apps. Due to the
size of the model, MindSpore cannot be deployed on moble phones.
True
F a l s e irAM , .-7 !
frIVAW: False
4, When collecting images for image processing, we usually use numbers (such as 0, 1, 2, 3, and 4) to indicate
image types. Because of this, we must perform one-hot or
dummy encoding on data before training the model.
T r u e CD False 1E,131
Ef1M14: False
5, Object Storage Service (OBS) can be widely used in scenarios with abundant relational data such as social
apps, enterprise relationship analysis, knowledge graph, risk
control, and recommendations.
True
F a l s e tikAisy,fit.V44: False
6, The added cost of Li regularization is proportional to the absolute value of the weight coefficient.
• True
F a l s e a, N7!
TiA1014: True
7, The architecture of CPUs Is different from that of GPUs. CPUs are used for logic control in instruction
execution, while GPUs have outstanding performance in matrix and
parallel computation.
o True
F a l s e TiffiNV:
True
8, Mini-batch gradient descent (MBGD) cannot balance the robustness of stochastic gradient descent and the
efficiency of batch gradient descent, and we may risk getting
stuck at local minima. It Is not commonly used In practice.
True
False
Mf I IrAMz: False
9, MindSpore supports Ascend processors and CPUs, but does not support GPUs.
True
• False
MAR, Nk17! ENIL.M: False
10, As the cornerstone of the Huawei full-stack all-scenario Al solution, Atlas provides modules, cards, and
servers powered by Kunpeng Al processors to meet demands for
computing power in all scenarios.
True
False
MAP, UM-7! INV.: False
11, With the continued development of Al processors and Al frameworks, new data types, such as BFloat16, are
emerging.
True
0False
VAT! ETLAIg: True
12, Distributed training is a feature of the TensorFlow framework.
0 True
0 False
lk,J7! MAI : True
Different label selection standards generate diffe rent
decision tree algorithms.
01 True
0False
MT!06=4:False
14, AIR eliminates model differences between different backends through unified operator IR definitions. You
can coordinate tasks on all platforms (device, edge, and cloud)
based on the same model file.
True
False
7! M i t l I M : T r u e
15, Device, edge, cloud can respectively refer to smartphones, Huawei intelligent edge devices, and HUAWEI
CLOUD.
0 True
False
a, lAny! EfAIRAC True
16, OCR automatically extracts key information about invoices to help employees automatically fill in
reimbursement forms. In addition, the RPA automatic robot can
improve the efficiency of financial reimbursements.
True
False
lg*-17! Ef*INV: True
17, The softplus function has a continuous derivative and defines a smooth curved surface.
True
False
MAO, AMA! ETAV,44: True
18, The spatial relationship of images is that the local pixels are closely related, and the pixels that are far away
from each other are weakly related.
True
False
A O, 1 F ,V7 I EIMIC True
19, The tree model used at the bottom layer of the random forest and GBDT algorithms is only the classification
and regression tree (CART).
True
False
&AM M4-
7! EIAN44: Fal se
20, The dropout technology can be used only during Al model training.
True
False
• Igki7! ETLAIRA: F a l s e
21, Which of the following statements is true about the low-latency design of CPUs?
o A. The low clock frequency reduces the latency.
o B. Complex logic control units increase the latency.
o C. A CPU is equipped with few ALUs and caches. The caches merge access requests to the same data and
then access the DRAM, reducing the
latency.
D. The powerful ALU can complete the computation in a short clock cycle.
MAO,.Nni!ISA4:D
22, Which of the following is a math operator in MinciSpore?
• A. Mul
• B.ControlDepend
C. Callback
• D. Select
•1
.E.7! ETICIM: A
23, Compared with the random forest algorithm, which of the following is false about the GBDT algorithm?
o A. Training using the GBDT algorithm is faster than that of the random forest algorithm (under the same
dataset and tree quantity).
B. The GBDT algorithm is easier to overfit than the random forest algorithm.
o C. The random forest algorithm tends to reduce the variance, while the GBDT algorithm tends to reduce the
bias.
o D. Both the GBDT and random forest algorithms can use the CART as the basic tree model.
IhM7 ! IMAM: A
24, The Relu function is often used in deep learning and neural networks. Which of the following statements is
true about this
function?
o A. The value range is [-1,1].
o B. The value range is [0,-
,Y).
o C. The value range of the derivatives is (0,1).
o D. The value range of the derivatives is [-1,0].
ANY! Eri B
25, Which of the following determines the upper limit of machine learning?
▪ A. Data
ID B. Algorithms
▪ c Scenarios
• D. Computing power
1
,4M71 EfANW: A
26, Which of the following statements is false about the softmax function?
A. The softmax function turns a vector of K real numbers into another vector of K real numbers,
B. The value of each element in the vector obtained from the softmax function ranges from 0 to 1.
C. The elements of the vector obtained from the softmax function do not add up to 1.
D. The softmax function is often used as the output layer of a multiclass classification task.
it1M71 ETANW: C
27, When the vi compatibility package of TensorFlow 2.0 is used to inherit the TensorFlow 1.x code, the eager
operation needs to be disabled. Which of the following
commands is used to disable it?
• A. tfdisable_eager run()
CD B. Um eager_execution0
• CALuneager_execution0
D. tf.disable_eager_execution()
ufta, ISR,13-1 D
28, Which of the following is the type of labels predicted by ensemble learning algorithms?
A. Discrete and continuous
• B. Continuous
® C. Discrete
D. Strain and discrete
.4M7! EfLf0S,A: A
29, The core of an Al framework and one of the decisive factors of the programming paradigm Is the automatic
differentiation used by the Al framework. Which of the
following is used by MindSpore?
A . Gr a p h k e r ne l f us i o n
B. Operator overloading
C. Source to Source (525)
D. Just-in-time compilation
AMT! 0011.g: C
30, Which of the following is covered by HUAWEI CLOUD El Health?
A. Genome
B. Smart surgery
C. Pati ent car e
D. Doctor-patient dispute resolution
1Sin7! ETiffila: A
31, Which of the following is NOT a main function of HUAWEI CLOUD GeoGenius?
A. Natural resource survey
B. Ecological environment detection
▪ C. Weather forecast
® D. Real-time traffic detection
ex1-7! TTLAV,: D
A. The ReLU function is not differentiable at x = 0 and a derivative is forcibly defined at this point.
B. The ReLU function effectively alleviates the vanishing gradient problem.
problems.
D. Compared with Sigrnoid and tanb, the convergence of the ReLU function is slow.
V,V7! TTLANAC D
33, Which of the following statements is true about the Li buffer in Ascend Al Processors?
• A. Data in the Li buffer needs to be read outside the Al Core over the bus interface each time.
B. The Ll buffer can permanently retrain data that needs to be reused.
C. The Li buffer decreases data accesses over the bus and avoids bus congestion.
D. The Li buffer is used to store the initial values in a neural network.
k,MY! TfifOSA: C
34, Which of the following is the main function of HUAWEI CLOUD GeoGenius?
A. Ecological environment detection
B. Heat demand forecast
C. Real-time traffic detection
▪ D. Traffic incident detection
AY! EfLAIM: A
35. Which of the following is not a feature of the Mind pore core architecture?
0 A. Automatic differentiation
B. Automatic parallelism
• C. Automatic deployment
D. Automatic tuning
.E,X11! EfilM14: C
36, Which one of the following actions can the function ttsqueeze be used for in TensorFlow 2.0?
▪ A. Element-wise addition
▪ B. Computation of the absolute value
▪ C. Tensor concatenation
• D. Dimensionality reduction
*M7! D
37, Which of the following is not supported by data management on ModelArts?
A. A wide range of data formats B.
No-manual labeling 0 C.
Intelligent data filtering
D. Intelligent data analysis
liMIT,
, NAY! ENV44: B
38, How does Al perform in the cognitive intelligence phase?
A. Machines begin to compute and transmit information like humans.
B. Machines begin to see, listen, understand, make judgments, and take simple actions.
D. Machines begin to record information, such as pictures and voices.
V.M71 1N,1
=: C
39, Which of the following statements are true about activation functions?
A. The existence of the activation function introduces linearity into the network.
E4 B. If we do not use an activation function, the output signals will be just a simple linear function.
• C. There are many activation functions. You need to select one based on the actual situation.
E4 D. Activation functions play an important role in learning and understanding complex and nonlinear functions
of neural network models.
tkftif.1•,
,
517! TA *: BCD
40, Which of the following products can be equipped with Ascend 110 Processors?
K1A. Atlas 200 DK Al developer kit
ID B. Atlas 500 Al edge station
EaeC. Atlas 200 Al accelerator module
D. Atlas 300 Al accelerator card
W7! ENNW: ABCD
41, Which of the following statements are true about tensors?
• A. The most basic data structure in TensorFlow is tensor.
Li B. A first-order tensor can be considered as a vector.
D C. A tensor is a data matrix.
.2 D. A tensor is a generalization based on the vector and matrix.
,a$,M7! AB D
42, Which of the following operations are involved when TensorFlow is used for model training?
ReA. Define input and output nodes.
Eal, B. Train and save models.
Li C. Obtain and cleanse data.
S,M7! 1LrA,4g: ABD
43, Which of the following cells are predefined in MindSpore?
R A. MAELoss
III B. TFRecord
E4 c cos
III D. WithGradCell
Rif 7i EraW, W : AD
44, Which of the following are supported by data management on ModelArts?
I A. A wide range of data formats
D B. Team labeling
C. Intelligent data filtering
D D. Automatic data migration
M7! EffaM: A B C
45, Which of the following statements are true about regression analysis?
✓
1 A. Regression analysis is a statistical analysis method used to determine the quantitative relationship
between two or more
variables. 111 B. Regression analysis is a type of unsupervised learning.
I C. Regression analysis is a type of supervised learning.
0 D. Linear regression with an absolute loss (L2 regularization) is called Lasso regression.
ih,MI7! ETAVAZ: AC
46, Which of the following use the Da Vinci Architecture?
A. Ascend 310
El B. RTX3080
C. Ku npeng 920
El D. Ascend 910
S,M7! Er04,: AD
47, Which of the following are the advantages of the softsign activation function over the sigmoid function?
▪ A. The maximum value of a derivative is 0.5.
Iy*.. C. The output is zero-centered.
E l D. It is an unsaturated activation function.
VX47! 1-E =: BC
48, Which of the following statements are false about the combination of the model bias and variance?
A. A good model requires a high variance and high bias.
B. A high bias and high variance may result in overfitting.
D C. A high bias and low variance may result in underfitting.
D. A model with a high variance is of no application value.
VA7! EfitiA4: ABD
49, Which of the following statements are true about gradient descent?
E4 A. Stochastic gradient descent (SGD) randomly chooses samples for each training job, causing unstability.
As a result, the loss function fluctuates or even encounters reverse displacements
during the process of dropping to the minimum value.
❑ B. Batch gradient descent (BCD) is the most unstable method and consumes too many compute resources.
Ke, C. Mini-batch gradient descent (MBGD) is a balance between BGD and SGD, and is the optimal choice for
all datasets.
Ke D. BGD uses all samples for each training job.
14,4V.v, lan7! EfIVX=: ACD
50, What are the theories of cognitive intelligence?
A. Threshold theory
B.The theory of evolution
C.Behavioral theory
D. Logic theory
intn7! EfLATS,: ABD
51, Which of the following statements are true about the Mind pore components?
Re, A. communication: processes data flow communication between the CPU and memory.
▪ B. dataset: dataset processing module that loads and preprocesses data.
Re. C. nn: neural network cells in MindSpore that define loss functions and optimizers.
Re, D. train: relates to training model and model quantization module.
Efo4,4,: BC
52, Which of the following statements are true about using OCR to train an invoice recognition model with the
support of a dedicated Al pr ocessor?
Ke, A. Training requires a large amount of data.
Re, B. Supervised learning is often used.
Ke, C. The processor has demanding requirements on the computing power.
E.', D. The trained model may be a complex deep neural network.
la17! WifalM A E C D
53, Which of the following methods provided by TensorFlow 2.0 cannot be used to check whether an object is a
tensor?
❑A.istensor
Ke, B. isinstance
E", C. device
Kt' D. tftypes
4XAM.
, Igki71 EifolgV: BCD
54, Which of the following statements are false about Keras?
I?, A. Like TensorFlow, Keras is a multi-layer neural network development package. However, Keras has simpler
syntax and is easier to use.
FY; B. Keras is a neural network development package used to build CNN sequential models. It cannot be used
to build other neural networks.
• C. The neural network model built using Keras must be compiled before data can be input into it for training.
• Keras can be used as the backend of TensorFlow.
2kM7! Erki,24=: BD
55, As the neural network increasingly deepens, two issues may be encountered during network training: one is
0, and the other is gradient vanishing/exploding.
=71 f6 (1) t6t; MAIXV: (1) Network degradation; network degradation: Network Degradation
56, At a convolution layer, there are 256 5 x 5 convolution kernels, the size of an input feature map is 32 x 32 x
5, the convolu tion step is 2, and zero padding is
employed(padding size=1). So, the size of the output feature map is O * 0 * 0. (Fill in each bl ank with a
number.)
4#147! 1J2 (1), (2). (3) MR; MIVAV: (1) 15 (2) 15 (3) 256
57, HUAWEI CLOUD () helps enterprises build invoice management systems and self-service reimbursement
systems. (Fill in the blank with the abbreviation of the service
name.)
gm / ! 15 (2(11 MN; A7lM4: (1) OCR;ocr
58, The average value of tensors generated using ttrandom.normal([4], 2, 1, tf.float32) is 0. (Fill in the blank with
a
number.)
*MT! 11A (1) MiR; HiLIZM: (1) 2
59, In an Magrad optimizer, the initial value of the cumulative square gradient is 0. (Fill in the blank with a
number.)
At (1) MX; fffia.44: (1) 0
60, The Naive Bayes algorithm needs to obtain the 0 probability of the dataset.
r.,7! (1) MR; EflifiVA: (1) prior
1, The PIL module in TensorFlaw 2.0 is used to perform operations on images.

(.,) True
False
So r ry ! Co r r ec t A nsw e r: F al s e
2, Assuming a dataset contains the areas and prices of 21,613 housing units in a city, a classification model can
predict the prices of
other housing units in the city
True
CU False
Cong ratulation s! Co rre ct Answ er : F alse
• True
False
S o rr y ! C o r re c t A n s w e r: F a l s e
4, Grid search randomly selects hyperparameter combinations. Therefore, it has a high efficiency and accuracy.
True
0 False
C o n g rat u la tio n s! Co rre c t An sw e r: F als e
5, During model training, a larger value of the epoch parameter indicates a better training result.
0 True
False
So r ry ! Co r re c t An sw e r: F al se
6, The added cost of L1 regularization is proportional to the absolute value of the weight coefficient.
True
0 False
S o rr y ! C o r re c t A n s w e r: T ru e
7, ModelArts allows visualized management throughout the Al development lifecycle, including data preparation,
training, modelin g,
and inference. It does not support resumed training, training result comparison, or model lineage.
True
False Sorry ! Correct Answer:
False
8, Ascend Al Processors are a series of Huawei-proprietary Al processors that feature superior computing
power.
True
False
Sorry!CorrectAnswerTrue
significantly improving
performance.
True
False
C o n g r a t u l a t i o n s ! C o rr e c t A n s w e r. T ru e
10, As the cornerstone of the Huawei full-stack all-scenario Al solution, Atlas provides modules, cards, and
servers powered by
Kunpeng Al processors to meet demands for computing power in all scenarios.
True
False
11, The ModelArts inference platform implements fast and efficient model inference and image recognition.
T ru e
False
S o r r y ! C o r r e c t A n s w e r - T ru e
12, HUAWEI HiAI 3.0 supports the development of P50 smartphones.
True
False
Congratulations! Correct Answer: True
13, The Tensor Boost Engine (TBE) is a processor enablement layer built by Huawei for deep neural networks
and Ascend Al
Processors. It provides a processor operators library and a highly automated operators development toolkit.
True
False
14, In a convolutional neural network, the size of each kernel is not necessarily the same as that of a pooling
layer window, but their
steps should be the same.
True
False
S o rr y ! C o r re c t A n s w e r: F a l s e
15, Filter methods for feature selection are restricted they do not tend to select redundant variables because
they consider the
relationships between features.
0True
False
C o n g r a t u l a t i o n s ! C o rr e c t A n s w e r F a l s e
16. Although each industry and application faces unique challenges, they are all gradually adopting Al-based
workflows and
solutions.
C:) True
0 False
Congratulations!ConanAnswerTrue
17, The spatial relationship of images is that the local pixels are closely related, and the pixels that are far away
from each other are
weakly related.
True
•False
S o rr y ! C o r re c t A n s w e r: T ru e
Different label selection
standards generate different decision tree algorithms.
0 True 0
False
Sorry!CorrectAnswer:False
19, One of the filter methods is using the Chi-Squared test to analyze the mapping between features and labels
to implement feature
selection.
T ru e
False
Sorry ! Corre ct An swe r: True
20, Dropout randomly discards certain neurons during training. so the parameters corresponding to the
discarded neurons are not
involved in the next round of training.
True
F a l s e Sorry ! Correct
Answer: False
o A. The ReLU function is not differentiable at x = C and a derivative is forcibly defined at this point.
o B. The ReLU function effectively alleviates the vanishing gradient problem.
problems.
o D. Compared with Sigmoid and tank, the convergence of the ReLU function is slow.
Sorry!CorrectAnswer:D
22, Which of the following statements is true about optimizers?
o A. The RMSprop optimizer inherits the advantages of the Adam optimizer. The learning rate is automatically
updated.
B. Unlike the RMSprop optimizer, the Adagrad optimizer does not automatically update the learning rate.
o C. The SGD and momentum optimizers use the same learning rate for each iteration.
o D. The learning rate of the momentum optimizer does not need to be manually set.
sorry ! Correct Answer: C
23, "Deep learning is a complex, comprehensive discipline. It includes Al and machine learning." Which of the
follo wing is
true about this statement?
A. This statement is correct. Deep learning includes Al.
o B. This statement is correct. Deep learning includes machine learning.
o C. This statement is incorrect. Machine learning has nothing to do with deep learning.
o D. This statement is incorrect. Al includes deep learning and machine learning.
Sorry ! Correct Answer: D
24, Dirty data refers to data with quality problems. Which of the following statements is false about the data
quality?
o k Noise: Data contains incorrect records or exceptions.
• B. Incompleteness: Data tacks attributes or contains some missing values.
o C. Unprocessed: Data for which feature engineering has not been performed.
o a inconsistency. Data contains conflicting records.
Sorry!CorrectAnswerC
25, Which of the following indicators cannot be used to evaluate a model?
A Generalization capability
B. Ex pla i na bil it y
C. Prediction rate
D. Code complexity
Sorry!CorrectAnswerD
26, Compared with the random forest algorithm, which of the following is false about the GBDT algorithm?
• A. Training using the MT algorithm is faster than that of the random forest algorithm (under the same dataset
and tree quantity).
❑ B. The GBDT algorithm is easier to overfit than the random forest algorithm.
❑ C. The random forest algorithm tends to reduce the variance, while the GBDT algorithm tends to reduce the
bias.
D. Both the GBDT and random forest algorithms can use the CART as the basic tree model.
Sorry I Correct Answer: A
27, Which of the following commands can be executed to use the code of TensorFlow 1.X in TensorFlow 2?
• A. import tensorflow.compatvl as tf
❑ B. import tensorflow.vl as tf
• C. import tensorflow as tf
D. import v1 as tf
Sorry ! correct Answer: A
28, concat_sample_1 = ttrandom.normal([4,100,100,3])811tAr>concat_sample_2 ,-
tfrandom.normal([40.100,100.3])&itlbr>concated_sample_l =
aconcat([concat_sample_Lconcat_sample_2],axis=0)8(itArgegt;print(concated_sample_1.shape)8tItibrgigtWhich
of the f ollowing
is the output for this code?
A. (44.200200,6)
▪ B. (44,100,100,3)
❑ C (40,100,100,6)
a ( 4 4 1 0 0, 1 0 0, 0
Sorry ! Co rrec t A n swer B
29. Which of the following is the purpose of running ocr_client = HWOcrClientAKSK(ak, sk, region) when using
HUAWEI CLOUD
OCR?
A. Initialize the client.
B. Transfer images for recognition.
❑ C. Delete image data
❑ D. Encode image data.
Sorry 1 Correct Answer. A
30, Which of the following is the correct shape of tensor 110,11. [2,3], [4,5]]?
D 1, (2,3)
❑8.(3,2)
❑C.(1.6)
❑D.(6,1)
Sorry ! Correct Answer! B
31, Which of the following statements is true about classification and regression models in machine learning?
A. For classification problems, the output variables are discrete values with a limited quantity. For regression
problems, the output variables are
continuous values.
B. The cross entropy loss function is required for both regression and classification problems.
C. There may be underfitting in both regression and classification problems. Classification models are more
prone to underfitting than regression
models.
D. The logistic regression model can be used to predict housing unit prices.
Sorry!CorrectAnswer:A
32, The core of an Al framework and one of the decisive factors of the programming paradigm is the automatic
differentiation used
by the Al framework. Which of the following is used by MindSpore?
A. Graph kernel fusion
B. Operator overloading
® C. Source to Source (S25)
o D. Just-in-time compilation
Congratulations! correct Answer: C
33, When the v1 compatibility package of TensorFlow 2.0 is used to inherit the TensorFlow 1.x code, the eager
operation needs to be
disabled. Which of the following commands is used to disable it?
A. tf.disable_eager_runO
B. tf.noeager execution°
C. tfuneager_execution0
D. tidisable_eager execution()
34, Which of the following commands is used to install TensorFlow?
k conda create install tensorflow
B. python tensorflow setup.py
C. conda setup.py tensorflow
Congratulations! Correct Answer, D
35, Which of the following statements is true about model validity?
A. The goal of machine learning is for a trained model to perform well on new samples, not just on samples
used for training.
B. Variance refers to the difference between the prediction of a learned model on a sample and the actual
result of the sample.
C. During model testing, errors can be classified into training errors and sample variances.
D. Bias refers to errors you get when you run the model on new samples.
sorry ! Correct Answer: A
36, Which of the following is not a hyperparameter for model training?
A. batch_size in a training neural network
❑ B . K i n K •m e a n s
71 C. b in the one dimensional linear regression model Y -wX c b
❑ D. Number of trees in a random forest
Cong rat ul ations ] Co rre ct Answ er C
37. Which of the following is the type of labels predicted by ensemble learning algorithms?
A- Discrete and continuous
C B. Continuous
C.Discrete
Sorry ! Correct Answer: A
38, Which of the following is the correct shape of tensor [1[0.1],[2,3]],[[4,5],[6,7]]]?
A. [3,3,2]
B. [32,4] )
C. 12,3,4]
• D. [2,22]
Sorry!CorrectAnswerD
39. Which of the following are used for backpropagation?
❑ A. Number of network layers
❑ B. Activation functions
• C. Number of neurons at a single Payer
❑ D. Neuron weight
SonyCorrectAnswerBD
40. Which of the following statements are false about TensorFlow?
❑ A. TensorFlow can only be used in a deep learning algorithm.
✓B. TensorFlow supports only GPUs and CPUs.
❑ C. TensorFlow 2.0 uses the dynamic graph mechanism by default. with a higher running efficiency than the
static graph mechanism
❑ D. TensorFlow 1.0 uses the graph session mechanism
Sorry I Correct Answer: ABC
41. Which of the following statements are false about the combination of the model bias and variance?
❑ A. A good model requires a high variance and high bias. -
-' B. A high bias and high variance may result in overfitting.
❑ C. A high bias and low variance may result in underfitting
✓D. A model with a high variance is of no application value.
SorryCorrectAnswerABD
42, Which of the following functions are supported by the HiAl Engine platform?
❑ A. Form recognition
El. B. Keyword extraction
C . V id eo s um mar iz at io n
✓ D Automatic generation of video thumbnails
S o r r y ! C o rre c t A n s w e r: A B C D
43, Which of the following are CNN-based network structures?
•ResNet
❑B.Transformer
✓C. Inception
❑ D. BERT
Congratulations? Correct Answer: AC
44, Which of the following statements are true about linear regression?
• A. The error of linear regression is affected by many independent factors. According to the central limit
theorem (CLT), the err or follows normal
distribution.
❑ B. The loss function of linear regression can be obtained using the normal distribution function an d maximum
likelihood estimation (MLE).
✓ C. The result of multivariate linear regression analysis may be a hyperplane in a high -dimensional space.
0. Due to algorithm complexity, linear regression cannot use the gradient descent method to calculate the
weigh t parameter if the loss function
reaches the minimum value.
S o r r y ! C o r r e c t A n sw e r : A 8C
45, Which of the following statements are true about data conversion?
❑ A. Image data processing, such as color space conversion. grayscale image conversion, and geometric
conversion, is a part of data conversion.
at" B. in machine learning, features are often normalized to ensure that the value ranges of different variables i
nput to the same model are the same.
✓ C. When the data dimension is high, the existing variables need to be combined or converted to generate new
features, such as averages.
❑ D. All data needs to be converted.
S o r r y ! C o r r e c t A n s w e r. A B
46, Which of the following statements are true about MindSpare and Huawei all -scenario solutions?
❑ k The full scenario deployment solution includes model generation and efficient execution.
❑ B. MindSpare Serving supports batching.
• C. Distributed training of ultra-large models and ultra-large datasets requires only data parallelism.
❑ D. MindIR is the key to 'Train Once, Deploy Everywhere.'
Sorry ! Correct Answer: ABD
47, Which of the following statements are true about using OCR to train an invoice recognition model with the
support of a
dedicated Al processor?
Ee A. Training requires a large amount of data
❑ B. Supervised learning is often used
• C The processor has demanding requirements on the computing power
❑ D The trained model may be a complex deep neural network.
Sorry I Correct Answer: ABCD
48, Which of the following operations are involved when TensorFlow is used for model training?
❑ A. Define input and output nodes.
DB. Train and save models. C. Obtain and
cleanse data 1? D. Define learning parameters
and operations.
Sorry 1 Correct Answer ABD
49, Which of the following statements are true about decision trees?
❑ A. A decision tree is a binary tree structure-
B. Each non-leaf node in a decision tree represents a test on a feature attribute. each branch represents the
output of the feature attribute within a
value range, and each leaf node holds a class label.
• C Except the root node, all nodes in a decision tree are called leaf nodes. C D Building a decision
tree means selecting and measuring feature attributes and determining their topology.
Sorry ! Correct Answer. BD
SO, Which of the following are the advantages of the softsign activation function over the sigmoid function?
A. The maximum value of a derivative is D.S.
❑ B. Gradient saturation slows down.
C The output is zero centered.
❑ D. It is an unsaturated activation function.
So rry ! Co rrec t Answe r, BC
Sorry Blan k ite ms CO is inco rre ct. Co rrect A nswe r: (11 C allba ck ; c allba ck
51, Which of the following statements are true about the MindSpore components?
❑ A_ comm unication: process es data flow comm unication betw een the CPU and m emory.
• B. datas et datas et proces si ng modul e that loads and preprocess es data.
✓ C. nn: neural network cel ls i n Mi ndSpore that defi ne l os s f uncti ons and opti mi z ers .
❑ D. train: relates to training model and model quantization module.
C o n g rat u la tio n s! Co rre c t An sw e r: B C
52, The accuracy of a trained model is high on a training dataset but low on a test dataset. In this case, which of
the follo wing
operations can be performed?
A Use the sigmoid activation function.
E". B. Perform data augmentation.
❑ C . A d d r e g ul a r t e r m s .
❑ D. Delete part of the training data
sorry ! coned A nsw e r: B C
53, Which of the following statements are true about the three main schools of Al?
❑ A. Behaviorism states that intelligence requires no knowledge, representation, or reasoning.
C B. Connectionisrn states that Al can evol ve like hum an i ntelligence.
1-2' C. Behaviorism states that intelligent behavior can only be expressed through constant int eraction between
humans and the surrounding
environment in the real world.
D. Symbolism states that Al focuses on behavior control, adaptive computation, and evolutionary computation.
Sorry!CorrectAnswer:AC
54, Which of the following statements are true about activation functions?
A. The existence of the activation function introduces linearity into the network
C B. If we do not use an activation function, the output signals will be just a simple linear function.
• -•" C. There are many activation functions. You need to select one based on the actual situation.
C D. Activation functions play an important role in learning and understanding complex and nonlinear functions
of neural network m odels.
sorry I Correct Answer- BCD
55. As the neural network increasingly deepens, two issues may be encountered during network training: one is
0, and the othe r is
gradient vanishing/exploding.
S o rr y ! B l a n k i t e m s 1 1 1 i s i n c o r re c t . C o r re c t A n s w e r: ( 1 1 N e t w o rk d e g r a d a t i o n ; n
e t w o rk d e g r a d a t i o n ;N e t w o rk D e g r a d a t i o n
56, Linear regression using a loss function wit h the LO -norm reg ular term is also called ridge re gression.
S o rry I B la n k i t e ms Ill i s in c o rre c t . C o rre ct A n sw e r, 2
57, In MindSpore, if you want to define a function for collecting information after each step is complete, you
need to inherit th e 0
class.
58. At a convolution layer, there are 128 3 x 3 kernels and the size of an input feature map is 28 x 28 x 64
(without padding and with a
step of 1). So, the depth of the convolution kernels is Q. (Fill in the blank with a number.)
Sony ! Blank items {1111 incorrect. Correct Answer (1 64
59, Assume that there are 10,000 data pieces in a cancer data sample, of which 100 pieces are from cancer
patients, and the other
9900 pieces are normal. If a classification model predicts that 9000 out of the 9900 pieces are normal, and 90
out of the 100
pieces are from cancer patients, the accuracy rate of the model is 0%.
Sorry Blank items is incorrect. Correct Answer' Ill 90.9
60, The 0 algorithm (abbreviation) is used when RNN performs backpropagation (8P).
Sat ry Blank items 111 is incorrect. Correct Answer: 11) BPTT:bptt
The tree model used at the bottom layer of the random forest and GBDT algorithms is only the classification and
registration tree (CART)
False
Distributed training is a feature of the TensorFlow framework.
True
During model training, a larger value of the epoch parameter indicates a better training result.
False
Mini-batch gradient descent (MGBD) cannot balance the robustness of stochastic gradient descent and the
practice.
False
MindSpore accelerates model convergence through automatic parallelism and second-order optimization.
True
Air eliminates model differences between different backends through unified operator IR definitions. You can
coordinate tasks on all platforms (device, edge, and cloud) based on the same model file.
True
Assuming the number of hyperparameters is the same, stochastic gradient descent (SGD) combined with
manual adjustment achieves a better effect than adaptative learning rates.
False
An AI processor is also referred to as an AI accelerator. It is a functional module used to process a large
quantity of computing tasks in an AI application. Ascend 910 is typical product.
True
Federated learning can protect user privacy to some extent.
True
In the convolutional layer, the dropout ration is the ratio of features whose values are set to 0. Generally, the
dropout ratio ranges from 0.2 to 0.5.
True
As the compute core of an Ascend AI Processor, the AI Core is responsible for complex matrix computation.
True
A decision tree selects a label from features of the trining data to work as the node splitting standard. Different
label selection standards generate different decision tree algorithms.
False
In a convolutional neural network, the size of each kernel is not necessarily the same as that of a pooling layer
window, but their steps should be the same.
False
The ModelArts inference platform implements fast and efficient model inference and image recognition.
True
A model constructed using machine learning algorithms can be a function that conforms to actual data
distribution.
False
Convolutional neural networks cannot be used to process text data.
False
If a 32 x 32 image (without padding) is input and a 5 x 5 kernel is used for convolution with a step of 2, the size
of the output image is 13 x 13.
False
One of the filter methods is using the Chi-Squared test to analyze the mapping between features and labels to
implement feature selection.
True
The softplus function has a continuous derivative and defines a smooth curved surface.
True
MindSpore supports Ascend processors and CPUs, but does not support GPUs.
False
Which of the following statements is true about rule-based approaches?
In TensorFlow 2.0, x = tf.constant([1,2,3])y = tfibroadcast_to(x,[3,3])print(y) Which of the
B. [[1,2,3], [1,2,3], [1,2,3]]
How does Al perform in the cognitive intelligence phase?

Which of the following indicators cannot be used to evaluate a model?

d. Code complexity
concat_sample_1 = tfrandom.normal([4,100,100,3])concat_sample_2 = tfrandom.normal([40,100,100,3))

concated_sample_1 = tf.concat([concat_sample_1,concat_sample_2],axis=0)
print(concated_sample_1.shape) Which of the following is the output for this code?
B. (44,100,100,3)
An image recognition experiment uses 42,510 training images. There are more than 200 categories, each of
which contains only 10 images. There are also categories with only 20 to 50 images. Which of the following data
problems does this phenomenon fit into?
B. Data imbalance
"Knowledge representation is the unique method of representing knowledge using a set of symbols in a
structure that can be understood by computers." Which of the following is true about this statement?
A. This statement is false. The method of representing knowledge is not unique.
Which of the following statements is true about optimizers?
c. The SGD and momentum optimizers use the same learning rate for each iteration.
Which of the following is the basic unit of data storage in PyTorch, MindSpore, and TensorFlow?
b. tensor
Which of the following determines the upper limit of machine learning?
A Data
Which of the following HUAWEI CLOUD services is NOT a basic Al platform service?
C. Relational Database Service (RDS)
Which of the following statements is false about the ReLU function?
Which of the following is not a hyperparameter for model training?
C. b in the one-dimensional linear regression model Y=wX+b
The core of an Al framework and one of the decisive factors of the programming paradigm is the automatic
differentiation used by the Al framework. Which of the following is used by MindSpore?
C. Source to Source (2S)
Which of the following is NOT a main function of HUAWEI CLOUD GeoGenius?

D. Real-time traffic detection
Which of the following statements is true about neural networks?
C. LSTMis suitable for processing events with long interval and delay in the time sequence.
Which of the following commands is used to install TensorFlow?
Compared with the random forest algorithm, which of the following is false about the GBDT algorithm?
A. Training using the GBDT algorithm is faster than that of the random forest algorithm (under the same dataset
and tree quantity).
The accuracy of a trained model is high on a training dataset but low on a test dataset. In this case, which of the
following operations can be performed?
B. Perform data augmentation.
C. Add regular terms.
Which of the following format conversion operations are not performed by the storage control unit of the DaVinci
Architecture?
B. Video stream compression
C. Color space conversion
Which of the following statements are false about GANs?

a. The discriminator input is mainly noise and real sample data.
d. The input of the discriminator is mainly sample data provided by the generator.
Which of the following statements are true about gradient descent?
A. Stochastic gradient descent (SGD) randomly chooses samples for each training job, causing unstability. As a
result, the loss function fluctuates or even encounters reverse displacements during the process of dropping to
the minimum value.
C. Mini-batch gradient descent (MBGD) is a balance between BGD and SGD, and is the optimal choice for alll
datasets.
Which of the following are used for backpropagation?
B. Activation functions
D. Neuron weight
Which of the following statements are true about tensors?

a.The most basic data structure in TensorFlow is tensor.
b.A first-order tensor can be considered as a vector.
d. A tensor is a generalization based on the vector and matrix.
Which of the following statements are false about Keras?
B. Keras is a neural network development package used to build CNN sequential models. It cannot be used to
build other neural networks.
D. Keras can be used as the backend of TensorFlow.
Which of the following statements are true about decision trees?
B. Each non-leaf node in a decision tree represents a test on a feature attribute; each branch represents the
output of the feature attribute within a value range; and each leaf node holds a class label.
D. Building a decision tree means selecting and measuring feature attributes and determining their topology.
Which of the following cells are predefined in MindSpore?
a.MAELoss
d. WithGradCell
Which of the following are the advantages of the softsign activation function over the sigmoid function?
Which of the following statements are true about common activation functions in deep learning?
C. The tanh function is symmetric with respect to the origin, and the mean of its output is closer to 0.
D. During training of a deep neural network, the sigmoid, tanh, and softsign functions cannot prevent the
vanishing gradient problem.
Which of the following operations are involved when TensorFlow is used for model training?
A. Define input and output nodes.
B. Train and save models.
Which of the following statements are true about linear regression?

A. The error of linear regression is affected by many independent factors. According to the central limit theorem
(CLT), the error follows normal distribution.
B. The loss function of linear regression can be obtained using the normal distribution function and maximum
likelihood estimation (MLE).
C. The result of multivariate linear regression analysis may be a hyperplane in a high-dimensional space.
Which of the following are common ensemble learning algorithms in machine learning?
A. Random forest
B. GBDT
D. Adaboost
Which of the following statements are true about using OCR to train an invoice recognition model with the
support of a dedicated AI processor?
ABCD
Which of the following are model hyperparameters?
A. Learning rate, iteration count, and batch size in a training neural network
B. K in the k-nearest neighbors (KNN) algorithm
C. Number of trees in a random forest
The average value of tensors generated using tf.random.normal([4], 2, 1, tf.float32) is 0. (Fill in the blank with a
number.)
2
At a convolution layer, there are 256 5 x 5 convolution kernels, the size of an input feature map is 32 x 32 x 5,
the convolution step is 2, and zero padding is employed(padding size=l). So, the size of the output feature map
is * 0 * 0. (Fill in each blank with a number.)
(1) 15
(2) 15
(3) 256
In MindSpore, if you want to define a function for collecting information after each step is complete, you need to
inherit the 0 class.
(1) Callback;callback
In TensorFlow 2.0, if tf.keras.layers.RNN is used to process timing information and you want to obtain the output
status at each moment, set () to True.
(1)return_sequence
HUAWEI CLOUD 0 helps enterprises build invoice management systems and self-service reimbursement
systems. (Fill in the blank with the abbreviation of the service name.)
(1)OCR;ocr
HUAWEI HiAl built an open smart ecosystem based on the " device, chip" three-layered open architecture. It
enables developers to quickly utilize Huawei's powerful A1 processing capabilities to provide a better smart
application experience for users.
(1)Cloud;cloud;CLOUD
1. Grid search randomly selects hyperparameter combinations false
2. The Tensor Boost Engine (TBE) is a processor enablement layer built byHuawei for deep neutral networks
and Ascend AI Processors. False
3. With the continued development of AI processors and AI frameworks, new data types, such as BFloat16,
are emerging True
4. Assuming a dataset contains the areas and prices of 21,613 housingunits in a city, a classification model
can predict the prices of otherhousing units in the city
False
5. Poisson distribution is generally regarded ass the common distributionof errors in linear regression.
False
6. Distributed training is a feature of the TensorFlow framework. TRUE
7. When collection images for image processing, we usually use numbers(such as 0,1,2,3 and 4) to indicate
image types.
False
8. Although each industry and application faces unique challenges, theyare all gradually adopting
True
9. In TensorFlow 2.0, tf.data contains methods to write data True
10. The softplus function has a continuous derivative and definesasmooth curved surface.
True
11. Convolutional neural networks cannot be used to process text datafalse
12. In convolutional neural networks, the size of each kernel is not necessarily the same as that of a pooling
layer window,
False
13. MindSpore supports automatic graph splitting, which splits graphs false
14. In the Convolutional layer, the dropout ratio of features whosevaluesare set to 0.Generally, the dropout ratio
ranges from0.2 to 0.5 true
15. Ascend AI Processors are a series of Huawei-proprietaryAI processors that feature superior computing
power
True
16. The PIL module in TensorFlow 2.0 is used to performoperations onimages.

False
17. MindSpore supports Ascend processors and CPUs, but does not support GPUs.
False
18. The softmax function turns a vector of K real numbers intoanothervector of K real numbers,
False
19. One of the filter methods is using the Chi-Squared test to analyzethemapping between features and labels
to implement features andlabelsto implement feature selection.
True
20. The Relu function is often used in deep learning andneutral networks:
B the valuse range is [0, +)
21. Which operation is not a step in the network defenitionprocessduring application development:
D model compression
22. Which of the following is covered by HUAWEI CLOUD EI Health: B Genome
23. Which of the following is the reasoning method of the productionsystem that draws conclusions through a
rule library: A Froward
24. Which of the following HUAWEI CLOUD is NOT a basic AI platform

service :
Relational Database Service (RDS)
25. A vendor wants to provide an intelligent EMR systemfor a hospital: Neutral language processing
26. Which of the following statements is true about variance andbias: A model with high bias and variance
performs poorly and will beused. 27) A generative adversarial is like a game system. The
generatorproduces fake samples, and the discriminator tries to distinguish: The discriminator cannot
distinguish real data fromthe data createdbythe generator.
27. Which of the following is the correct shape of tensor [[0,1],]: (3,2)
28. Math operator in MindSpore :

Mul
29. True about about optimizers:

The SGD and momentum optimizers
30. Is true about model validity:

The goal of machine learning is for a tr
31. Which of the following is NOT an application of speech processing: Named entity recognition
32. Which of the following is the purpose of running ocr_client =HWOcrClientAKSK(ak, sk, region) when using
HUAWEI CLOUDOCR
Initialize the client
33. Which of the following statements is true about the low-latencydesign of CPUs.
The powerful ALU
34. "Deep learning is a complex, comprehensive discipline. AI includes deep learning and machine learning 36)
When the v1 compatibilty package of TensorFlow2.0 is usedtoinherit the TensorFlow 1.x code, the eager
operations needs tobedisabled.
Tf.disable_eager_execution()
35. Which of the following is NOT a main function of HUAWEI CLOUDGEOGenius

Real-time traffic detection
36. Which of the following statements are true about grid searchbasedon hyperparameter tuning ?
NOT grid search suits neutral networks well
37. Which of the following are the advantages of the softsign activationfunction over the sigmoid function ?
Gradient sat
The output
38. Which of the following methods provided by TensorFlow2.0cannot be used to check whether an object is a
tensor ?
Not is-tensor
39. Which of the following statements are true about ensemble learning?The voting
A batch of features
40. Are true about the three main schools main schools of AI ? Behaviorism (zouz)
41. Which of the following statements are true about gradient descent ?Not batch gradient descent BGD
42. Which of the following cells are predefined in MindSpore ? MAELoss

cos
43. Are true about using OCR to train an invoice recognition model lkol
44. Which of the following statements are true about hidden layers The more hidden layers,
Excessive hidden layers may cause overfitting 47) Which of the following statements are false about the
GatedRecurrent Unit (GRU)?
GRU is a variant
GRU combines
45. In TensorFlow 2.0, which of the following cannot be used tocreatean all-zero tensor ?
Tf.zeros_frame() / tuple()
46. Which of the following functions are supported by HiAI Engineplatform?

lkol
47. Which of the following are common application scenarios of HUAWEI CLOUD OCR?
Not photo recog
48. Common application scenarios of HUAWEI CLOUD OCR? Not behavioral theory
49. WhAT are the theories of cognitive intelligence? Ila the existence
50. Are true about activation functions?

The tanh function
During training
51. Which of the following statements are true about commonactivations in deep learning?
Callback;
52. In MindSpore,if you want to define a function for collectinginformation after each step is complete, you need
to inhert the () class. 2
53. Linear regression using a loss function with the L()-normregular termis also called ridge regression
2
54. Generally, a mode with a () volume has a higher precision, andamodel

Large
small
55. HUAWEI HiAI built an open smart ecosystem based on the " Cloud
56. In MindSpore Python APIs, mindspore.() provides various neural network layers.
NN
57. At a convolution layer, there are 128 3 x 3 kernels and the sizeof aninput feature map is 28 x 28 x 64 .
64
By "all Artificial Intelligence scenarios", Huawei means different deployment scenarios for Al, including public
clouds, private clouds, edge computing in all forms, industrial IoT devices, and consumer devices.
true
Which of the following are Al application fields?
Smart household
Smart healthcare
Smart city
Smart education
Gradient descent iteration is the only method of machine learning algorithms.
false
Which of the following algorithms is not supervised learning?

K-means
Compared with the recurrent neural network, the convolutional neural network is more suitable for image
recognition.
true
GAN is a deep learning model, which is one of the most promising methods for unsupervised learning of
complex distribution in recent years.
True
There are many types of deep learning neural networks. Which of the following is not a deep learning neural
network?
Logistic
There are many important "components in the convolutional neural network architecture. Which of the following
are the convolutional neural network components"?
Activation function
Convolutional kernel
Pooling
Fully connected layer
In TensorFlow 2.x, eager execution is enabled by default.

true
Which of the following statements about tf.keras. Model and tf.keras. Sequential is incorrect when the tf.keras
interface is used to build a network model?
D. tf.keras. Sequential is recommended for model building when a sharing layer exists on the network.
In MindSpore, which of the following is the operation type of nn?
Network
What are the main applications of Ascend 310?

Model inference
What are the advantages of the HiAl mobile computing platform?
Comprehensive documents
Source code enabling quick start
Different types of APIs
Various tool chains
Which of the following scenarios can El be applied to?
Intelligent government
Intelligent city
Intelligent manufacturing
Intelligent finance
1. Faced with the challenge of achieving efficient distributed training for ultra-large-scale models, MindSpore is
handled as?
• Automatic parallel
2. Which of the following is not the difference between Python 2 and Python 3? import
Probability theory
4. The Python dictionary is widely identified by "1". and the internal data consists of the key and its
corresponding value
True
5. In the process of training the neural network, we use the gradient descent method to continuously update
which value, which makes the loss Function minimization?
Parameter
1. Huawei Machine learning Service MLS MLS is a one-stop platform that supports the entire process of data
analysis. Which of the following is not a feature of MLS?
• Support for the R language but does not support the Python language
2. GBDT Compared with the random forest algorithm, which of the following statements is wrong?
GBDT Algorithm is easier to overfit than random forest.
1. The following code was used when compiling the model: model.compile(optimizer='Adam,
loss='categorical.crossentropy',metrics=[tf.keras.metrics. accuracy]), currently using evaluate When the
method evaluates the model, which of the following indicators will be output?
• accuracy
• loss
2. According to Huawei Cloud El Intelligent platform, which of the following solutions can be provided?
•Crowd statistics program
• Policy query scheme based on knowledge graph
• Vehicle identification scheme
• Intrusion recognition scheme
3. Which of the following does not belong to long-term memory LSTM (long Short Term Memory) network
architecture?
• Memory door
Function.
• False
5. Which of the following options is not the Tensorflow build process?

• Update weights
6. Artificial intelligence is the research and development of theories, methods and application systems used to
simulate, extend and expand human intelligence Of a new technological science.
• TRUE
7. TensorFlow Which of the following companies first developed?
Google
1. In Huawei Cloud El Enterprise Intelligence, which basic platform services are included? (Multiple Choice)
• Machine learning
• Deep learning
• Graph engine
• Batch processing
2. The activation function plays an important role in the neural network model learning and understanding of
very complex problems. The following statement about the activation function is correct.
• Activation functions are non-linear functions
3. PyTorch Which of the following functions does not have?
• Inline keras

• Python modules can define functions classes and variables but the module does not contain executable
code.
Sigmoid
Tanh
relu

• numbers • string • list • tuple • dictionary
2. When the voice recognition service is successfully called, which field is the recognition result stored in?
• result
3. What does factors that promote the development of artificial intelligence not include?
• Block chain
4. The image data of the passport recognition service is not needed base64 Coded.
• FALSE
5. The loss function of logistic regression is the cross-entropy loss function.

• TRUE
6. Pytorch Which company launched it first?
Facebook
1. It is fine to pass in only one image when calling the face comparison service.
FALSE
1. Which of the following options is not the- session mode used by Tensorflow?
Multiple POST queries
1. twenty two. Among machine learning algorithms, which of the following is not unsupervised learning?
Xgboost
Image style migration
legibiltty
Simplicity
Scalabiliy
1. What are the results returned by the if conditional statements in the Python language? (Multiple choice)
TRUE
FALSE
The verification set can coincide with the test set.
1. Which of the following can optimization problem be classified according to the constraints? (Multiple choice)
Equality constraints
Inequality constraints
Unconstrained conditions
1. Which of the following are the characteristics of the Python language? (Multiple choice)
Object-oriented
Dynamic data type
1. Which of the following are the elements of artificial intelligence? (Multiple choice)
Algorithm
Computing power
Data
Scene
1. Which of the following is not a specific technology of artificial intelligence?
Riemann geometry
1. Which of the following statements about overflow and underflow in numerical calculations is correct?
(Multiple choice)
Underflow occurs when the number close to zero is rounded to zero
Overflow occurs when a large number of levels are approximated as ~ or -~
1. Which of the following description about Python regular expression is not correct?
re regular expressions can only process string data, not numeric data.
1. In which year was the first introduction of “artificial intelligence”?

©1956
2. Which of the following description about the relationship between the Poisson distribution and the binomial
distribution is correct?
Poisson distribution is an approximation of the binomial distribution when n is very large and p is small.
1. Python tuples are identified by “()” and internal elements are separated by “:”.
False
int (integer type)
long (long integer type)
float (floating point type)
complex (complex number type)
1. There are a lot of data generated during the training of the neural network. What mechanism does
placeholder
1. What is the English abbreviation for Artificial Intelligence?
Artificial Intelligence
1. There are many commercial applications for machine learning services. What are the main business
• Financial product recommendation
• Predictive maintenance
• Telecom customer retention
• Retailer grouping

• Python modules can define functions, classes, and variables, but the module does not contain executable
code.
3. Which of the following options is not the session mode used by TensorFlow?
• Multiple POST queries
4. Principal Component Analysis (PCA) is a statistical method. A set of variables that may be related to each
other is transformed into a set of linearly related variables by orthogonal transformation. The converted set
of variables is called the principal component.
• False
5. The history of robots really began.

According to the development process of the robot, it is usually divided into three generations, respectively
are: (Multiple Choice)
Teaching Reproduction Robot
Robot with sensation
Intelligent robot
Beautiful
Explicit
Simple
1. What are the service categories included in Huawei Cloud El Enterprise Intelligence? (Multiple Choice)
El visual cognition
El speech semantics
El industry scene
VGGNet
ResNet
AlexNet
GoogleNet
The larger the standard deviation, the steeper the normal distribution curve.
1. In May 1997, the famous “Human-Machine Wars” final computer defeated Kasparov, the world chess king,
with a total score of 3.5 to 2.5. Is this computer called?
Dark blue
1. What is wrong description of backpropagation?
The backpropagation phase sends training inputs to the network to obtain an stimuli response.
1. Which of the following options is not a reason for traditional machine learning algorithms to promote the
development of deep learning?
Feature Engineering
1. The commonly used functions for mathematical operations in Python are basically in the math module and
the cmath module.
True
1. What dose not belong to supervised learning?
Principal component analysis
function.
False
1. What is the subject of artificial intelligence?
Comprehensive interdisciplinary and marginal disciplines
Answer: False
The biggest difference between logic regression and perceptron is that logistic regression uses the Sigmoid
activation function at the output layer, whereas perceptron uses the sign function. In addition, neither of the two
models can resolve the XOR problem.
Answer: True
Answer: True
Answer: False
Answer: False
Answer: True
The HiAI mobile computing platform architecture supports various mainstream front-end frameworks, such as
TensorFlow and Caffe.
Answer: True
Answer: True
Answer: False
Answer: True
In a medical image recognition task, we usually are more concerned about the recall ratio than the precision
ratio when evaluating model performance.
Answer: True
Answer: False
Similar to hidden layer, the quantity of neurons at both an input layer and an output layer of a neural network is a
hyperparameter that can be adjusted.
Answer: False
Answer: True
Answer: False
Answer: True
Answer: False
Answer: True
Answer: False
Answer: False
Factory A produces 60% of an automotive company’s products, and factory B accounts for 40%. The defective
rate of factory A is 1%, and that of factory B is 3%. If a vehicle is defective, what is the possibility that the vehicle
is produced by factory A?
Answer: B. 1/3
Input_shape = (4, 28, 28, 1)
= tf.random.normal(input_shape)
= tf.keras.layers.Conv2D(filters=2, kernel_size=3, strides=(1, 1), padding=’valid’, activation=’relu’,
print(y.shape)
Answer: B. (4,26,26,2)
Without Padding, use a 3X3 convolution kernel to convolve a 5X5 image with stride 1. Which of the following
options is the size of the output image?
Answer: A. 3X3
Answer: A. 1950
Answer: D. Da Vinci
In the process of training the neural network, which of the following methods is used to update the parameters
so as to minimize the loss function?
Answer: D. Several
import numpy as np
Answer: A. (2,3,2,2)
Which of the following options is the artificial intelligence development framework launched by Huawei?
Answer: C. XOR
Which of the following models is not suitable for the scenario where the covariance between data features is
large?
Answer: D. Atlas
A company wants to use the HiAI platform to develop its own APP. Which of the following IDEs can be used in
conjunction with the HiAI platform?
Answer:
Sigmoid
Softmax
Softplus
ReLu
Answer:
C. If peephole is used in LSTM, the vector of the output and memory unit of the previous time sequence are
concatenated with the input vector.
D. In LSTM, if the dimension of the input vector is 200 and the dimension of the hidden layer is 100, and the
peephole is not considered, the parameter quantity at this layer is calculated as follows: 4x ((100 + 200) X 100 +
100)
Answer: A,B,C
Convenient for users
Modular and combinable
Easy to expand
Answer: B & D

Answer: B,C,D
C. The training set includes a large amount of dirty data D. The number of training sets is small
Answer: C & D
matplotlib
cv2
Answer: A, B, D
HiAI Service
HiAI Foundation
D. HiAI Engine
Answer:
Algorithms
Computing power
Data
Scenarios
Answer:
Operator-level automatic differentiation
Automatic parallelization
Automatic operator generation
Semi-automatic data labeling
Which of the following are the problems you may encounter when training a generative adversarial network?
Answer:
An overview learning rate leads to model divergence
The discriminator is too well trained at the beginning of training. As a result, the generator cannot be trained.
The mode collapse results in insufficiently diversified images.
The number of data sets is small, causing model overfitting
Answer:
Chip system control CPU
AI computing engine
Multi-level System-On-a-Chip cache or buffer
Digital vision pre-processing module
Answer:
Atlas 200DK
Ascend 310
Ascend 910
Answer:
Smart speaker
Smart customer service
Autonomous driving
Alpha Go
Answer:
Keras is a high-level API of TensorFlow
TensorFlow 1.0 uses static graphs
TensorFlow 2.0 uses Eager Execution by default
Compared with the Sigmoid activation function, which of the following options are the advantages of ReLU?
Answer:
D. Save calculation
Answer:
TensorFlow
PyTorch
MXNet
MindSpore
1. Backpropagation makes gradient descent more efficient by means of dynamic programming.

O True
2. In some image and voice fields, Al performs better than humans, so artificial intelligence has entered the
stage of strong artificial intelligence.
O False
3. Automatic graph segmentation is one of the key technologies of MindSpore.

O True
4, One of the disadvantages of deep learning is that the interpretability is very weak, but itis still possible to
visualize the model using methods such as Local Interpretable Model-Agnostic Explanations (LIME).
True
5, Atlas series chips need the support of CUDA software package like other GPU accelerator cards.
False
6, The speed of grid search must be faster than random search.

False
7. In a medical image recognition task, we usually are more concerned about the recall ratio when evaluating
model performance.
O True
8. Voice recognition refers to the recognition of text data as audio data.
O False
1. In convolutional neural networks, the convolutional layer can reduce the dimension of feature maps.
O True
2. Laplace smoothing can be used to address the “zero probability” problem in the naive Bayes model.
True
1. Knowledge graph can be applied to social networking, gene sequencing, and risk control.
O True
12, Softmax can't be used for multiple classification problems.
O False
2. Da Vinci architecture is the core of the Ascend Al computing engine and Al processor.
O True
14, Ina feedforward network, neurons in the same layer are connected to each other.
False
15, The generalization error will continue to decrease as the complexity of the model increases.
False
16, If the metrics parameter is not specified in Model.compile, no loss value will be output.
False
17, Polynomial regression cannot resolve the nonlinear problem.

False
18, The biggest difference between logistic regression and perceptron is that logistic regression uses the
Sigmoid activation function at the output layer, whereas perceptron uses the sign function. In addition, neither of
the two models can resolve the XOR problem.
True
19, The HiAl mobile computing platform architecture supports various mainstream front-end frameworks, such
as TensorFlow and Caffe.
True
20, To install TensorFlow of the GPU version, you need to install CUDA and cuDNN.
True
In a feedforward network, neurons in the same layer are connected to each other. Answer: False
Answer: True
Answer: True
Answer: False
Similar to hidden layer, the quantity of neurons at both an input layer and an output layer of a neural network is a
hyperparameter that can be adjusted.
Answer: False
Cifar10 is a built-in data set of Keras. Answer: True
The generalization error will continue to decrease as the complexity of the model increases. Answer: False
Answer: True

Answer: False
PyTorch supports GPU acceleration. Answer: True
In the KNN model, a small K value may cause model underfitting. Answer: False
Malformed data is not dirty data. Answer: False
Backpropagation makes gradient descent more efficient by means of dynamic programming. 0 True
In some image and voice fields. Al performs better than humans, so artificial intelligence has entered the stage
of strong artificial intelligence.
Correct Answer: False
, Automatic graph segmentation is one of the key technologies of MindSpore.
Correct Answer = True
One of the disadvantages of deep learning is that the interpretability is very weak, but it is still possible to
visualize the model using methods such as Local Interpretable Model-Agnostic Explanations (LIME)
Correct Answer: True
Atlas series chips need the support of CUDA software package like other GPU accelerator cards_
Voice recognition refers to the recognition of text data as audio data

In convolutional neural networks, the convolutional layer can reduce the dimension of feature maps
Laplace smoothing can be used to address the 'zero probability' problem in the naive Bayes model
Knowledge graph can be applied to social networking, gene sequencing, and risk control.
Softmax can't be used for multiple classification problems.
Da Vinci architecture is the core of the Ascend Al computing engine and Al processor.
in a feeclforward network, neurons in the same layer are connected to each other.
If the metrics parameter is not specified in Model.compile, no loss value will be output.

The biggest difference between logistic regression and perceptron is that logistic regression uses the Sigmoid
activation function at the output layer, whereas perceptron uses the sign function. In addition, neither of the two
models can resolve the XOR problem.
The HiAI mobile computing platform architecture supports various mainstream front-end frameworks, such as
TensorFlow and Gaffe,
. To install TensorF low of the GPU version, you need to install CUDA and cuDNN.
Similar to a hidden layer, the quantity of neurons at both an input layer and an output layer of a neural network is
a hyperpararmeter that can be adjusted.
, Cifar10 is a built-in data set of Keras.
Correct Answer False
Malformed data is not dirty data

PyTorch supports GPLJ acceleration

The Adam optimizer in TensorFlow 2.0 is the built-in optimizer of the keras module.
Jobs that are innovative and have high requirements on social capabilities cannot be easily replaced by Al.
Logical reasoning is a part of Al.

Scikit-learn is a free software machine learning library for Python. It has powerful tools for data analysis and
mining.
HifLens can help small and micro companies realize smart access control.
Al is the core research area of machine learning.


, Graph Engine Service (GES) depends on structured data and facilitates querying and analysis of graph-
structure data based on various relationships.
The trace operation returns the sum of the diagonal elements of the matrix. So the trace the same as that of its
transposed matrix,
Deep learning can be used to easily export simple mathematical functions from a large amount of multi-
dimensional data to describe complicated decision-making interfaces.
Currently, Al is suitable for scenarios with a known environment, clear targets, and predictable actions. Deep
learning deals with image recognition, speech recognition, and translation, with Al having the same or in some
cases, better recognition abilities than human beings. These abilities are applied in many scenarios like medic
treatment and public safety but are still weak in reasoning and cognition.
DataTensorFlow is represented by tensors and computational graphs.
TensorFlow is Google's first-generation proprietary machine learning system.

Convolutional neural networks are more suitable for speech recognition problems.
Currently, Al is suitable for scenarios with a known environment, clear targets, and prediaabl
Deep learning deals with image recognition, speech recognition, and translation, with Al having the same or in
some cases, better recognition abilities than human beings. These abilities are applied in many scenarios like
medical treatment and public safety but are still weak in reasoning and cognition.
Python lists can be delimited by square brackets ([ ]), with the first element at index 0
Time intervals in Python are floating point numbers in units of second

Convolutional neural networks (CNNs) are better than recurrent neural networks(RNN)
Python lists can be delimited by square brackets ([]), with the first element at index 0
Recurrent neural networks are more suitable for image recognition problems
For scalar k and matrix A, the scalar multiplication of a matrix means each entry in the matrix A is multiplied by
the given scalar k.
"print" in Python 3 must be enclosed in parentheses (( )).

Operations and computation graph of TensorFlow do not necessarily run in the Session
The TensorFlow framework does not support the Windows OS

Python tuples are enclosed in parentheses (( )) and the internal elements are separated by semicolons (;).
Al is a new science of researching and developing theories, methods, and systems to imitate and enhance
human intelligence. It is one of the core research fields of machine learning.
Huawei believes that the value of Al is reflected in addressing enterprise business problems. Regarding the
development of AI, guided by industry solutions, Huawei focuses on basic technologies and the enabling
platform, and works with industry leaders to develop industry solutions.
Matrix A and matrix B can be added only if they have the same number of rows and columns.
Unlike machine learning, deep learning does not have unsupervised algorithms.
True
Unlike machine
continue statements in Python are used for jumping out the current loop and entering the next loop.
TensorFlow supports multiple TPU clusters computing : True

GAN is a deep learning model. Recently, it is one of the most promising methods for unsupervised learning in
complex distribution : False
Common mathematical functions in Python are stored in the math and cmath modules: False
Principal components Analysis (PCA) is a statistical method. It uses orthogonal transformation to convert a set
of possibly correlated variables into a set of values of linearly variables called principal components: True
Comments In Python start with a hash # mark in the beginning of each line: True
Deep learning algorithms can be categorized into supervised learning or unsupervised learning: False
Convolutional neural networks are more suitable for speech : True
The Python author designed the restrictive syntax to filter out bad programming habits : True
1. Grid search randomly selects hyperparameter combinations

false
2. The Tensor Boost Engine (TBE) is a processor enablement layer built by

Huawei for deep neutral networks and Ascend AI Processors. False
3. With the continued development of AI processors and AI frameworks, new data types, such as BFloat16,
are emerging
True
4. Assuming a dataset contains the areas and prices of 21,613 housing

units in a city, a classification model can predict the prices of other
housing units in the city
False
5. Poisson distribution is generally regarded ass the common distribution

of errors in linear regression. False
6. Distributed training is a feature of the TensorFlow framework. TRUE
7. When collection images for image processing, we usually use numbers

(such as 0,1,2,3 and 4) to indicate image types. False
8. Although each industry and application faces unique challenges, they

are all gradually adopting
True
9) In TensorFlow 2.0, tf.data contains methods to write data
True
10) The softplus function has a continuous derivative and defines a
smooth curved surface. True
11) Convolutional neural networks cannot be used to process text data
false
12) In convolutional neural networks, the size of each kernel is not
necessarily the same as that of a pooling layer window, False
13) MindSpore supports automatic graph splitting, which splits graphs
false
14) In the Convolutional layer, the dropout ratio of features whose values
are set to 0.Generally, the dropout ratio ranges from 0.2 to 0.5
true
15) Ascend AI Processors are a series of Huawei-proprietary AI
processors that feature superior computing power
True
16) The PIL module in TensorFlow 2.0 is used to perform operations on
images. False
1. MindSpore supports Ascend processors and CPUs, but does not

support GPUs. False
2. The softmax function turns a vector of K real numbers into another

vector of K real numbers, False
3. One of the filter methods is using the Chi-Squared test to analyze the
mapping between features and labels to implement features and labels
to implement feature selection. True
4. The Relu function is often used in deep learning and neutral

networks:
B the valuse range is [0, +)
5. Which operation is not a step in the network defenition process

during application development:
D model compression
6. Which of the following is covered by HUAWEI CLOUD EI Health :
B Genome
7. Which of the following is the reasoning method of the production

system that draws conclusions through a rule library:
A Froward
8. Which of the following HUAWEI CLOUD is NOT a basic AI platform
service :
Relational Database Service (RDS)
25) A vendor wants to provide an intelligent EMR system for a hospital:
Neutral language processing
26) Which of the following statements is true about variance and bias:
A model with high bias and variance performs poorly and will be used. 27) A generative adversarial is like a
game system. The generator
produces fake samples, and the discriminator tries to distinguish:
The discriminator cannot distinguish real data from the data created by
the generator. 28) Which of the following is the correct shape of tensor [[0,1],]:
(3,2)
29) Math operator in MindSpore :
Mul
30) True about about optimizers:
The SGD and momentum optimizers
31) Is true about model validity:
The goal of machine learning is for a tr
32) Which of the following is NOT an application of speech processing:
Named entity recognition
33) Which of the following is the purpose of running ocr_client = HWOcrClientAKSK(ak, sk, region) when using
HUAWEI CLOUD OCR
Initialize the client

34) Which of the following statements is true about the low-latency
design of CPUs. The powerful ALU
35) "Deep learning is a complex, comprehensive discipline. AI includes deep learning and machine learning
36) When the v1 compatibilty package of TensorFlow 2.0 is used to
inherit the TensorFlow 1.x code, the eager operations needs to be
disabled. Tf.disable_eager_execution()
37) Which of the following is NOT a main function of HUAWEI CLOUD
GEOGenius
Real-time traffic detection
38) Which of the following statements are true about grid search based
on hyperparameter tuning ?
NOT grid search suits neutral networks well
39) Which of the following are the advantages of the softsign activation
function over the sigmoid function ?
Gradient sat
The output
40) Which of the following methods provided by TensorFlow 2.0 cannot
be used to check whether an object is a tensor ?
Not is-tensor
41) Which of the following statements are true about ensemble learning ?
The voting
A batch of features
42) Are true about the three main schools main schools of AI ?
Behaviorism (zouz)
43) Which of the following statements are true about gradient descent ?
Not batch gradient descent BGD
44) Which of the following cells are predefined in MindSpore ?
MAELoss
cos
45) Are true about using OCR to train an invoice recognition model
lkol
46) Which of the following statements are true about hidden layers
The more hidden layers, Excessive hidden layers may cause overfitting
47) Which of the following statements are false about the Gated
Recurrent Unit (GRU)?
GRU is a variant
GRU combines
48) In TensorFlow 2.0, which of the following cannot be used to create
an all-zero tensor ?
Tf.zeros_frame() / tuple()
49) Which of the following functions are supported by HiAI Engine
platform?
lkol
50) Which of the following are common application scenarios of HUAWEI
CLOUD OCR?
Not photo recog
51) Common application scenarios of HUAWEI CLOUD OCR?
Not behavioral theory
52) WhAT are the theories of cognitive intelligence?
Ila the existence
53) Are true about activation functions?
The tanh function
During training
54) Which of the following statements are true about common
activations in deep learning?
Callback;
55) In MindSpore,if you want to define a function for collecting
information after each step is complete, you need to inhert the () class. 2
56) Linear regression using a loss function with the L()-norm regular term
is also called ridge regression
2
57) Generally, a mode with a () volume has a higher precision, and a
model
Large
small
58) HUAWEI HiAI built an open smart ecosystem based on the "
Cloud
59) In MindSpore Python APIs, mindspore.() provides various neural
network layers. NN
60) At a convolution layer, there are 128 3 x 3 kernels and the size of an
input feature map is 28 x 28 x 64 .
1, Boosting independently builds multiple basic learners and then averages their predictions.
0 True
False
Congratulations! Correct Answer: False
2, In TensorFlow 2.0, tf.data contains methods to write data in TFRecord format.
True
0 False
3, An Al processor is also referred to as an Al accelerator. It is a functional module used to process a large
quantity of computing tasks in an Al application. Ascend 910 is a
typical product.
e True
False
Congratulat ion s!Correct Answer: True
0True
0 F al s e
significantly improvi ng performance.
True
CD False
6, OCR automatically extracts key information about invoices to
help employees automatically fill in reimbursement forms. In
addition, the RPA automatic robot can improve the efficiency of
financial reimbursements.
0 True
False
Sorry ! Correct Answer: True
7, MindSpore can be quickly deployed on the cloud and edge,
improving resource utilization and privacy protection, and
helping developers focus on Al apps. Due to the size of the
model, MindSpore cannot be deployed on mobile phones.
0True
0 False
Sorry ! Correct Answer: False
8, Poisson distribution is generally regarded as the common
distribution of errors in linear regression.
0 True
0 False
9, Grid search randomly selects hyperpararneter combinations.
Therefore, it has a high efficiency and accuracy.
True
0 False
10, The tree model used at the bottom layer of the random forest
and GBDT algorithms is only the classification and regression
tree (CAST).
True
False
11, Powered by Ascend Al Processors, the Huawei Atlas Al
computing platform delivers products that come in various
forms, such as modules, cards, edge stations, servers, and
clusters, to build Al solutions for all scenarios across the
device, edge, and processor.
• True
• False
12, TensorFlow is Microsoft's second-generation opensource software library built for digital computing.
True
0 False
13, Creative work, such as writing, will not be easily replicated by
Al.
0 True
0 False
14, Although each industry and application faces unique
challenges, they are all gradually adopting Al-based
0 Tru e
0 False
15 TensorFlow 2.x is compatible with all TensorFlow 1.x modules.
0True
0 False
16, Object Storage Service (OBS) can be widely used in scenarios
with abundant relational data such as social apps, enterprise
relationship analysis, knowledge graph, risk control, and
recommendations.
0True
0False
17, In actual data processing, there are multiple trained models
available so that you can choose the optimal one.
0 True
0 False
18, One of the filter methods is using the Chi-Squared test to
analyze the mapping between features and labels to
implement feature selection.
0 True
0 False
True
False
20, Mini-batch gradient descent (MB D) cannot balance the
robustness of stochastic gradient descent and the efficiency of
batch gradient descent, and we may risk getting stuck at local
minima. It is not commonly used in practice.
True
e False
21, Which of the following is the correct shape of tensor [[0,11r
[213], [4-,5]]?
A. (23)
• B. (32)
.(11
6)
• D. (6,1)
Sorry! Correct Answer: B
22, Which of the following statements is true about neural
networks?
A. The full name of LSTM is Less Short-Term Memory.
• B. LSTM has nothing to do with RNINIs.
• C. LSTM is suitable for processing events with long interval and dE
in the time sequence.
0 ID. Neural networks can only propagate information forward.
Sorry ! Correct Answer: C
Congratulations! Correct Answer: D
23, Which of the following statements is true about the chip
enablernent layer in the Ascend Al software stack?
A. The Ascend Al software stack encapsulates the framework invoking
and offline model generation capabilities.
B. The Ascend Al software stack is the hardware compute basis of
Ascend Al Processors.
0 C. The Ascend Al software stack is nearest to the underlying compute
resources and outputs operator-layer tasks to the hardware.
D. The Ascend Al software stack can invoke an appropriate deep
learning framework based on the features of related algorithms.
Congratulations! Correct Answer: C
24, Which of the following statements is false about the ReLU
function?
A. The ReLLI function is not differentiable at x = 0 and a derivative is
forcibly defined at this point.
B. The ReLLI function effectively alleviates the vanishing gradient
problem.
C. The surface defined at the zero point of the ReLU function is not
smooth enough in some regression problems.
D. Compared with Sigmoid and tanh, the convergence of the ReLLJ
function is slow.
25 Which of the following is the type of labels predicted by
ensemble learning algorithms?
CI A. Discrete and continuous
B. Continuous
C. Discrete
Congratulations! Correct Answer: A
26, Which of the following HUAWEI CLOUD services is NOT a
basic Al platform service?
A. ModelArts
B. Huawei HiLens
C. Relational Database Service (RD)
D. Graph Engine Service (GES)
27, Dirty data refers to data with quality problems. Which of
the following statements is fake about the data quality?
0 B. Incompleteness: Data lacks attributes or contains some
missing values.
C. Unprocessed: Data for which feature engineering has not
been performed.
28, Which of the following statements is true about rulebased approac hes?
A. Rules are obtained through machine learning.
0 B Rule-based approaches use explicit programming to solve
problems.
0 C Rule-based approaches use samples for training.
D. Decision-making rules are complex or difficult to
describe. Congratulations! Correct Answer: B
29, Which of the following is a math operator in IvlindSpore?
A. Mul
B. ControlDepend
C. Callback
D. Select
.30, Intelligent quality inspection is based on the cloud-edgedevice synergy of deep learning algorithms. Which
of the
following operations is performed on edge devices?
C) A. Model training
B. Data labeling
C. Model delivery
• D. Onsite inference
31, A generative adversarial network is like a game system. The
generator produces fake samples, and the discriminator tries
to distinguish real data from the data created by the
generator. What is the ideal result?
A The generator is trained to fool the discriminator.
B. The discriminator cannot distinguish real data from the data created
by the generator.
C. The discriminator distinguishes real data from the data created by
the generator,
D. Develop a high-precision discriminator and a generator that cannot
fool the discriminator,
Sorry ! Correct Answer: B
32, Which of the following statement is false about the running
process of the MinciArrnour subsystem?
A. Configuration policies: Define test policies based on threat vectors
and trustworthiness certification requirements and select appropriate
test data generation methods.
B. Fuzzing execution: Generate trusted test data randomly based on
the model coverage and configuration policies.
C. Evaluation report: Generate an evaluation report based on built-in
or user-defined trustworthiness metrics.
• D. Trustworthiness enhancement: Use preset methods to enhance the
trustworthiness of Al models.
Congratulations! Correct Answer: B
.33, Which one of the following actions can the function tf,squeeze
be used for in TensorFlow 2.0?
0 C. Tensor concatenation
• D. Dimensionality reduction
34, In the Da Vinci Architecture, which of the following
computation data types is supported by the vector unit?
A. INT16
B. BFloat16
0 C. FP 32
D. BFloat32
35, Which of the following is a lightweight and high-performance
service module that helps MindSpore developers efficiently
deploy online inference services in production environments?
A. rvlindIR
B. NilindSpore Serving
C. MindArrnour
D. Mindlnsight
36, Which of the following is NOT an application of speech
processing?
• A. Speech recognition
C) B. Text to speech
C. Voiceprint recognition
D. Named entity recognition
37, Which of the following statements is true about variance and
bias?
0 A. A model with low bias but high variance has high robustness.
C. A model with low bias and variance has low precision.
D. A model with high bias and variance performs poorly and will not
be used.
Sorry ! Correct Answer: BCD
38, Which of the following is covered by HUAWEI CLOUD El
Health?
• A. Genome
• B. Smart surgery
• C. Patient care
• D. Doctor-patient dispute resolution
Sorry ! Correct ArmArer: A
.39, Which of the following statements are true about activation
functions?
Ef A. The existence of the activation function introduces linearity into the
network.
111 B. If we do not use an activation function, the output signals will be
just a simple linear function.
El C. There are many activation functions. You need to select one based
on the actual situation.
El D. Activation functions play an important role in learning and
understanding complex and nonlinear functions of neural network
models.
40, Which of the following development modes are supported by
the ModelArts training platform?
...I' A. In-cloud development (CodeLab)
...le B. In-cloud development (Notebook-ESDK)
▪ C. On-premises development (PyChanin+PyCharm Toolkit)
...I,' D. On-premises development (IDE+SDK)
congratulations! Correct Answer: ABCD
41, Which of the following statements are true about hidden
layers?
..tie A. The more hidden layers, the stronger the identification capability.
E4 B. The more hidden layers, the better.
E4 C. Excessive hidden layers may cause overfitting.
E4 D. Insufficient hidden layers may cause overfitting.
Sorry! Correct Answer: AC
42, In TensorFlow 2.0, which of the following cannot be used to
create an all-zero tensor?
A. tf.zeros0
B. iffill([1],0)
C. tf.zeros frame°
❑ D. tfieros tupleo
Sorry ! Correct Answer: CD
43, Which of the following statements are true about MindSpore
and Huawei all-scenario solutions?
0 A. The full -scenario deployment solution includes model generation
and efficient execution.
.
7
/ B. find pore Serving supports hatching.
C. Distributed training of ultra-large models and ultra-large datasets
requires only data parallelism.
D. I'VlindIR is the key to "Train Once, Deploy Everywhere."
Sorry ! Correct Answer: ABD
44, Which of the following statements are true about gradient
descent?
A. Stochastic gradient descent (SGD) randomly chooses samples for
each training job, causing unstability. As a result, the loss function
fluctuates or even encounters reverse displacements during the
process of dropping to the minimum value.
B. Batch gradient descent (BGD) is the most unstable method and
consumes too many compute resources.
C. Mini-batch gradient descent (MBGD) is a balance between BCD and
SGD, and is the optimal choice for all datasets.
Sorry ! Correct Answer: ACD
45, Which of the following are the advantages of the softsign
activation function over the sigrnoid function?
111 A. The maximum value of a derivative is 0.5.
El D. It is an unsaturated activation function.
Congratulations! Correct Answer: BC
0 A. Like TensorFlow, Keras is a multi-layer neural network
development
package. However, Keras has simpler syntax and is easier to use.
B. Keras is a neural network development package used to
build CNN sequential models. It cannot be used to build other
neural networks.
C. The neural network model built using Keras must
be compiled before data can be input into it for training.
0 D. Keras can be used as the backend of
TensorFlow. Sorry ! Correct Answer: ED
47, Which of the following statements are true about
common activation functions in deep learning?
U A. The sigmoicl function is monotonic continuous, and easy to
derive. Its output is bounded. making the network converge
better.
LI B. The sigrnoid function can easily cause the gradient to explode.
&'C. The tank function is symmetric with respect to the origin,
and the mean of its output is closer to 0.
D. During training of a deep neural network, the
sigmoid, tanhr and softsign functions cannot prevent the
vanishing gradient problem.
Congratulations! Correct Answer: CD
48, Which of the following are part of Mind pore?
0 A. Mind Expression API layer
B. Mind Compiler's real-time graph-level compilation subsystem
C.CUDA and third-party chips
D.!AndData
Sorry ! Correct AnsiArer: AB D
49, Which of the following statements are false about the
combination of the model bias and variance?
A.A good model requires a high variance and high bias.
B.A high bias and high variance may result in overfitting.
C.A high bias and low variance may result in underfitting.
&' D. A model with a high variance is of no application value.
Sorry ! Correct Answer: ,BBD
Sorry ! Correct Answer: ABC
50, Which of the following methods provided by TensorFlow 2.0
cannot be used to check whether an object is a tensor?
A. is tensor
Ef B. isrnstance
El C. device
III D. tftypes
Sorry! Correct Answer: BCD
51, Which of the following statements are true about linear
regression?
A. The error of linear regression is affected by many
independent factors. According to the central limit theorem (LT), the
error follows normal distribution.
B. The loss function of linear regression can be obtained using
the normal distribution function and maximum likelihood
estimation (M LE).
111 C. The result of multivariate linear regression analysis may be a
hyperplane in a high-dimensional space.
El D. Due to algorithm complexity, linear regression cannot use the
gradient descent method to calculate the weight parameter if the loss
function reaches the minimum value.
52, Which of the following are common tensor operations in
MindSpore?
A. dtype()
B. astype(
) *7 C. switch°
EI D. abso
53, Which of the following statements are true about data
conversion?
El A. Image data processing, such as color space conversion, grayscale
image conversion, and geometric conversion, is a part of data
co nversion.
Li B. In machine learning, features are often normalized to ensure that
the value ranges of different variables input to the same model are the
same.
C. When the data dimension is high, the existing variables need to
be combined or converted to generate new features, such as
averages.
Sorry ! Correct Answer: AB
KieD. All data needs to be converted.
54, Which of the following products can be equipped with Ascend
310 Processors?
ID A. Atlas 200 DK Al developer kit
ID B. Atlas 500 Al edge station
❑ C. Atlas 200 Al accelerator
module D. Atlas 300 Al accelerator
card Sorry ! Correct Answer: ABCD
55, As the neural network increasingly deepens, two issues may be
encountered during network training: one is 0, and the other is grad
vanishing/exploding.
lank items (1) is incorrect. Correct Answer: (1) Network degradation;network
egradation;Network Degradation
56, The Atlas 200 DK is designed to run on the Ascend Al Processor. (
using Arabic numerals.)
Sorry ! Blank items (1) is incorrect. Correct Answer: (1) 310
57 Tensor R[2,3]]] is a/an 0-dimensional tensor. (Fill in the blank with a
number.)
(1) Please fill in the answer here
Sorry! Blank items (1) is incorrect. Correct Answer: (1) 3
58, In an Adagrad optimizer, the initial value of the cumulative square
gradient is Q. (Fill in the blank with a number)
Sorry! Blank items (1) is incorrect. Correct Answer: (1) 0
59, At a convolution layer, there are 256 5 x 5 convolution kernels, the si an
input feature map is .32 x 32 x 5, the convolution step is 2, and ze padding
is employed(padding size=i). So, the size of the output feamap is * () *
0. (Fill in each blank with a number.)
SorBlemk items (1) and (2) and (3) are incorrect. Correct Answer: (1) 15 (2) 15 (3) 25(
60, In TensorFlow 2.0, if ttkeras.layers.RNIN is used to process timing
information and you want to obtain the output status at each momE set
0 to True.
Sorry! Blank items (1) is incorrect. Correct Answer: (1) return_sequence
E.
1, Boosting independently builds multiple basic learners and then averages their predictions.
0 True
False
2, In TensorFlow 2.0, tf.data contains methods to write data in TFRecord format.
True
0 False
3, An Al processor is also referred to as an Al accelerator. It is a functional module used to process a large
quantity of computing tasks in an Al application. Ascend 910 is a typical product.
True
False
Congratulat ion s!Correct Answer: True
0 True 0 False
significantly improving performance.
True
CD False
6, OCR automatically extracts key information about invoices to help employees automatically fill in
reimbursement forms. In addition, the RPA automatic robot can improve the efficiency of financial
reimbursements.
0 True
False
Sorry ! Correct Answer: True
protection, and helping developers focus on Al apps. Due to the size of the model, MindSpore cannot be
deployed on mobile phones.
0 True 0 False
0 True
0 False
Grid search randomly selects hyperpararneter combinations. Therefore, it has a high efficiency and accuracy.
True 0 False
The tree model used at the bottom layer of the random forest and GBDT algorithms is only the classification and
regression tree (CAST).
True
False
11, Powered by Ascend Al Processors, the Huawei Atlas Al computing platform delivers products that come in
various
forms, such as modules, cards, edge stations, servers, and clusters, to build Al solutions for all scenarios across
the device, edge, and processor.
True
False
12, TensorFlow is Microsoft's second-generation open-source software library built for digital computing.
True 0 False
13, Creative work, such as writing, will not be easily replicated by Al.
0 True 0 False
14, Although each industry and application faces unique challenges, they are all gradually adopting Al-based
0 Tru e
0 False
15 TensorFlow 2.x is compatible with all TensorFlow 1.x modules.

0 True 0 False
Object Storage Service (OBS) can be widely used in scenarios with abundant relational data such as social
apps, enterprise relationship analysis, knowledge graph, risk control, and recommendations.
0 True 0 False
In actual data processing, there are multiple trained models available so that you can choose the optimal one.
0 True
0 False
18, One of the filter methods is using the Chi-Squared test to analyze the mapping between features and labels
to implement feature selection.
0 True
0 False
True
False
20, Mini-batch gradient descent (MB D) cannot balance the robustness of stochastic gradient descent and the
practice.
True
False
21, Which of the following is the correct shape of tensor [[0,11r [213], [4-,5]]?
A. (23)
B. (32)
.(116)
D. (6,1)
22, Which of the following statements is true about neural networks?
A. The full name of LSTM is Less Short-Term Memory.
B. LSTM has nothing to do with RNINIs.
C. LSTM is suitable for processing events with long interval and dE in the time sequence.
0 ID. Neural networks can only propagate information forward.
Congratulations! Correct Answer: D

23, Which of the following statements is true about the chip enablernent layer in the Ascend Al software stack?
A. The Ascend Al software stack encapsulates the framework invoking and offline model generation capabilities.
B. The Ascend Al software stack is the hardware compute basis of Ascend Al Processors.
0 C. The Ascend Al software stack is nearest to the underlying compute resources and outputs operator-layer
tasks to the hardware.
D. The Ascend Al software stack can invoke an appropriate deep learning framework based on the features of
related algorithms.
A. The ReLLI function is not differentiable at x = 0 and a derivative is forcibly defined at this point.
B. The ReLLI function effectively alleviates the vanishing gradient problem.
problems.
D. Compared with Sigmoid and tanh, the convergence of the ReLLJ function is slow.
25 Which of the following is the type of labels predicted by ensemble learning algorithms?
CI A. Discrete and continuous
B. Continuous
C. Discrete
26, Which of the following HUAWEI CLOUD services is NOT a basic Al platform service?
A. ModelArts
B. Huawei HiLens
C. Relational Database Service (RD)
D. Graph Engine Service (GES) Congratulations! Correct Answer: C
27, Dirty data refers to data with quality problems. Which of the following statements is fake about the data
quality?
0 B. Incompleteness: Data lacks attributes or contains some missing values.
C. Unprocessed: Data for which feature engineering has not been performed.
28, Which of the following statements is true about rule-based approac hes?
A. Rules are obtained through machine learning.
Rule-based approaches use samples for training.
D. Decision-making rules are complex or difficult to describe. Congratulations! Correct Answer: B
29, Which of the following is a math operator in IvlindSpore?

A. Mul
B. ControlDepend
C. Callback
D. Select
.30, Intelligent quality inspection is based on the cloud-edgedevice synergy of deep learning algorithms. Which
of the following operations is performed on edge devices?
C) A. Model training
B. Data labeling
C. Model delivery
D. Onsite inference
31, A generative adversarial network is like a game system. The generator produces fake samples, and the
discriminator tries to distinguish real data from the data created by the generator. What is the ideal result?
A The generator is trained to fool the discriminator.
B. The discriminator cannot distinguish real data from the data created by the generator.
C. The discriminator distinguishes real data from the data created by the generator,
D. Develop a high-precision discriminator and a generator that cannot fool the discriminator,
32, Which of the following statement is false about the running process of the MinciArrnour subsystem?
A. Configuration policies: Define test policies based on threat vectors and trustworthiness certification
requirements and select appropriate test data generation methods.
B. Fuzzing execution: Generate trusted test data randomly based on the model coverage and configuration
policies.
C. Evaluation report: Generate an evaluation report based on built-in or user-defined trustworthiness metrics.
D. Trustworthiness enhancement: Use preset methods to enhance the trustworthiness of Al models.
Congratulations! Correct Answer: B
.33, Which one of the following actions can the function tf,squeeze be used for in TensorFlow 2.0?
0 C. Tensor concatenation
D. Dimensionality reduction
34, In the Da Vinci Architecture, which of the following

computation data types is supported by the vector unit?
A. INT16
B. BFloat16
0 C. FP 32
D. BFloat32
35, Which of the following is a lightweight and high-performance service module that helps MindSpore
developers efficiently deploy online inference services in production environments?
A. rvlindIR
B. NilindSpore Serving
C. MindArrnour
D. Mindlnsight
36, Which of the following is NOT an application of speech processing?

C) B. Text to speech
C. Voiceprint recognition
D. Named entity recognition
37, Which of the following statements is true about variance and bias?
0 A. A model with low bias but high variance has high robustness.
C. A model with low bias and variance has low precision.
D. A model with high bias and variance performs poorly and will not be used.

38, Which of the following is covered by HUAWEI CLOUD El Health?
A. Genome
B. Smart surgery
C. Patient care
D. Doctor-patient dispute resolution
Sorry ! Correct ArmArer: A
.39, Which of the following statements are true about activation functions?
Ef A. The existence of the activation function introduces linearity into the network.
111 B. If we do not use an activation function, the output signals will be just a simple linear function.
El C. There are many activation functions. You need to select one based on the actual situation.
El D. Activation functions play an important role in learning and understanding complex and nonlinear functions
of neural network models.
40, Which of the following development modes are supported by the ModelArts training platform?
...I' A. In-cloud development (CodeLab)
...le B. In-cloud development (Notebook-ESDK)
C. On-premises development (PyChanin+PyCharm Toolkit) ...I,' D. On-premises development (IDE+SDK)
congratulations! Correct Answer: ABCD
41, Which of the following statements are true about hidden layers?
..tie A. The more hidden layers, the stronger the identification capability. E4 B. The more hidden layers, the
better.
E4 C. Excessive hidden layers may cause overfitting.
E4 D. Insufficient hidden layers may cause overfitting.
Sorry ! Correct Answer: AC
42, In TensorFlow 2.0, which of the following cannot be used to create an all-zero tensor?
A. tf.zeros0
B. iffill([1],0)
C. tf.zeros frame°
D. tfieros tupleo
Sorry ! Correct Answer: CD
43, Which of the following statements are true about MindSpore and Huawei all-scenario solutions?
0 A. The full-scenario deployment solution includes model generation and efficient execution.
.7/ B. find pore Serving supports hatching.
Distributed training of ultra-large models and ultra-large datasets requires only data parallelism.
I'VlindIR is the key to "Train Once, Deploy Everywhere." Sorry ! Correct Answer: ABD
44, Which of the following statements are true about gradient descent?
Stochastic gradient descent (SGD) randomly chooses samples for each training job, causing unstability. As a
result, the loss function fluctuates or even encounters reverse displacements during the process of dropping to
the minimum value.
Batch gradient descent (BGD) is the most unstable method and consumes too many compute resources.
Mini-batch gradient descent (MBGD) is a balance between BCD and SGD, and is the optimal choice for all
datasets.
BGD uses all samples for each training job.
Sorry ! Correct Answer: ACD
45, Which of the following are the advantages of the softsign activation function over the sigrnoid function?
111 A. The maximum value of a derivative is 0.5.
Gradient saturation slows down.
The output is zero-centered.
El D. It is an unsaturated activation function.
Congratulations! Correct Answer: BC
0 A. Like TensorFlow, Keras is a multi-layer neural network development
package. However, Keras has simpler syntax and is easier to use.
Keras is a neural network development package used to build CNN sequential models. It cannot be used to
build other neural networks.
The neural network model built using Keras must be compiled before data can be input into it for training.
0 D. Keras can be used as the backend of TensorFlow. Sorry ! Correct Answer: ED
47, Which of the following statements are true about common activation functions in deep learning?
U A. The sigmoicl function is monotonic continuous, and easy to derive. Its output is bounded. making the
network converge better.
LI B. The sigrnoid function can easily cause the gradient to explode.
&' C. The tank function is symmetric with respect to the origin, and the mean of its output is closer to 0.
During training of a deep neural network, the sigmoid, tanhr and softsign functions cannot prevent the vanishing
gradient problem.
Congratulations! Correct Answer: CD
48, Which of the following are part of Mind pore?

0 A. Mind Expression API layer
Mind Compiler's real-time graph-level compilation subsystem
CUDA and third-party chips
!AndData
Sorry ! Correct AnsiArer: AB D
49, Which of the following statements are false about the combination of the model bias and variance?
A good model requires a high variance and high bias.
A high bias and high variance may result in overfitting.
A high bias and low variance may result in underfitting. &' D. A model with a high variance is of no application
value.
Sorry ! Correct Answer: ABC

50, Which of the following methods provided by TensorFlow 2.0
cannot be used to check whether an object is a tensor?
A. is tensor
Ef B. isrnstance
El C. device
III D. tftypes
51, Which of the following statements are true about linear regression?
The error of linear regression is affected by many independent factors. According to the central limit theorem
(LT), the error follows normal distribution.
The loss function of linear regression can be obtained using the normal distribution function and maximum
likelihood estimation (M LE).
111 C. The result of multivariate linear regression analysis may be a hyperplane in a high-dimensional space.
El D. Due to algorithm complexity, linear regression cannot use the gradient descent method to calculate the
weight parameter if the loss function reaches the minimum value.
52, Which of the following are common tensor operations in MindSpore?

dtype()
astype() *7 C. switch° EI D. abso
53, Which of the following statements are true about data conversion?
El A. Image data processing, such as color space conversion, grayscale
image conversion, and geometric conversion, is a part of data
co nversion.
Li B. In machine learning, features are often normalized to ensure that the value ranges of different variables
input to the same model are the same.
When the data dimension is high, the existing variables need to be combined or converted to generate new
features, such as averages.
Sorry ! Correct Answer: AB
Kie D. All data needs to be converted.
54, Which of the following products can be equipped with Ascend 310 Processors?
ID A. Atlas 200 DK Al developer kit ID B. Atlas 500 Al edge station
❑ C. Atlas 200 Al accelerator module D. Atlas 300 Al accelerator card Sorry ! Correct Answer: ABCD
55, As the neural network increasingly deepens, two issues may be encountered during network training: one is
0, and the other is grad vanishing/exploding.
lank items (1) is incorrect. Correct Answer: (1) Network degradation;network egradation;Network Degradation
56, The Atlas 200 DK is designed to run on the Ascend Al Processor. ( using Arabic numerals.)
57 Tensor R[2,3]]] is a/an 0-dimensional tensor. (Fill in the blank with a number.)
58, In an Adagrad optimizer, the initial value of the cumulative square gradient is Q. (Fill in the blank with a
number)
59, At a convolution layer, there are 256 5 x 5 convolution kernels, the si an input feature map is .32 x 32 x 5,
the convolution step is 2, and ze padding is employed(padding size=i). So, the size of the output feamap is * () *
0. (Fill in each blank with a number.)
Please fill in the answer here
SorBlemk items (1) and (2) and (3) are incorrect. Correct Answer: (1) 15 (2) 15 (3) 25(
60, In TensorFlow 2.0, if ttkeras.layers.RNIN is used to process timing information and you want to obtain the
output status at each momE set 0 to True.
Sorry ! Blank items (1) is incorrect. Correct Answer: (1) return_sequence

Answer: False
The biggest difference between logic regression and perceptron is that logistic regression uses the
Sigmoid activation function at the output layer, whereas perceptron uses the sign function. In addition,
neither of the two models can resolve the XOR problem.
Answer: True
Answer: True
Answer: False
Answer: False
Answer: True
The HiAI mobile computing platform architecture supports various mainstream front-end frameworks,
such as TensorFlow and Caffe.
Answer: True
Answer: True
Answer: False
Answer: True
In a medical image recognition task, we usually are more concerned about the recall ratio than the
precision ratio when evaluating model performance.
Answer: True
Answer: False
Similar to hidden layer, the quantity of neurons at both an input layer and an output layer of a neural
network is a hyperparameter that can be adjusted.
Answer: False
Answer: True
Answer: False
Answer: True
Answer: False
Answer: True
Answer: False
Answer: False
Factory A produces 60% of an automotive company’s products, and factory B accounts for 40%. The
defective rate of factory A is 1%, and that of factory B is 3%. If a vehicle is defective, what is the
possibility that the vehicle is produced by factory A?
Answer: B. 1/3
Input_shape = (4, 28, 28, 1)
X = tf.random.normal(input_shape)
Y = tf.keras.layers.Conv2D(filters=2, kernel_size=3, strides=(1, 1), padding=’valid’, activation=’relu’,
print(y.shape)
Answer: B. (4,26,26,2)
Without Padding, use a 3X3 convolution kernel to convolve a 5X5 image with stride 1. Which of the
following options is the size of the output image?
Answer: A. 3X3
Answer: A. 1950
Answer: D. Da Vinci
In the process of training the neural network, which of the following methods is used to update the
parameters so as to minimize the loss function?
Answer: D. Several
import numpy as np
Answer: A. (2,3,2,2)
Which of the following options is the artificial intelligence development framework launched by
Huawei?
Answer: C. XOR
Which of the following models is not suitable for the scenario where the covariance between data
features is large?
Answer: D. Atlas
A company wants to use the HiAI platform to develop its own APP. Which of the following IDEs can be
used in conjunction with the HiAI platform?
Answer:
A. Sigmoid
B. Softmax
C. Softplus
D. ReLu
Answer:
C. If peephole is used in LSTM, the vector of the output and memory unit of the previous time
sequence are concatenated with the input vector.
D. In LSTM, if the dimension of the input vector is 200 and the dimension of the hidden layer is 100,
and the peephole is not considered, the parameter quantity at this layer is calculated as follows: 4x
((100 + 200) X 100 + 100)
Answer: A,B,C
A. Convenient for users
B. Modular and combinable
C. Easy to expand
Answer: B & D
Answer: B,C,D
C. The training set includes a large amount of dirty data
D. The number of training sets is small
Answer: C & D
C. matplotlib
D. cv2
Answer: A, B, D
A. HiAI Service
B. HiAI Foundation
D. HiAI Engine
Answer:
A. Algorithms
B. Computing power
C. Data
D. Scenarios
Answer:
A. Operator-level automatic differentiation
B. Automatic parallelization
C. Automatic operator generation
D. Semi-automatic data labeling
Which of the following are the problems you may encounter when training a generative adversarial
network?
Answer:
A. An overview learning rate leads to model divergence
B. The discriminator is too well trained at the beginning of training. As a result, the generator
cannot be trained.
C. The mode collapse results in insufficiently diversified images.
D. The number of data sets is small, causing model overfitting
Answer:
A. Chip system control CPU
B. AI computing engine
C. Multi-level System-On-a-Chip cache or buffer
D. Digital vision pre-processing module
Answer:
A. Atlas 200DK
B. Ascend 310
C. Ascend 910
Answer:
A. Smart speaker
B. Smart customer service
C. Autonomous driving
D. Alpha Go
Answer:
A. Keras is a high-level API of TensorFlow
B. TensorFlow 1.0 uses static graphs
C. TensorFlow 2.0 uses Eager Execution by default
Compared with the Sigmoid activation function, which of the following options are the advantages of
ReLU?
Answer:
D. Save calculation
Answer:
A. TensorFlow
B. PyTorch
C. MXNet
D. MindSpore

1 Ken Radh3touh Its Ok U Have Bilgacem For Further Assistance If You Dont Support Bilgacem Timchi T

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1 Ken Radh3touh Its Ok U Have Bilgacem For Further Assistance If You Dont Support Bilgacem Timchi T

Uploaded by

Copyright:

Available Formats

1) ken radh3touh its ok u have bilgacem

for further assistance ,if you dont support

A. Intrusion recognition scheme

B. Vehicle identification scheme

C. Crowd statistics program

D. Policy query scheme based on knowledge graph

D. Basic platform services

A. Robot with sensation

B. Teaching Reproduction Robot

C. Robots that will think

Loss function and model function are the same thing.

D. Write a file format reader

A. Avoid local extremes

C. Speed up algorithm convergence

Self-encoder is an unsupervised learning algorithm

What research areas does artificial intelligence have? (Multiple choice)

B. Natural language processing

D. Analysis and positioning task

B. The activation function is partly a nonlinear function, partly a linear function

Which of the following are AI Application areas?

A. New programming language

A scalar k is multiplied by matrix A equal to k and each of the numbers in A is multiplied.

A. Quick source code

D. Diverse tool chain

The following belongs to TensorFlow2.0 The characteristic is?

A. Introduce Keras interface

B. Support more platforms and more languages

C. Continue to be compatible Tensorflowl.x Related modules

D. Support static image mechanism

Where is the main foothold of symbolism?

B. The number of rows is less than the number of columns

B. use Relu Activation function

C. Higher order function

A. It has been installed Cmake Compilation tool

B. It has been installed CUDA Software package

C. Complete environment configuration

D. carry out Atlas Driver Installation

Which of the following products is related to artificial intelligence? (Multiple Choice)

B. Huawei Mate mobile phone

B. words Representative text block recognition result

D. The incoming image data needs to go through base64 coding

B. Increase the amount of data

D. The results of each trial are independent of each other.

B. Exponential loss function

B. Know the actual problem

C. Computer calculation results

E. Numerical calculation method

C. Logically wrong value

Which of the following statements about supervised learning is correct?

B. Supervised learning cannot use cross-validation for training

D. Decision tree is a supervised learning

HUAWEI HiAI What are the supported models?

A. Strong artificial intelligence

B. Convolutional Neural Network

Tensorflow supports multi-TPU cluster computing

C. Backpropagation can only be used in feedforward neural networks

A. Forward propagation algorithm

Which description is wrong about the hyperparameter?

B. The value of the hyperparameter is not learned by the algorithm itself.

D. Most machine learning algorithms have hyperparameters.

TensorFlow It is one of the most popular deep learning frameworks.

Which of the following is not an application of image recognition services?

A. The entire training set will be divided into 6 Share