Banksoal Digabungkan

1. Which of the following are sub-fields of AI?
----------------------------------- ABCD
A. Machine learning
B. Computer Vision
C. Speech recognition
D. Natural language processing
(Judgment) 2. Voice recognition refers to recognizing audio data as text data. ------------------------- A
A.TRUE
B.FALSE
(Single choice) 3. The three stages of artificial intelligence include computational intelligence, () and
cognitive intelligence. ----------------- B
A. Weak artificial intelligence
B. Perceptual Intelligence
C. Behavioral Intelligence
D. Strong artificial intelligence
(Single choice) 4. Which of the following options does not belong to the scope of Huawei's full-stack
solution? ------------------------ B
A. Application enable
B. Edge computing
C. Open source framework
D. Chip enable
(Multiple choice) 5. What are the elements of artificial intelligence? -------------------------------------- - ABCD
A. Algorithm
B. Scene
C. Computing power
D. Data
(Judgment) 6. In the neural network based on connectionism, each node can express a specific meaning.
--------- B
A.TRUE
B.FALSE
9. Which of the following aspects belong to Huawei's full-stack AI solution?----------------------- ABCD

A. Ascend
B. CANN
C. ModelArts
D. MindSpore
(Single choice) 7. According to the Society of Automotive Engineers (SAE), autonomous driving is divided
into
A.L1~L4
B.L1~L5
C.L0~L4
D.L0~L5
(Judgment) 8. Computer vision is the science of studying how to make computers "see". ---------------------
----A
(Judgment) 10. Jobs that are highly repetitive and require weak social skills are the jobs most easily
replaced by AI. ------------ A
A.TRUE
B.FALSE
(Judgment) 11.Huawei's AI scenarios include public cloud, private cloud, various edge computing, IoT
industry terminals, and consumer
The deployment environment of terminal, edge, and cloud such as fee-based terminals. ----------------------
----------- A
A.TRUE
B.FALSE
(Single choice) 12. What is the performance of artificial intelligence in the perceptual intelligence stage?
--------------------------- C
A. Machines begin to understand, think and make decisions like humans

B. Machines begin to calculate and transmit information like humans
C. The machine starts to understand and understand, make judgments, and take some
simple actions
(Judgment) 13. Under the premise of ensuring data privacy and security, federated learning uses
different data sources to cooperatively train models to make progress and break through the bottleneck
of data.
------------------------------------------------------------A
(Single choice) 14. Where is the foothold of symbolism? --------------------------------B
A. The foothold is in neuron networks and deep learning.

B. The foothold is in reasoning, symbolic reasoning and machine reasoning.
C. The foothold is perception and action.
D. The foothold is behavior control, adaptive and evolutionary computing.
(Judgment) 15. The current stage of artificial intelligence is still in the weak artificial intelligence stage. --
------------------------- A
16. Which of the following are the application areas of AI? -------------------------------ABCD
A. Wisdom Education
B. Smart City
C. Smart Home
D. Smart medical
(Judgment) 17. Artificial intelligence is the research and development of theories, methods and
application systems for simulating, extending and expanding human intelligence Of a new technological
science. --------------------------------------------- A
(Single choice) 18. Which of the following schools does neural network research belong to? -----------------
-------------------------------------------B
A. Symbolism
B. Connectionism
C. Behaviorism
D. None of the above
19. What are the main technical applications of artificial intelligence? ----------------------------------------------
------------ACD
A. Natural language processing

B. Control system
C. Computer Vision
D. Speech recognition
(Judgment) 20. Huawei's AI full scenarios include public cloud, private cloud, various edge computing,
IoT industry terminals, and consumer terminals and other end, edge, and cloud deployment
environments.
------------------------------------------------------------A
2. Overview of machine learning
(Judgment) 2. A model composed of machine learning algorithms cannot represent the true data
distribution function on a theoretical level. ------------------------------------------------ A
(Single choice) 3. Among the machine learning algorithms, which of the following are not integrated
learning strategies? ---------------------D
A.Boosting
B. Stacking
C. Bagging
D. Marking
(Judgment) 4. The loss function of logistic regression is the cross-entropy loss function. -----------------------
------- A
A.TRUE
B.FALSE
(Judgment) 5. Regular term can also be added to logistic regression to avoid overfitting. ----------------------
A
A.TRUE
B.FALSE
6. What is the fitting surface of linear regression in more than 3 dimensions? ----------------------------- C
A. Surface
B. Plane
C. Hyperplane
D. Hypersurface
(Judgment) 7. K-fold cross-validation refers to dividing the test data set into K sub-data sets. ----------------
---- B
A.TRUE
B.FALSE
(Judgment) 8. Grid search is a method of parameter adjustment. ---------------------------------- B
A.TRUE
B.FALSE
(Single choice) 9. Which of the following statements about supervised learning is correct? -------------------
--------- A
A. Decision tree is a supervised learning
B. Supervised learning cannot use cross-validation for training
C. Supervised learning is a rule-based algorithm
D. Supervised learning can be trained without labels
(Multiple choice) 10. When dealing with actual problems, when should machine learning be used? ACD
A. The data distribution itself changes over time and requires continuous re-adaptation of the
program, such as predicting the trend of merchandise sales.
B. The complexity of the rules is low and the problem is small.
C. The rules of the task will change over time, such as the production line. Defect detection
on
D. The rules are very complicated or cannot be described, such as face recognition and
speech recognition
11. Regularization is an important and effective technique for reducing generalization errors in
traditional machine learning. The following techniques are regular The technology is: -------------
-------------------------------- ---ABC
A. L1 regularization
B. L2 regularization
C.Dropout
D. Momentum optimizer
12. Which of the following is correct about the difference between machine learning algorithms
and traditional rule-based methods? -------- ABC
A. Traditional rule-based methods, in which the rules can be explicitly clarified manually
B. Traditional rule-based methods use explicit programming to solve problems
C. The mapping relationship of the model in machine learning is automatically learned
D. The mapping relationship of the model in the machine learning institute must be implicit
(Single choice) 13. In Bagging integrated learning, what is the relationship between each base
learner? -------------------- B
A. Additive relationship
B. Correlation
C. The latter model must be built on top of the previous model
D. Independent
(Single choice) 14. After the data has completed the feature engineering operation, in the process
of constructing the model, which of the following options is not a step in the decision tree
construction process?
-------------------------------------------------- --------- C
A. Pruning
B. Feature selection
C. Data cleaning
D. Decision tree generation
15. What are the commonly used kernel functions in SVM? --------------------------------- ABCD
A. Gaussian kernel function
B. Polynomial kernel function
C. Sigmiod kernel function
D. Linear kernel function
(Single choice) 16. In the gradient descent algorithm, which algorithm is the most confusing
trajectory on the loss function surface? -----A
A. SGD
B.BGD
C.MGD
D.MBGD
(Judgment) 17. In the two-classification process, we can set any category as a positive example. -
------------------------ A
A.TRUE
B.FALSE
(Judgment) 18. When we describe housing, we often use the attributes of residential area, house
type, decoration type, etc. If we use plain
With Bayes as a model, we assume that there is no relationship between attributes. -----------------
-A
(Judgment) 19. Loss function and model function are the same thing. ---------------------------------
---B
(Single choice) 20. "From historical data with labels, predict whether the retail sales of the next
quarter will be 200,000-300,000 or 300,000-400,000",
(Judgment) 27. In polynomial regression, there is a square term in the formula of the model, so it
is not linear. ----------- B
A.TRUE
B.FALSE
28. What are the common types of dirty data? ---------------------------------- ABCD
A. Malformed value
B. Duplicate value
C. logically wrong value
D. Missing values
(Single choice) 29. What is the most important difference between batch gradient descent, small
batch gradient descent, and stochastic gradient descent? ------ D
A. Gradient size
B. Gradient direction
C. Learning rate
D. Number of samples used
(Single choice) 30. Which of the following evaluation indicators belong to the regression
algorithm? ---------------------------------- C
A. Recall rate
B. Confusion matrix
C. Mean square error
D. Accuracy
(Single choice) 31. Which of the following descriptions of machine learning is correct? -----------
---------------------- A
A. Deep learning is a branch of machine learning
B. Deep learning and machine learning are mutually inclusive
C. Deep learning and machine learning are both artificial intelligence but have no relationship
with each other
(Multiple choice) 32. Which of the following descriptions of the k value in the KNN algorithm is
correct?--------------------------- BC
A. The larger the K value, the easier the model is to overfit
B. The larger the K value, the smoother the segmentation surface of the classification
C. K value is a hyperparameter
D. The value of k can be set to 0
is not linear.
-------------------------------------------------- ---------- B
A.TRUE
B.FALSE
(Judgment) 34. Add the L1 regular term to the loss function of linear regression. At this time, the
regression is called Lasso regression.
-------------------------------------------------- ----------- A
(Single choice) 35. What is the label predicted by the regression algorithm? -------------------------
------------ ------D
A. Self-deformation
B. Discrete
C. Strain type
D. Continuous
(Single choice) 36. Global gradient descent algorithm, stochastic gradient descent algorithm and
batch gradient descent algorithm are all gradient descent algorithms. The following statements
about their advantages and disadvantages are wrong:
-------------------------------------------------- -----------C
A. The global gradient algorithm can find the minimum value of the loss function
B. Batch gradient algorithm can solve the local minimum problem
C. Stochastic gradient algorithm can find the minimum value of the loss function
D. The convergence process of the global gradient algorithm is time-consuming
37. In the random forest, what strategy does the final ensemble model use to determine the model
result?
-------------------------------------------------- ----------BC
A. Cumulative system
B. Find the average
C. Voting system
D. Multiplying system
(Judgment) 38. The test error will keep getting smaller as the complexity of the model increases.
-------------------------------------------------- --------- B
A.TRUE
B.FALSE
39. What are the common clustering algorithms?--------------------------------------- ABCD
A. Density clustering
B. Hierarchical clustering
C. Spectral clustering
D.Kmeans
(Single choice) 40. In supervised learning, which of the following models is used to describe "the
one who is close to the red and the one who is red to the black"?
-------------------------------------------------- ------------ C
A.K-Means
B. SVM
C.KNN
D. Neural network
(Judgment) 41. The naive Bayes algorithm does not require independent and identical
distribution among sample features. ------------------- B
A.TRUE
B.FALSE
Three, deep learning overview
1. In a convolutional neural network, different layers have different functions. Which of the
following layers can play a dimensionality reduction effect?
-------------------------------------------------- ----------------- BCD

A. Input layer
B. Fully connected layer
C. Convolutional layer
D. Pooling layer
2. Which of the following activation functions are prone to vanishing gradient problems? ---------
---------------- CD
A. ReLU
B.Softplus
C. Tanh
D. Sigmoid
(Single choice) 3. Which of the following statements about the standard RNN model is correct? -
---------------------------- D
A. There is no one-to-one model structure
B. The time direction is not considered when backpropagating
C. There is no many-to-many model structure
D. There will be a problem of attenuation of long-term transmission and memory information
(Single choice) 4. In the process of training the neural network, we use the gradient descent
method to continuously update which value, and then make the loss
Minimize function? ---------------------------------------------- -----D
A. Number of samples
B. Characteristic value
C. Hyperparameters
D. Parameters
(Single choice) 5. 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. Which of the following statements about feedforward neural networks are
correct:
-------------------------------------------------- ------------D
A. Neurons with computing power are connected to the upper and lower layers
B. Its input nodes have computing capabilities
C. Connecting neurons in the same layer
D. Information between layers is only transmitted in one direction
(Single choice) 6. In the neural network, which of the following methods are used to update the
parameters when training the network to minimize
The loss function? --------------------------------------------- ------D
A. Forward propagation algorithm
B. Pooling calculation
C. Convolution calculation
D. Backpropagation algorithm
(Single choice) 7. The optimizer is an important part of training neural networks. The purpose of
using the optimizer does not include which of the following:
-------------------------------------------------- --------------------C
A. Speed up algorithm convergence
B. Reduce the difficulty of setting manual parameters
C. Avoid overfitting
D. Avoid local extremes
(Single choice) 8. Can the perceptron be displayed in space? ----------------------------------- ----C
A. Line
B. Plane
C. Hyperplane
D. Point
(Single choice) 9. The loss function reflects the error between the target output of the neural
network and the actual output. The loss commonly used in deep learning
The loss function is: ---------------------------------------------- -------B
A. Exponential loss function
B. Mean square loss function
C. Log loss function
D. Hinge loss function
(Single choice) 10. There are many types of neural networks in deep learning. The following
neural network information is one-way propagation: ---------B
A. LSTM
B. Convolutional Neural Network
C. Recurrent Neural Network
D. GRU
(Single choice) 11. In the deep learning neural network, the perceptron is the simplest neural
network. The correct statement about its structure is:
-------------------------------------------------- --------------------B
A. Its hidden layer has only two layers
B. Its hidden layer has only one layer
C. The Sigmoid activation function is used in the network
D. Relu activation function is used in its network
12. Generative confrontation networks are currently widely used. Which of the following
scenarios can use this network?
-------------------------------------------------- -----------------ABCD
A. Data enhancement
B. Semantic segmentation
C. Information retrieval
D. Image generation
(Single choice) 13. Recurrent neural network is different from convolutional neural network.
Which of the following problems is it better at solving?
-------------------------------------------------- ---------A
A. Sequence Related Issues
B. Image classification
C. Image detection
D. Recommendation question
14. Weight sharing is common in neural networks. Which of the following neural networks will
share weights?
-------------------------------------------------- ---------BD
A. Perceptron
C. Fully connected neural network
D. Recurrent neural network
15. In deep learning tasks, when encountering data imbalance problems, which of the following
methods can we use to solve the problem?
-------------------------------------------------- --------- BCD

A. Batch delete
B. Random oversampling
C. Synthetic sampling
D. Random undersampling
(Single choice) 16. Regarding backpropagation, is the following statement wrong? -----------------
---------------- A
A. Backpropagation can only be used in feedforward neural networks
B. Backpropagation can be combined with gradient descent algorithm to update network weights
C. Backpropagation will go through the activation function
D. Back propagation refers to the back propagation of errors through the network
17. What are the commonly used loss functions in deep learning? ------------------------------------
BC
A. L1 loss function
B. Mean square error loss function
C. Cross entropy error loss function
D. Self-decreasing loss function
(Single choice) 18. The following applications that are not part of the adversarial generation
network? -------------------------------- C
A. Text generation
B. Image generation
C. Image recognition
D. Data enhancement
(Judgment) 19. All convolution kernels of the same convolution layer in the convolutional neural
network are weight-sharing. --------------- A
TRUE
FALSE
20. Knowing that the total number of parameters of a layer of the fully connected neural network
is 330, then the number of neurons in the previous layer and this layer
May be? ----------------------------------------------- -----BC
A.32 and 10
B.10 and 33
C.33 and 10
D. 9 and 33
(Single choice) 21. God will often encounter many problems when training the network. For the
problem of gradient disappearance, we can choose to use
Which of the following functions should be used to alleviate this problem? --------------------------
-------------A
A. Relu function
B. Sigmoid function
C.tanh function
D. Softsign function
(Single choice) 22. The number of hidden layers of the deep learning neural network has a
certain impact on the performance of the network. Which of the following statements about its
impact is correct: ------------------ ------------------------------------B
A. The number of hidden layers is appropriately reduced, and the resolution ability of the neural
network remains unchanged
B. Increase the number of hidden layers appropriately, the stronger the resolution ability
of the neural network
C. The number of hidden layers is appropriately reduced, the stronger the resolution ability of the
neural network is
D. The number of hidden layers increases appropriately, the weaker the resolution ability of the
neural network
23. In the process of deep learning model training, what are the common optimizers?
-------------------------------------------------- -------ABCD
A. Adam
B. Adagrad
C. SGD
D. Momentum
(Single choice) 24. In the classic convolutional neural network model, what hidden layer does
the Softmax function follow?
-------------------------------------------------- ----------C
A. Convolutional layer
B. Pooling layer
C. Fully connected layer
D. All of the above
(Single choice) 25. What is wrong about the following statement about cyclic neural network? ---
---------------------------B
A. Cycle The neural network can be expanded according to the time axis
B. LSTM cannot solve the problem of gradient disappearance
C. LSTM is also a recurrent neural network
D. Recurrent neural network can be abbreviated as RNN
(Single choice) 26. For the image classification problem, which of the following neural networks
is more suitable to solve this problem? ------------C
A. Perceptron
B. Recurrent neural network
C. Convolutional neural network
D .Fully connected neural network
(Single choice) 27. The activation function plays an important role in learning and understanding
very complex problems of neural network models. The following statement about the activation
function is correct.
-------------------------------------------------- ----------B
A. Activation functions are linear functions
B. Activation functions are non-linear functions
C. The activation function is partly a non-linear function and part is a linear function
D. Most of the activation functions are nonlinear functions, and a few are linear functions
(Single choice) 29. 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:
-------------------------------------------------- -----------C
A. Motherboard
B. RAM
C. GPU
D. CPU
30. Which of the following steps in deep learning are automatically completed by the model?
-------------------------------------------------- ------BD
A. Model training
C. Analysis and positioning task
D. Feature extraction
31. Is the following description correct about the convolutional neural network pooling layer?
-------------------------------------------------- ------ABCD
A. The pooling operation uses a scanning window to achieve
B. The pooling layer can play a role in dimensionality reduction
C. Commonly used pooling methods include maximum pooling and average pooling
D. After the characteristics of pooling Image becomes smaller
(Single choice) 32. 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
A. Chain rule
B. The rule of accumulation
C. Law of Equivalence
D. Rule of Normalization
(Single choice) 33. L1 and L2 regularization are methods commonly used in traditional machine
learning to reduce generalization errors. The following is about the two
The correct method is: --------------------------------------------- --------A
A. L1 regularization can be used for feature selection
B. Both L1 and L2 regularization can be used for feature selection
C. L2 regularization can do feature selection
D. Neither L1 nor L2 regularization can be used for feature selection
(Single choice) 34. Traditional machine learning and deep learning are the core technologies of
artificial intelligence, and there is a slight difference in the engineering process. The following
steps
What you don't need to do in deep learning is -----------------------------------B
A. Model evaluation
B. Feature Engineering
C. Data cleaning
D. Model construction
35. The pooling layer in the convolutional neural network can reduce the size of the lower layer
input. Common pooling are:
-------------------------------------------------- ----------CD
A. Minimal stratum
B. Product pooling layer
C. Maximum pooling layer
D. Average pooling layer
(Single choice) 36. 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?
-------------------------------------------------- ------------A
A. Back propagation update parameters
B. Forward propagation update dreams
C. Back propagation calculation results
D. Forward propagation calculation results
(Judgment) 37. The cyclic neural network can capture the dynamic information in the serialized
data.
-------------------------------------------------- -----------A
(Single choice) 38. Where should the labeled data be placed in the adversarial generative
network?
-------------------------------------------------- ----------B
A. As the output value of the generative model B. As the input value of the discriminant model
C. As the output value of the discriminant model D. As the input value of the generative model
39. If the deep learning neural network has the problem of gradient disappearance or gradient
explosion, our common solution is.
-------------------------------------------------- --------ACD
A. Gradient shear
B. Random undersampling
C. Use Relu activation function
D. Regularization
(Single choice) 40. Input a 32*32 image, use a 5*5 convolution kernel to do the convolution
calculation with a step size of 1, and output the image
The size is ------------------------------------------------ ------ B
A.28*23
B.28*28
C.29*29
D.23*23
(Single choice) 41. There are often many problems in the training of the gods network. For the
problem of gradient disappearance, we can choose which of the following functions to alleviate
the problem?
-------------------------------------------------- -----------B
42. Which of the following are the activation functions of the deep learning algorithm?
-------------------------------------------------- --------ABC
A. Sigmoid
B. ReLU
C. Tanh
D. Sin
(Single choice) 43. 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
A. Gradient descent
B. Dropout
C. Cross validation
D. Regularization
(Single choice) 44. What is wrong with what people say about neural networks?--------------------
------------- -A
A. With the increase in the number of hidden layers of the neural network, the classification
ability of the model gradually weakens
B. The limitation of the single-layer perceptron is that it cannot solve the XOR problem
C. The feedforward neural network can be represented by a directed acyclic graph
D. Front There is no connection between neurons in the same layer of the feed neural network
Fourth, the industry's mainstream development framework
(Single choice) 1. What methods can be used for tensor merging in TensorFlow 2.0? --------------
-----------B
A.join
B.concat
C.split
D.unstack
(Judgment) 2. The three main advantages of the Keras interface in TensorFlow 2.0 are user-
friendly, modular, and
Combination, easy to expand. -------------------------------------------------- A

A.TRUE
B.FALSE
(Judgment) 3.tf.keras.datasets You can view the built-in data sets in keras. ------------------ A
A.TRUE
B.FALSE
4. As of the end of 2019, what version of TensorFlow has released? ---------------------- ACD
A.1
B.3
C.2
D.0.1
(Judgment) 5. TensorFlow is an end-to-end open source platform for machine learning and deep
learning. ---------- A
A.TRUE
B.FALSE
(Judgment) 6. TensorFlow is one of the most popular deep learning frameworks. -------------------
---- A
A.TRUE
B.FALSE
7. The following code was used when compiling the model:
model.compile(optimizer='Adam,loss='categorical.crossentropy',
metrics=[tf.keras.metrics.accuracy]),
When using the evaluate method to evaluate the model, which of the following indicators will be
output?
-------------------------------------------------- -----------AC
A.accuracy
B.categorical_ 1oss
C.loss
D.categorical accuracy
(Single choice) 8. What methods are not included in the built-in indicators of keras.metrics of
TensorFlow 2.0? ----------C
A. Accuracy
B. Recall
C. Sum
D. Mean
(Judgment) 9. The graph and session mechanism is cancelled in TensorFlow 2.0. -------------------
--------B
(Single choice) 10. When building a neural network using the keras interface of TensorFlow 2.0,
the network needs to be compiled
Which of the following methods should be used to work? ---------------------------------------A
A.compile
B.write
C.join
D.fit
(Single choice) 11. The operators that are not supported by TensorFlow 2.0 among the following
options are ----------------------C
A.pow
B.@
C.^
D.//
(Single choice) 12. What data type does TensorFlow use? ------------------------------------C
A. Scalar
B.Vector
C.Tensor
D.Matrix
(Judgment) 13.TensorFlow 2.0 supports GPU acceleration mainly in the parallel operation of
matrix addition, subtraction, multiplication and division
accelerate. -------------------------------------------------- ------- A
A.TRUE
B.FALSE
14. What are the following features of TensorFlow 2.0? ---------------------------- ACD
A. Introduce Keras interface
B. Support static image mechanism
C. Support more platforms and more languages
D. Continue to be compatible with related modules of Tensorflowl.x
(Judgment) 15. PyTorch's current version supports Python2 and Python3? ---------------------- B
A.TRUE
B.FALSE
(Judgment) 16. TensorFlow2.0 does not support tensor transfer between GPU and CPU. ----------
-- B
A.TRUE
B.FALSE
(Single choice) 17. What are the following features that are not part of TensorFlow 2.0? ----------
------------------- A
A. Multi-core CPU acceleration
B. Distributed
C. Multilingual
D. Multi-platform
(Single choice) 18. What is the way to view the tensor dimension in TensorFlow 2.0? -------------
------------C
A.dimens
B.dtype
C.ndim
D.device
(Single choice) 19.Code model.fit(mnist.train.inmage,mnist.train.labels.epochs=5)
What does the epochs parameter represent? --------------------------------------------A
A. The entire training set will be trained 5 times
B. The entire test set will be tested 5 times
C. The entire training set will be divided into 6 parts
D. The entire training set will be divided into 5 parts
20. Which of the following options belong to the estimator package method under keras? ---------
--------ABC
A. Evaluation
B. Training
C. Forecast
D. Output model
(Single choice) 21. Among the following attributes, TensorFlow 2.0 does not support the method
to create tensor? ---------------C
A.zeros
B.fill
C.create
D.constant
(Single choice) 22. Which of the following functions does PyTorch have? ---------------------------
-------A
A. Embedded keras
B. Support dynamic graph
C. Automatic derivative
D. GPU acceleration
23. What is the built-in optimizer of the keras module in TensorFlow?----------------------ABCD
A. Adam
B. SGD
C. Adaboost
D. Adadelta
24. Is there a way to check whether it is a tensor in TensorF1ow2.0?

-------------------------------------------------- ---------BC
A.dtype
B.isinstance
C.is_tensor
D.device
(Single choice) 25. Which of the following companies developed TensorF1ow first?
-------------------------------------------------- ----------D
A. Oracle
B. Facebook
C. Nvidia
D. Google
(Single choice) 26.Which company did Pytorch first launch?

-------------------------------------------------- ----------C
A. Baidu
B. Google
C. Facebook
D. Huawei
(Single choice) 27.What is the role of keras.preprocessing of TensorFlow 2.0?
-------------------------------------------------- ---------A
A.keras data processing tool
B.keras built-in optimizer
C.keras model deployment tool
D.Keras generating model tool
(Judgment) 28. The tf.contrib method can continue to be used in TensorFlow 2.0.
-------------------------------------------------- ---------B
(Single choice) 29. The following options do not support TensorFLow2.0 for dimension
transformation.
-------------------------------------------------- ---------C
A.squeeze
B.reshape
C.gather
D.transpose
5. Huawei's AI development framework MindSpore

(Judgment) 1. Tensor is the data storage component in MindSpore. ----------------------------A
(Single choice) 2. Which of the following is not a commonly used Operation in MindSpore? -----
---------------------A
A.signal
B.math
D.nn
C.array
(Single choice) 3. On-Device execution, that is, the entire image is offloaded and executed, and
the computing power of the Yiteng chip can be fully utilized, which can greatly reduce the
interaction overhead, thereby increasing the accelerator occupancy rate. What is wrong about the
implementation of the following description of on-Device?
-------------------------------------------------- -------------C
A.MindSpore uses gradient data-driven adaptive graph optimization to achieve decentralized
autonomous A11 Reduce, Gradient aggregation is in the same step, and calculation and
communication are fully streamlined.
B. Challenges of model execution under super chip computing power:
Memory wall problems, high interaction overhead, and difficulty in data supply. Partly executed
on the Host and partly executed on the Device, the interaction overhead is even much greater
than the execution overhead, resulting in low accelerator occupancy
C. MindSpore uses chip-oriented depth map optimization technology to minimize
synchronization waiting and maximize "data computing The parallelism of “trust”, the training
performance is the same as that of the Host side graph scheduling method.
D. The challenge of distributed gradient aggregation under super-powered chip computing
power: ReslNet50 will be generated in a single iteration of 20ms.
The synchronization overhead of heart control and the communication overhead of frequent
synchronization. The traditional method requires 3 simultaneous completion of A11 Reduce, and
the data-driven method is autonomous A11 Reduce, without control overhead
(Single choice) 4. Which of the following is not a feature of MindSpore's core architecture?
-------------------------------------------------- -------------C
A. Automatic differentiation
B. Automatic tuning
C. Automatic coding
D. Automatic parallel
5. MSIR is a concise, efficient and flexible graph-based functional IR, a functional semantic type
that can be expressed
Have?------------------------------------------------ --------ABD
A. Free variables
B. Higher-order functions
C. Interrupt
D. Recursion
(Single choice) 6. Faced with the challenge of implementing efficient distributed training for
ultra-large-scale models, how does MindSpore handle it?
-------------------------------------------------- -------------A
A. Automatic parallel
B. Serial
C. Manual parallel
(Single choice) 7. Ce11 provides basic modules for defining and executing calculations. The
objects of Ce11 can be executed directly. What is wrong with the following statement? -----------
--------- ---------------------------------ACD
A.__init__, initialization parameter (Parameter), submodule (Ce11), Operators (Primitive)
and other groups
B.Construct, which performs initialization verification, defines the execution process. In the
graph mode, it will be compiled into a graph for execution, and there is no grammatical
restriction.
C. There are also some predefined optim commonly used optimizers, wrap commonly used
network packaging functions, etc. Cel1
D.bprop (optional), the reverse of custom module
(Single choice) 8. Facing the huge gap between industry research and full-scenario AI
applications, which of the following is not included in the technological innovation of
MindSpore to help inclusive AI across the application gap?
-------------------------------------------------- ------------B
A. New collaboration method
B. New programming language
C. New programming paradigm
D. New execution mode
(Single choice) 9. During application development, which of the following operations is not a
typical network definition?
-------------------------------------------------- -------------C
A. Network definition
B. Network execution
C. Knowledge distillation
D. Weight initialization
Six, Atlas artificial intelligence computing platform

1. What conditions need to be checked when compiling and running programs using a server
equipped with Atlas300 (3000) accelerator card?
-------------------------------------------------- -----------------ACD
A. Complete Atlas driver installation
B. CUDA software package installed
C. Cmake compilation tool is installed
D. Complete environment configuration
2. From the perspective of technical architecture, the classification of AI chips includes? ----------
-------------------ABCD
A. FPGA
B. CPU
C. GPU
D. ASIC
(Judgment) 3.Root users cannot install DDK (Device Device) on the server with Atlas300 (3000)
accelerator card.
Development Kit), can only be installed by ordinary users. ------------------------A
(Single choice) 4. What is the main application of the Yiteng 310 chip based on Atlas 200 DK? -
-------------------A
A. Model reasoning
B. Build the model
C. Training model
5. What are the main computing resources contained in the Da Vinci architecture computing
unit? ------------------------ABD
A. Vector calculation unit
B. Scalar calculation unit
C. Tensor calculation unit
D. Matrix calculation unit
(Judgment) 6. AI chips are also called AI accelerators, which are specially designed to handle a
large number of computing tasks in artificial intelligence applications
functional module. -------------------------------------------------- --- A
A.TRUE
B.FALSE
(Single choice) 7. Which of the following options is not included in the logical architecture of
the flatten AI processor? ---------------------- B
A. DVPP
B. GPU
C. AI calculation engine
D. Chip system control CPU
(Single choice) 8. Check if the Atlas300 (3000) accelerator card driver is installed successfully,
which command should be used? ------- A
A.npu-sim info
B.npu info
C.atlas-driver info
D.atlas info
(Judgment) 9. Install DDK (Device Development Kit) in the server with Atlas300 (3000)
accelerator card
Only Python3 needs to be installed when the environment is dependent. ------------------------------
------ A
A.TRUE
B.FALSE
10. AI chips are divided into business applications and can be divided into? -------------------------
---- AD
A. Training
B. GPU
C. Model building
D. Reasoning
(Single choice) 11.Which command can check whether the Atlas 300 (3000) accelerator card is
in place? ---------------- B
A.1spci | grep'npu'
B.1spci | grep'd100'
C.1spci grep'atlas'
D.atlas info
(Judgment) 12. GPU is good at computationally intensive and easy to parallel programs. ----------
-------------------- A
A.TRUE
B.FALSE
(Multiple choice) 13. Atlas 800 AI server has multiple models, which is based on the Kunpeng
processor platform? ------- B
A. Atlas 800 Model: 9000
B. Atlas 800 Model: 3000
C. Atlas 800 Model: 3010
(Single choice) 14. Which of the following options is the same as Shengteng 310 and Shengteng
910? -----------------------B
A. Maximum power
B. Architecture
C. Integer precision
D. 7nm process
(Judgment) 15. Huawei Shengteng AI chip is one of the typical representatives of NPU (Neural
Network Processor). ------------A
(Single choice) 16. What processor does Atlas use to accelerate AI inference?
-------------------------------------------------- -----------B
A. Isotron 910 processor
B. Yiteng 310 processor
C. GPU
D. FPGA
17. Which of the following libraries are installed in the server with Atlas300 (3000) accelerator
card DDK (Device Development Kit) which needs to be installed before?
-------------------------------------------------- -------ABC
A.decorator
B.setuptools
C.numpy
D.tensorflow
(Judgment) 18. The CPU mainly improves AI performance by increasing instructions and
increasing the number of cores.
-------------------------------------------------- ------------A
7. Huawei Smart Terminal AI Open Platform
(Single choice) 1. Which module does the face detection in the HUAWEI HiAI platform belong
to? -----------------------A
A. HiAI Engine
B.HiAI Framework
C.HiAI Foundation
D.HiAI Service
(Judgment) 2. HUAWEI HiAI Engine can easily integrate multiple AI capabilities with App. ----
---------A
(Judgment) 3. The HiAI mobile computing platform provides a wealth of upper-level functional
business APIs, which can run efficiently on mobile devices.
-------------------------------------------------- --------------------A
4. What models does HUAWEI HiAI support? --------------------------------ABC
A. P30
B. Mate 20
C. Honor V20
D. iPhone10
(Single choice) 5. Which tool does the HUAWEI HiAI platform support integration with? --------
---------------------C
A. Jupyter Notebook
B. MyEclipse
C. Android Studio
D. Spider
(Single choice) 6. Which HUAWEI HiAI module does Huawei chip support? -----------------------
--B
A. HiAI Engine
B.HiAI Foundation
C.HiAI Framework
D.HiAI Service
7. What is the value of HUAWEI HiAI empowering APP? -------------------------------ABCD

A. Security
B. Stable
C. Real time
D. Anytime
8. What are the advantages of the HiAI mobile computing platform?---------------------------------

ABCD
A. Quick start source code
B. Diverse tool chains
C. Complete documentation
D. Rich API
(Single choice) 9. Which of the following is a function of the HUAWEI HiAI Foundation
module?
-------------------------------------------------- ------------C
A. App integration
B. Let the service actively find users
C. Quickly convert and migrate existing models
D. Push the service at the right time and according to the needs of users
8. Huawei Cloud Enterprise Intelligent Application Platform

1. Which statement about the general form recognition service is correct? ------------------------
ABCD
A.rows represents the row information occupied by the text block, the number starts from
0, in the form of a list
B.colums represents the column information occupied by the text block, the number starts
from 0, in the form of a list
C. The incoming image data needs to be base64 encoded
D.words represents the result of text block recognition
(Judgment) 2. There can be multiple tags returned by the image tag service. -------------------------
------A
(Single choice) 3. When the voice recognition service is successfully called, which field is the
recognition result stored in?
-------------------------------------------------- ---------A
A.result
B.content
C.data
D.text
(Single choice) 4. When the universal text recognition service is successfully invoked, the
returned text recognition result is stored in which of the following fields?
-------------------------------------------------- ---------B
A.text
B.result
C.content
D.words
5. According to different industries and application scenarios, there are many kinds of HUAWEI
CLOUD EI agents. What are the currently available HUAWEI CLOUD EI agents?
-------------------------------------------------- ----------ABCD
A. Industrial intelligence
B. Traffic intelligence
C. Park intelligence
D. Automobile intelligence
6. Which of the following items are included in the results returned when the face search service
is successfully invoked?
-------------------------------------------------- ---------ABC
A. The similarity of the searched face
B. The searched face id
C. The searched face position
D. The searched face number
(Single choice) 7. The combination of ModelArts service and () service can easily deploy the
model to the "end"?
-------------------------------------------------- ------------D
A.OBS
B.OCR
C.ECS
D.HiLens
(Single choice) 8. Which of the following is not an application of image recognition service? ----
-------------------------D
A. Target detection
B. Smart photo album
C. Scene analysis
D. Speech synthesis
9. HUAWEI CLOUD ModelArts is a one-stop development platform for AI development.

Which of the following functions are in
Is it available on Mode1Arts? --------------------------------------- ABCD
A. Data governance
B. AI market
C. Visual workflow
D. Automatic learning
(Multiple choice) 10. Which of the following statements about the type field returned by the
universal form recognition service is correct?
-------------------------------------------------- -------- BD
A.type represents the type of text recognition area
B. When type is text, it represents the text recognition area
C.type represents the table type
D.type is table, it represents the table recognition area
(Judgment) 11. Only one image can be passed in when calling the face comparison service. ------
------------------- B
A.TRUE
B.FALSE
(Multiple choice) 12. HUAWEI CLOUD EI allows more enterprise boundaries to use AI and big
data services to accelerate business development and benefit society.
In which of the following aspects can Huawei Cloud EI service serve enterprises? ------------------
------ ABCD
A. Industry data
B. Industry wisdom
C. Algorithm
D. Computing power
(Multiple choice) 13. Which of the following services belong to the Huawei EI service family? --
-------------------------- ABCD
A. Dialogue robot
B.EI basic services
C. Natural language processing
D. EI big data service
(Multiple choice) 14. Which of the following capabilities does the gene knowledge map have? --
---------------------------- ACD
A. Auxiliary case diagnosis
B. Disease prediction and diagnosis
C. Gene test report generation
D. Entity query
15. Is the following statement about passport identification service correct? -------------------------
-ABCD
A.country_code represents the country code of the country where the passport was issued
B.nationality represents the nationality of the holder
C.passport_number represents the passport number
D.confidence The confidence information of related fields. The greater the confidence, the
higher the reliability of the corresponding field identified this time. In a statistical sense, the
greater the confidence, the higher the accuracy.
(Judgment) 16. In the face search service, if we want to delete a face set, we can use this code:
firs_cliet.get_v2().get_face_set_service().delete_face_set("**"),
Where "**" is to fill in the actual face set name. ------- ----------------------A
A. Target detection
C. Scene analysis
D. Speech synthesis
17. What happens to the label sugestion when the image remake detection service is called
successfully? -----------BCD
A.certainty
B.FALSE
C.uncertainty
D.TRUE
18. Which of the following are the topics of speech processing research? -----------------------------
---BCD
A. Voice processing
B. Voiceprint recognition
C. Speech recognition
D. Voice wake up
(Judgment) 19. The face detection service will report an error when there is no face in the test
image. ------------------------B
(Judgment) 20. The image data of the passport recognition service does not need to be base64
encoded. -------------------B
(Single choice) 21. What is wrong about the image content review service? --------------------------
---C
A.politics is the detection result of politically sensitive persons
B.terrorism is the test result of political violence and terrorism
C.confidence represents the degree of confidence, ranging from 0-100
D.labe1 represents the label of each test result
22. Regarding the suggestion field returned when the image content review service is called
successfully, which statement is correct?
-------------------------------------------------- ---------ABCD
A.pass means no sensitive information, pass
B.review represents the need for manual review
C.suggestion field represents whether the test passed
D.block means that it contains sensitive information and does not pass
23. According to Huawei Cloud EI Intelligent Platform, which of the following solutions can be
provided?
-------------------------------------------------- -------ABCD
A. People flow statistics plan
B. Policy query plan based on knowledge graph
C. Vehicle identification plan
D. Intrusion identification plan
(Single choice) 24. In the environment preparation for calling HUAWEI CLOUD services,
which of the following does not require confirmation of installation?
-------------------------------------------------- ----------D
A.requests
B.setuptools
C.websocket-client
D.Jieba
(Judgment) 25. 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
26. Which of the following options are need to be setting in Huawei Cloud
-------------------------------------------------- -------ABCD
A.ak
B.region
C.project_id
D.sk
27. Regarding the face search service, which of the following statements are correct?
-------------------------------------------------- -------ACD
A. When there is no face set, you need to create a face set first, and then add face data before you
can search B. The size of a face set cannot exceed 10,000 pictures C. There is a special interface
to delete a specified face set D. There is a dedicated interface to delete the face data in a certain
face set
(Single choice) 28. In HUAWEI CLOUD EI, which of the following can integrate AI technology
into the application scenarios of various industries and take advantage of the advantages of AI
technology, thereby improving efficiency and enhancing experience.
-------------------------------------------------- ----------A
A. EI agent
B. OBS Object Storage Service
C. Cloud database
(Single choice) 29. Can the traffic agent realize () hour all-time/global traffic perception?
-------------------------------------------------- ----------A
A.7×24
B.5×24
C.5×12
D.7×12
(Single choice) 30. An e-commerce company has always had problems with its official website
stalling, poor utilization of network resources, and difficulty in operation and maintenance. Then
which of the following EI services can the company use?
-------------------------------------------------- -----------C
A. Park agent
B. Traffic Agent
C. Network Agent
D. Industrial Agent
(Single choice) 31. Which of the following is not a face recognition technology?--------------------
-------------- D
A. Face search
B. Face comparison
C. Face detection
D. Remake recognition
(Single choice) 32. Which of the following is the URI of the image label service?
-------------------------------------------------- ---------D
A./v1.0/image/recognition
B./v1.0/image/celebrity-recognition
C./v1.0/image/recapture-detect
D./v1.0/image/tagging
(Single choice) 1. Feature is the dimension that describes the characteristics of the sample.
Regarding its interpretability in traditional machine learning and deep learning, which of the
following statements is correct: --------------- -------------------------------A
A. Features are interpretable in traditional machine learning, but weak in deep learning
B. Features are weak in traditional machine learning, but strong in deep learning.
C. Features are weak in traditional machine learning and deep learning.
D. Features are strong in traditional machine learning and deep learning.
What is the problem? -------------------------------------------- C
A. Regression problem
B. Rule issues
C. Classification issues
D. Clustering problem
(Single choice) 21. Compared with the random forest algorithm, which of the following
statements is wrong with the GBDT algorithm?
-------------------------------------------------- ---------- C
A. GBDT algorithm is easier to underfit than random forest
B. Random forest is calculated in parallel, while GBDT cannot
C. GBDT algorithm is easier to overfit than random forest
D. GBDT and random forest are both based on the CART tree
(Single choice) 22. Which of the following machine learning algorithms is not unsupervised
learning? ----------------------- B
A. GMM
B.Xgboost
C. Clustering
D. Association rules
(Judgment) 23. The training error will continue to decrease as the model complexity increases.
-------------------------------------------------- --------- A
A.TRUE
B.FALSE
(Single choice) 24. If a model has a large deviation on the test set and a small variance, then it
means the model?
-------------------------------------------------- --------- D
A. Overfitting
B. It may be overfitting or underfitting
C. Just fit
D. Underfitting
25. If there is a large amount of sales data but no labels, the company wants to identify VIP
customers. The following model is combined
Suitable? ---------------------------------------------- ------ CD
A. Logistic regression
B. SVM
C.K-Means
D. Hierarchical clustering
(Single choice) 26. In machine learning, what does the model need to input to train itself and
predict unknown? ---------------- D
A. Manual procedures
B. Neural Network
C. Training algorithm
D. Historical data
(Single choice) 28. Generating a confrontation network is like a game system. The generator
generates fake samples, and the discriminator judges whether it is true or false. Our ideal result
is?-------------- ------------------------C
A. The generator produces roughly the same samples
B. The discriminator efficiently distinguishes the true and false of the samples generated by the
generator
C. The discriminator cannot distinguish the true and false samples generated by the generator
D. The generator produces different samples
Write a custom catalog title here
Artificial Intelligence HCIAv2.0 Question Bank
Update date: 2020-6-23 Update content: Correct some wrong questions, the question bank score range
is 800-900
The question bank is expected to expire on October 1, 2020, please note
Q1. Which school does neural network research belong to?
A. Symbolism
B. Connectionism
C. Behaviorism
Answer: B
Q2. Smart speakers are one of the typical application scenarios for voice processing.
A.True
B.False
Answer: A
Q3. What are the main technical applications of artificial intelligence?
A. Control system
B. Speech recognition
C. Computer Vision

Answer: BCD
Q4. Which of the following is not a ModelArts development type?
A. Zero-based modeling
B. Agile development
C. Quick modeling
D. Standard model development
Answer: B
Q5. Which of the following are ModelArts development models?
A. Custom development
B. Iterative learning
C. Data annotation
Answer: ACD
Q6. Which of the following engines cannot be provided by HUAWEI HiAI Engine?
A. NLU engine
B. CV engine
C. ASR engine
D. DSP engine
Answer: D
Q7. When HiAI Foundation offline model calculation, the offline model is loaded from the file, and the
user’s input
Data (such as pictures) is copied and calculated on HiAI's NPU. When calculating, each inference only
Need to import and export user data from DDR to NPU once.
A.True
B.False
Answer: A
Q8. HiAI Foundation does not support sparse model acceleration.
A.True
B.False
Answer: B
Q9. Which of the following features does Python have?
A. Object-oriented
B. Open source
C. Simple
D. Explanatory
Answer:ABCD
Q10.
a=1
b=2
c = a if a>b else 0
What is the value of c?
A.2
B.0
C.None
D.1
Answer: B
Q11. Which of the following can generate a dictionary?
A.{{"C",3}:"c"}
B.{{"D":4}:"d"}
C.{[1, 2,3]:"a"}
D.{"b":["B’"]}
Answer: D
Q12. Which of the following output results are wrong?
A.sorted([(‘b’ ,2),(‘a’ ,1),(‘c’,3),(‘d’,4)], key=lambda x:x[1])
Output result: [( a’,1), (3),(‘d’,4)]
B.print (*filter (lambda x: x%2==1, [1, 2, 3]))
Output result: 1 3
C.print (map(lambda x:xx, [1,2,3]))
Output result: 1 4 9
D.print (*zip([1,2,3],["b"], ["A", "B","C"]))‘C’)
Output result: (1,’a’,""A";,"(2,b,‘B’)
Answer: AD
Q13. What is the result of 5/2 in Python3?
A.3
B.2
C.2.5
D.1
Answer: C
The data in Q14.json format is a dictionary in Python.
A.True
B.False
Answer: B
Q15. Which of the following fields can Python be used in?
A. Website development
B. Artificial Intelligence
C.APP
D. Data Science
Answer:ABCD
Q16. Which of the following identifiers are named incorrectly?
A.print
B.If
C._0a
D.a0*
Answer: AD
Q17. Concurrency refers to the simultaneous execution of multiple tasks, and concurrent operations can
be completed by using multiple processes in Python.
A.True
B.False
Answer: A
Q18. What are the following steps for Python3 operation MySQL database?
A. Establish a database connection
B. Execute SQL statement
C. Install PyMysq1
D. Close the database connection

Answer: CABD
Q19. Which of the following options describes the relationship between the Poisson distribution and the
binomial distribution correctly?
A. Poisson distribution can replace binomial distribution
B. Poisson distribution is an approximate calculation of binomial distribution when n is large and p is

small
C. The mathematical models of Poisson distribution and binomial distribution are both Lagrangian
probabilities
D. Poisson distribution has no relationship with binomial distribution
Answer: B
Q20. Which of the following options describes the numerical characteristics of random variables
correctly?
A. The greater the expectation, the smaller the variance.
B. Variance reflects the degree of deviation between random variables and their mathematical
expectations.
C. Expectation reflects the average level of random variables.
D. The correlation coefficient is used to measure the linear relationship between two variables.
Answer: BCD
Q21. The first-order tensor is a scalar.
A.True
B.False
Answer: A
Q22. Which of the following descriptions of the distribution function, distribution law, and density
function of random variables are incorrect?
A. Discrete random variables have no distribution function
B. The density function can only describe the value law of continuous random variables
C. The distribution function describes the value law of random variables
D. Distribution law can only describe the value law of discrete random variables
Answer: A
Q23. Regarding matrix decomposition, which of the following descriptions is correct?
A. Only square matrices can perform singular value decomposition.
B. Carry out singular value decomposition on matrix A, the left singular matrix and right singular matrix
are both orthogonal matrices.
C. After a matrix is multiplied by its eigenvector, the vector only undergoes scaling transformation.
D. The larger the singular value, the greater the amount of information contained in the corresponding
singular vector.
Answer: BD
Q24. Perform singular value decomposition on a matrix. The singular value matrix must be a square
matrix and the singular values are larger
Arrange in order of smallest.
A.True
B.False
Answer: B
Q25. After multiplying a vector by an orthogonal matrix. The matrix only produces expansion and
contraction effects on the vector, without rotation and
Space mapping function.
A.True
B.False
Answer: B
Q26. The direction of the gradient is the direction of the maximum directional derivative.
A.True
B.False
Answer: A
Q27. Which of the following statements about tensors are correct?
A. Tensors of order 4 and above, like vectors of 4 dimensions and above, have no direct geometric
meaning
B. Array is a tensor
C. The operation and model optimization process in TensorF1ow are all based on tensors
D. A matrix is a second-order tensor
Answer: DABC
Q28. What services does TensorFlow 2.0 support?

A.Lite
B. TensorFlow2.0 core library
C. JavaScript
D.Extended
Answer:ABCD
Q29. What kind of network layer is the pooling layer generally connected to?
A. Input layer
B. Output layer
D. Fully connected layer
Answer: C
Q30.TensorF1ow is one of the most popular deep learning frameworks.
A.True
B.False
Answer: A
Q31. In TensorFlow 2.0, gather is used to increase the dimension.
A.True
B.False
Answer: B
What are the built-in loss functions under Q32.tf.keras.losses?
A. Absolute percentage error
B. Similarity
C. Binary cross entropy loss
D. Error squared loss
Answer: ACD
Q33.TensorFlow is an end-to-end open source platform for machine learning and deep learning.
A.True
B.False
Answer: A
Q34. In Keras, model.summary() is a method used to print the network overview.
A.True
B.False
Answer: A
Q35. What is the role of tf.keras.preprocessing?
A.keras model deployment tool

B.keras data processing tool
C. Keras generative model tool
D.keras built-in optimizer
Answer: B
Q36. The fully connected layer is generally located in the middle of the entire network to reduce the
amount of parameters.
A.True
B.False
Answer: B
Q37.tf.keras.datasets can view the built-in data sets in keras.
A.True
B.False
Answer: B
Q38. Unbiased estimation in point estimation means that the mathematical expectation of the sampling
distribution of the estimator is equal to the true value of the overall parameter.
A.True
B.False
Answer: A
Q39. Which of the following options are the basis for the decision tree to divide nodes?
A.ID3
B. Gini coefficient
C. Information Entropy
D.CART
Answer: ABCD
Q40. The naive Bayes algorithm does not require independent and identical distribution among sample
features.
A.True
B.False
Answer: B
Q41. The higher the validity of the point estimate, the higher the degree of dispersion of the estimator.
A.True
B.False
Answer: B
Q42. What are the common clustering algorithms?
A.K-means
B. Spectral clustering
C. Density clustering
Answer: A
Q43. Which of the following activation functions are prone to vanishing gradient problems?
A. ReLU
B. Tanh
C.Softplus
D. Sigmoid
Answer: BCD
Q44. Which of the following descriptions about the convolutional neural network pooling layer is
correct?
A. The pooled feature image becomes smaller
B. Pooling operation is realized by scanning window
C. The pooling layer can play a role in dimensionality reduction
D. Commonly used pooling methods include maximum pooling and average pooling
Answer: ABCD
Q45. What are the strategies to avoid overfitting in model training?
A. Stop training early
B.Dropout
C. Delay to stop training
D. L2, L1 regular
Answer: ABCD
Q46. Can the perceptron be displayed in space?
A. Point
B. Line
C. Plane
D. Hyperplane
Answer: A
Q47. What is the mini-batch in mini-batch gradient descent?
A. Part of the error is extracted from the error for each parameter iteration
B. Extract a part of the hyperparameters for training at each parameter iteration
C. Extract a part of the feature for training at each parameter iteration
D. At each parameter iteration, a part of the data set is extracted for training
Answer: D
Q48. Which of the following structures belong to LSTM?
A. Input gate
B. Forgotten Door
C. Memory gate
D. Output gate
Answer: ABD
Q49. In the process of deep learning model training, what are the common optimizers?
A. Momenturm
B. Adagrad
C. SGD
D. Adam
Answer: ABCD
Q50. Which of the following statements about L1 regularization and L2 regularization are correct?
A. L2 regularization produces a sparser model than L1 regularization
B. L1 regular term is conducive to enhancing the generalization ability of the model
C. After adding the L2 regular term, the gradient descent algorithm cannot be used to iterate the
parameter values
D. L1, L2 regular terms cannot be applied to the loss function.
Answer: B
Q51. The Softmax function is often used as the output layer of multi-classification tasks.
A.True
B.False
Answer: A
Q52. The training method used by the recurrent neural network is called back-propagation, or BPTT for
short.
A.True
B.False
Answer: A
Q53. Regarding backpropagation, is the following statement wrong?
A. Backpropagation will go through the activation function
B. Backpropagation can only be used in feedforward neural networks
C. Backpropagation can be combined with gradient descent algorithm to update network weights
Answer: B
Q54. What is wrong about the following statement about convolutional neural networks?
A. Common pooling layers include maximum pooling and average pooling.
B. The convolution kernel cannot be used to extract global features of the picture.
C. When processing the image, the image is convolved by scanning the window.
D. Convolutional neural networks can include convolutional layers, pooling layers and fully connected
layers.
Answer: B
Q55. All convolution kernels of the same convolution layer in the convolutional neural network are
weight-sharing.
A.True
B.False
Answer: A
Q56. Which of the following applications are not part of the adversarial generation network?
A. Text generation
B. Data enhancement
C. Image generation
D. Image recognition
Answer: D
Q57. Recurrent neural network can capture dynamic information in serialized data.
A.True
B.False
Answer: A
Q58. What are the elements of artificial intelligence?
A. Algorithm
B. Data
C. Scene
D. Computing power
Answer: ABCD
Q59. The current stage of artificial intelligence is still in the stage of weak artificial intelligence.
A.True
B.False
Answer: A
Q60. Jobs that are highly repetitive and require weak social skills are the jobs most easily replaced by AI.
A.True
B.False
Answer: A
Q61. Which of the following platforms does HiAI Service include?
A. Quick service smart platform
B. Xiaoyi Dialogue Development Platform
C. IOS application development platform
D. Android application development
Answer: ABD
Q63. Which of the following statements about iterators and generators are correct?
A. Only the _iter_ method is implemented in the iterator object
B. yield is the same as return
C. List is an iterator
D. The generator will save the current state
Answer: CD
A. Artificial Intelligence
B.APP
C. Website development
D. Data Science
Answer: ABCD
Q65.Python supports coexistence of multiple versions.
A.True
B.False
Answer: A
Q66. Which is correct about Python?
A. Exceptions in Python can only be thrown by the system
B. Use the try statement to catch exceptions
C. Users can customize exceptions
D. Abnormal will not terminate the execution of the program
Answer:BC
Q67. Which of the following expressions can match 110 in "My phone number is 110, please find me"?
A.re.match("\d+", str1)
B.re.search("\d+", strl)
C.re.split("\d+", str1)
D.re.match("[0-9]+", str1)
Answer: B
Q68. Is the statement about *args in function parameters correct?
A. *args receives parameters in the form of a dictionary
B.*args can be defined before other parameters of the function
C.*args must appear at the end of all parameters
D.*args can receive multiple parameters
Answer: CD
Q69. Methods that start with a double underscore in Python classes may be private methods.
A.True
B.False
Answer: A
Q70. Which statement is correct?
A. If the function has a prescription guide number at a certain point, its directional derivative is unique.
B. The extreme point of the function must be a stationary point
C. The stagnation point of the function must be the extreme point
D. If the function has a gradient at a certain point, its gradient is unique.
Answer: AD
Q71. The covariance matrix is a symmetric matrix, and the elements on the diagonal are the covariances
that change randomly.
A.True
B.False
Answer: A
Q72. What is wrong about the description of the normal distribution?
A. The larger the standard deviation, the steeper the normal distribution curve.
B. A random variable t that obeys the standard normal distribution, its mean and variance are both 0.
C. The normal distribution takes the maximum value at the mean.
D. The larger the standard deviation, the slower the normal distribution curve.
Answer: A
Q73. Optimization problems can be divided into the following categories according to the constraints?
A. Unconstrained problem
B. Equality constraint problem
C. Inequality constraint problem
D. Semi-constrained problem
Answer:ABC
Q74. A matrix corresponds to a linear transformation, which is realized by matrix multiplication.
A.True
B.False
Answer: A
Q75. The probability of occurrence of a solution with a known cause is usually called conditional
probability or prior probability.
A.True
B.False
Answer: A
Q76. What is the built-in optimizer in keras?
A. SGD
B. Adaboost
C. Adadelta
D. Adam
Answer: ACD
Q77. The following are activation functions
A. Adam
B.sigmoid
C. Relu
D.softmax
Answer: BCD
Q78. model. sumnary in Keras (the method used to print the network overview)
A.True
B.False
Answer: A
Q79. Which of the following options are not supported by TensorFlow 2.0?
A.//
B.@
C.^
D.POW
Answer: C
Q80. When creating a network layer, what is the main purpose of adding a dropout layer?
A. Prevent overfitting
B. Extract image features
C. Act as a classifier
D. Increase nonlinearity
Answer: A
Q81. TensorFlow has released versions
A.1
B.2
C.0.1
D.3
Answer:ABC
Q82. For the products of an automobile company, factory A accounts for 40%, factory B accounts for
60%, the defective product rate of A is 1%, and the defective product of B
The rate is 2%. Now a car is a defective product. What is the probability that A will be produced?
A.0.25
B.0.5
C.0.2
D.0.3
Answer: A
Q83. The Naive Bayes algorithm does not require independent and identical distribution among sample
features.
A.True
B.False
Answer: B
Q84. Which of the following options is not an evaluation indicator for parameter estimation?
A. Consistency
B. Unbiased
C. Adaptive
D. Effectiveness
Answer: C
Q85. Which of the following are machine learning algorithms?
A. Supervised learning
B. Unsupervised learning
C. Reinforcement learning
D. Semi-supervised learning
Answer: ABCD
Q86. The higher the validity of the point estimation, the higher the degree of dispersion of the
estimated t.
A.True
B.False
Answer: B
Q87. What is wrong about the following statement about recurrent neural network?
A. Recurrent neural network can be abbreviated as RNN
B. LSTM cannot solve the problem of vanishing gradient
C.LSTM is also a recurrent neural network
D. Recurrent neural network can be expanded according to the time axis
Answer: B
Q88. Which of the following are the activation functions of deep learning algorithms?
A. ReLU
B.Sin
C. Tanh
D. Sigmoid
Answer: ACD
Q89. Which of the following steps in deep learning are automatically completed by the model?
A. Model training
C. Analysis and positioning tasks
Answer: AD
Q90. Which statement is wrong about the adversarial generative network?
A. The discriminant model needs to input labeled real samples
B. Adversarial generative network includes generative model and discriminant model
C. The adversarial generation network cannot be used to generate new pictures
D. The input to the generated model can be random numbers
Answer: C
Q91. Which of the following activation functions can solve the vanishing gradient problem well?
A. Relu
B. Tanh
C. Softsign
D. Sigmoid
Answer: A
Q92. In which of the following fields is deep learning widely used?
A. Decision-making assistance
B. Voice processing
D. Image processing
Answer: ABCD
Q93. Which statement about the optimizer is correct?

A. Adagrad optimizer learning rate is not updated automatically.
B. The common feature of SGD, MBGD and momentum optimizer is that each iteration is updated with
the same learning rate
C. There is no connection between the Adam optimizer and the Adagrad optimizer.
D. The learning rate of the momentum optimizer does not need to be manually set
Answer: B
Q94. In the adversarial generative network, where should the labeled data be placed?
A. As the input value of the generative model
B. As the input value of the discriminant model
C. As the output value of the discriminant model
D. As the output value of the generative model
Answer: C
Q95. In what year was "artificial intelligence" first proposed?
A.1956
B.1946
C.1950
D.1959
Answer: A
Q96. In a neural network based on connectionism, each node can express a specific meaning.
A.True
B.False
Answer: A
A. Symbolism
B. Connectionism
C. Behaviorism
Answer: B
Q98. Which of the following is not an application of image recognition service?
A. Smart photo album
B. Scene analysis
C. Target detection
D. Speech synthesis
Answer: D
Q99. Which of the following is not the core of odelArt s?
A. Cloud
B. End
C.API
D. Side
Answer: C
Q100. Which of the following functions does the data management module of modelArts include?
A. Data version management
B. Upload data
C. Create a data set
D. Annotate data
Answer: ABCD
Q101. The sys module is responsible for the interaction between the program and the operating system,
and provides an interface for accessing the bottom layer of the operating system.
A.True
B.False
Answer: B
Q102. How to get the help information of the function?
A.dir()
B.help()
C.del()
D.id()
Answer: B
Q103. Methods that start with double underscores in Python classes may be private methods.
A.True
B.False
Answer: A
Q104. What variables can only act on the current function?
A. Local variables
B. Anything
C. Global variables
D. Module variables
Answer: A
Q105. Which statement about the data structure in Python is correct?
A. float is an immutable data type.
B. Any tuple can be defined using ().
C. Neither sets nor dictionaries can use subscripts to get elements.
D. The strings defined using single, double and triple quotation marks are exactly the same.
Answer: BD
Q106. The role of the decorator is to add new functions to the function without changing any of the
function's properties.
A.True
B.False
Answer: A
Q107. Which of the following statements about tensors is correct?
A. A matrix is a second-order tensor
C. Tensors of order 4 and above, like vectors of 4 dimensions and above, have no direct geometric
meaning
D. The operations and model optimization processes in TensorFlow are all based on tensors
Answer: ABD
Q108. The determinant of square matrix A is a scalar.
A.True
B.False
Answer: A
Q109. When a function is at a certain point in the variable space, which direction has the greatest rate of
change?
A. Minimum directional derivative
B. Derivative
C. Gradient
D. Directional derivative
Answer: C
Q110. What is the way to view tensor dimensions in TensorFlow 2.0?
A.dtype
B.ndim
C.device
D.dimens
Answer: B
Q111. The mechanism of graphs and sessions is eliminated in TensorFlow 2.0
A.True
B.False
Answer: A
Q112. Which of the following attributes does TensorFlow 2.0 do not support creating tensors?
A.constant
B.create
C.fill
D.zeros
Answer: B
Q113. Why is TensorFlow becoming the mainstream framework for deep learning?
A. Automatic derivative
B. Back-end development
C. Deep Learning API
D. GPU parallel acceleration
Answer: ABCD
Q114. In what year was TensorFlow2.0 released?
A.2017
B.2018
C. 2019
D.2015
Answer: C
Q115. TensorFlow2.0 version is the second version of TensorFlow iteration.
A.True
B.False
Answer: B
Q116. TensorFlow2.0 supports GPU acceleration mainly in the parallel operation of matrix addition,
subtraction, multiplication, and division.
Calculated acceleration.
A.True
B.False
Answer: A
Q117. The three main advantages of Keras interface are user-friendly, modular and combinable, and
easy to expand.
A.True
B.False
Answer: A
Q118. If the accuracy of the model on the training set is very high, but the accuracy on the validation set
is very low. can
What methods are taken to improve the model?
A. Add a regularization term after the loss function
B. Dropout method
C. Increase the amount of training set data
D. Early stopping
Answer: ABCD
Q119. Which of the following models can be used for classification?
A. Decision tree
B. Linear regression
C. Logistic regression
D. Support Vector Machine

Answer: ACD
Q120. Which of the following statements about supervised learning is correct?
A. Supervised learning is a rule-based algorithm
B. Decision tree is a supervised learning
C. Supervised learning can be trained without labels
D. Supervised learning cannot use cross-validation for training
Answer: A
Q121. What is wrong about the SVM algorithm
A. SVM can be used to solve classification problems
B. Support vector is the vector point closest to the hyperplane
C. SVM algorithm can use multiple kernel functions
D. The SVM algorithm cannot be used to process nonlinear data sets
Answer: D
Q122. In the classic convolutional neural network model, what hidden layer does the softmax function
follow?
B. Pooling layer
D. All of the above
Answer: C
Q123. Which of the following is the offline mode in HiAI?
A. HiAI Service
B.HiAI Foundation
C.HiAI Engine
D.HiAI IDE
Answer: B
Q124. Which service is based on the one-stop deep learning platform service provided by Huawei
Cloud’s powerful high-performance computing?
A large number of optimized neural network algorithm models are built-in to allow users to easily use
deep learning technology.
Live scheduling provides model training, evaluation and prediction in an on-demand service-oriented
manner.
A. ModelArts
B.GES
C.Hilens
D.DLS
Answer: A
Q125. Which of the following is the big data service in Huawei EI?
A.DIS
B.MRS
C.HiAI
D. ModelArts
Answer: B
Q126. Which of the following conditions must be met to run a Python program?
A. Windows system
B. Compliance with PEP8
C. Comply with grammatical format
D. Python interpreter
Answer: DC
Q127. Is the Python IO operation correct?
A. After the file operation is completed, use the close method to close it.
B. You can use the context manager to open the file for operation.
C. In r mode, you can use the write method to write data to the file.
D. Use the flush method to write data into memory
Answer: B
Q128. Which of the following is correct?
A. The two object usage judgment results generated by shallow copy and deep copy are both True
B. The two objects of shallow copy use is to judge the result as True
C. The result of using the two objects of deep copy is False
D. The two objects of deep copy use is to judge the result as Irue
Answer: AB
Q129.
i=8
while i>0:
i -= 1
if i == 6:
continue
print (i)
if i == 2:
break
What is the output?
A.
B.
7
C.
D.
Answer: C
Q130. A vector is just a number.
A.True
B.False
Answer: B
Q131. What methods are not included in the built-in metrics of tf. keras. metrics?
A. Accuracy
B. Recall
C.mean
D.Sum
Answer: D
Q132. What are the main functions of the matplotlib library?
A. Drawing images
B. Build a network model
C. Used to process matrix operations
D. Data analysis
Answer: A
Q133. When creating a network, another convolutional layer cannot be directly connected after the
convolutional layer.
A.True
B.False
Answer: B
Q134. Which of the following options belong to the estimator package method under keras?
A. Training
B. Export
C. Evaluation
D. Forecast
Answer: ABCD
Q135. K-fold cross-validation refers to dividing the test data set into K sub-data sets.
A.True
B.False
Answer: A
Q136. Which statement about gradient descent is correct?
A. When using gradient descent, the learning rate should be consistent.
B. Global gradient descent is faster than stochastic gradient descent.
C. Stochastic gradient descent is sensitive to noise points.
D. Adam is not a gradient descent algorithm.
Answer: C
Q137. Which statement about the standard RNN model is correct?
A. There is no many-to-many model structure
B. There will be a problem of attenuation of long-term transmission and memory information

C. There is no one-to-one model structure
D. Do not consider the time direction when backpropagating
Answer: B
Q138. What are the commonly used loss functions in deep learning?
A. Mean square error loss function
B. L1 loss function
C. Self-declining loss function
D. Cross entropy error loss function
Answer: ABD
Q139. Artificial intelligence is the research and development of theories, methods and applications for
simulating, extending and expanding human intelligence.
A new technological science using the system.
A.True
B.False
Answer: A
Q140. Which of the following statements about HiAI Foundation is incorrect?
A. Provides an acceleration API that has nothing to do with processor hardware. Application vendors and
developers can
Accelerate model calculation and operator calculation on the acceleration system.
B. Provide commonly used artificial intelligence business function API, which can run efficiently on
mobile devices.
C. Supports basic operator calculation interfaces, but does not include interfaces such as convolution,
pooling, and full connection.
D. Support AI model management interfaces such as model compilation, model loading, model running,
and model destruction.
Answer: A
Q141. What is the operation that will not change the data after the following transformation?
A.list (set(list))
B.int (float)
C.str(int)
D.list(tuple(list))
Answer: DC
Q142. Which statement about the matrix is incorrect?
A. Any matrix multiplied by the identity matrix equals itself
B. The transpose of the orthogonal matrix A is equal to the inverse of A
C. Non-square matrices also have inverse matrices
D. The transpose of symmetric matrix A is still A itself
Answer: C
Q143. Singular value decomposition can be performed on non-square matrices.
A.True
B.False
Answer: A
Q144. Which of the following evaluation indicators belong to the regression algorithm?
A. Recall rate
B. Accuracy
C. Confusion matrix
D. Mean square error
Answer: D
A.ID3
B. Information Entropy
C.CART
D. Gini coefficient
Answer: ABCD
Q146. Is the following statement about neural networks wrong?
A. There is no connection between neurons in the same layer of the feedforward neural network
B. As the number of hidden layers of the neural network increases, the classification ability of the model
gradually weakens
D. The limitation of a single-layer perceptron is that it cannot solve the XOR problem
Answer:BC
Q147. The data annotation function in ModelArts only supports the modification of a single image tag,
and does not support batch modification.
Change the picture label.
A.True
B.False
Answer: B
A. MRS
B. ModelArts
C.DIS
D.HiAI
Answer: AC
Q149. What will the following code output?
If -1:
a=True+1
b=False+True
else:
a=True-False
b=False+I
print(a,b)
A.True
B.False
C.2 1
D.1 1
Answer: C
Q150. Which of the following options does not support TensorFlow2.0 for dimension transformation is
the attribute.
A.squeeze
B.transpose
C.gathed
D.reshape
Answer: ABD
Q151. The methods that can be used to check whether it is a tensor in TensorFlow 2.0 are:
A.dtype
B.is_ tensor
C.isinst ance
D.device
Answer:BC
Q152. What data type does TensorFlow use?
A.Vector
B.Tensor
C.Scalar
D.Matrix
Answer: ABCD
Q153. Which of the following descriptions of machine learning is correct?
C. Deep learning and machine learning are both artificial intelligence but have no relationship with each
other
Answer: A
Http://www.passcert.com
The safer , easier way to help you pass any IT exams.
Exam : H13-311_v3.0
Title : H13-311_V3.0: HCIA-AI

V3.0
Version : V8.02
1 / 61
1.Faced with the challenge of achieving efficient distributed training for ultra-large-scale models,
MindSpore is handled as?
B. Serial
C. Manual parallel
Answer: A
2.Which of the following is not the difference between Python 2 and Python 3?
A. print
B. Unicode
C. import
D. xrange
Answer: C
3.Which of the following options is not central to linear algebra?

A. Probability theory
B. Linear transformation
C. Matrix theory
D. Vector space
Answer: A
4.The Python dictionary is widely identified by "{}". and the internal data consists of the key and its
corresponding value
A. True
B. False
Answer: A
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?
B. Eigenvalues
C. Hyperparameter
D. parameter
Answer: D
6.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?
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
2 / 61
7.GBDT Compared with the random forest algorithm, which of the following statements is wrong?
A. GBDT Algorithm is easier to underfit than random forest
B. Random forest is calculated in parallel, while GBDT Can't
C. GBDT Algorithm is easier to overfit than random forest
D. GBDT And random forest are built on CART Based on the tree
Answer: C
8.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. accuracy
B. categorical_ 1oss
C. loss
D. categorical accuracy
Answer: A C
9.According to Huawei Cloud EI Intelligent platform, which of the following solutions can be provided?
A. Crowd statistics program
B. Policy query scheme based on knowledge graph
C. Vehicle identification scheme
D. Intrusion recognition scheme
Answer: A B C D
10.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
11.Functions are well-organized, non-reusable code segments used to implement a single, or associated
Function.
A. True
B. False
Answer: B
12.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 I Right Answers)
Answer: D
3 / 61
13.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.
A. TRUE
B. FALSE
Answer: A
14.TensorF1ow Which of the following companies first developed?

A. Oracle
B. Facebook
C. Nvidia
D. Google
Answer: D
15.In Huawei Cloud El Enterprise Intelligence, which basic platform services are included? (Multiple
Choice)
A. Machine learning
B. Deep learning
C. Graph engine
D. Batch processing
Answer: A B C D
16.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.
C. The activation function is partly a nonlinear function, partly a linear function
Answer: B
17.PyTorch Which of the following functions does not have?

A. Inline keras
D. GPU accelerate
Answer: A
18.What is wrong description of the Python module?

A. The Python module is a Python file that ends with .py and contains Python object definitions and
Python statements.
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.
4 / 61
Answer: C
19.What are the commonly used activation functions? (Multiple Choice)

A. sigmoid
B. tanh
C. relu
D. danish
Answer: A B C
20.What are the Python language data types? (Multiple Choice)

A. numbers
B. string
C. list
D. tuple
E. dictionary
Answer: A B C D E
21.When the voice recognition service is successfully called, which field is the recognition result stored
in?
A. result
B. content
C. data
D. text
Answer: A
22.What does factors that promote the development of artificial intelligence not include?
A. Big data
B. Computing ability
C. Algorithm theory
D. Block chain
Answer: D
23.The image data of the passport recognition service is not needed base64 Coded.
A. TRUE
B. FALSE
Answer: B
24.The loss function of logistic regression is the cross-entropy loss function.

A. TRUE
B. FALSE
Answer: A
25.Pytorch Which company launched it first?

A. Baidu
5 / 61
B. Google
C. Facebook
D. Huawei
Answer: C
26.It is fine to pass in only one image when calling the face comparison service.
A. TRUE
B. FALSE
Answer: B
27.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
28.twenty two.Among machine learning algorithms, which of the following is not unsupervised learning?
A. GMM
B. Xgboost
C. Clustering
Answer: B
29.Which of the following is not included in the recurrent neural network usage scenario?
A. Machine translation
C. Image style migration
D. Text generation
Answer: C
30.What are the characteristics of Python code? (Multiple Choice)

A. leg1bil1ty
B. Simpl1c1ty
C. Rapidity
D. Scalab1hty
Answer: A B D
31.In a convolutional neural network, different layers have different functions.

Which of the following layers can play the role of dimensionality reduction?
A. Input layer
B. Fully connected layer
D. Pooling layer
6 / 61
Answer: B C D
32.After the data has completed the feature engineering operation, in the process of constructing the
model, which of the following options is not a step in the decision tree construction process?
A. Pruning
C. Data cleaning
D. Decision tree generation
Answer: C
33.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
34.Which of the following descriptions about the Recurrent Neural Network (RNN) is correct?
A. Can be used to process sequence data.
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
35.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. Expert system
B. Machine learning
C. Neural Network
Answer: B
36.TensorFlow only supports Python and R language

A. True
B. False
Answer: B
37.The naive Bayes algorithm does not require independent and identical distribution among sample
features.
A. TRUE
B. FALSE
Answer: B
7 / 61
38.What are the scenarios or industries that are suitable for using Python? (Multiple Choice)
A. Artlfic1al intelligence
B. web development
C. Game development
D Hardware development
Answer: A B C
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.Huawei's chip support HUAWEI HiAI Which module of?

A. HiAI Engine
B. HiAI Foundation
C. HiAI Framework
D. HiAI Service
Answer: B
41.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. Client
B. feed
C. placeholder
D. fetch
Answer: C
42.The optimizer is an important part of training neural networks.

The purpose of using the optimizer does not include which of the following:
B. Reduce the difficulty of manual parameter setting
C. Avoid overfitting
Answer: C
43.The TensorFlow framework does not support Windows systems.

A. True
B. False
Answer: B
44.The following is the correct difference between machine learning algorithms and traditional rule-based
8 / 61
methods?
B. Traditional rule-based methods use explicit programming to solve problems
C. The mapping relationship of the model in machine learning is automatically learned
D. The mapping relationship of the model in the machine learning institute must be implicit
Answer: A B C
45.In machine learning, what input the model needs to train itself and predict the unknown?
A. Manual procedure
B. Neural Networks
D. historical data
Answer: D
46.Traffic agent can achieve () Hour full time/Global traffic awareness?

A. 7×24
B. 5×24
C. 5×12
D. 7×12
Answer: A
47.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. Gradient descent
B. Dropout
C. Cross-validation
D. Regularization
Answer: A
48.Deflated AI The logical architecture of the processor does not include which of the following options?
A. DVPP
B. GPU
C. AI Calculation engine
Answer: B
49.HUAWEI HiAI Engine Can easily combine multiple AI Ability and App integrated.
A. TRUE
B. FALSE
Answer: A
50.What is the subject of artificial intelligence?

A. Mathematics and Physiology
B. Psychology and Physiology
9 / 61
C. Linguistics
D. Comprehensive mterd1sc1pl1nary and marginal d1sc1phnes
Answer: D
51.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
C. Numerical calculation method
D. Programm1ng
E. Computer calculation results
Answer: A B C D E
52.What people say about neural networks is wrong?

A. As the number of hidden layers of the neural network increases, the classification ability of the model
gradually weakens
B. The limitation of a single-layer perceptron is that it cannot solve the XOR problem
D. There is no connection between neurons in the same layer of the feedforward neural network
Answer: A
53.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
54.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 A
55.The number or matrix is stored in the tensor
A. True
B. False
Answer: B
56.SVM What are the commonly used kernel functions in?

B. Polynomial kernel function
10 / 61
C. Sigmiod Kernel function

D. Linear kernel
Answer: A B C D
57.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
58.Which of the following does the convolutional neural network structure not include?
B. Pooling layer
C. Loop layer
Answer: C
59.If a model has a large deviation on the test set and a small variance, it means that the model?
A. Overfitting
B. May be overfitting may be underfitting
C. Just fit
D. Underfitting
Answer: D
60.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. TRUE
B. FALSE
Answer: A
61.Jobs that are repetitive and require weak social skills are the easiest to be AI Replaced work.
A. TRUE
B. FALSE
Answer: A
62.Regular term can also be added to logistic regression to avoid overfitting.

A. TRUE
B. FALSE
Answer: A
63.What is in the category of artificial intelligence? (Multiple choice)

A. Action
11 / 61
B. Perception
C. Cognition
Answer: A B C
64.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
Answer: A
65.enter 32*32 Image with size 5*5 The step size of the convolution kernel is 1 Convolution calculation,
output image Size is:
A. 28*23
B. 28*28
C. 29*29
D. 23*23
Answer: B
66.Which of the following is not MindSpore common Operation?

A. signal
B. math
C. nn
D. array
Answer: A
67.The current technology application directions of artificial intelligence mainly include?

B. Control System
C. Computer vision
D. Speech Recognition
Answer: A C D
68.In TensorFlow, data is represented in the form of tensors and calculation charts.
A. True
B. False
Answer: B
69.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
12 / 61
70.The following about the general form identification service returned type The field statement is correct?
A. type Representative text recognition area type
B. type for text Time represents the text recognition area
C. type Representative form type
D. type for table Time represents the form recognition area
Answer: B D
71.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: A C D
72.TensorFlow The data type used is?

A. Scalar
B. Vector
C. Tensor
D. Matrix
Answer: C
73.TensorFlow2.0 middle Keras The three main advantages of the interface are user-friendly, modular
and Combination, easy to expand.
A. TRUE
B. FALSE
Answer: A
74.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
C. aHa2+a3=0
D. 2a1+a2+a3=0
Answer: A
75.What does not belong to supervised learning?

B. Support vector machine
C. Decision tree
D. Princi1pal component analysis
Answer: D
76.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
13 / 61
of the following?
A. New ways of collaboration
B. New programming language
Answer: B
77.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
78.When training the network, we often encounter many problems. For the problem of gradient
disappearance, we can alleviate the problem by choosing which of the following functions to use?
A. Softsign function
B. Relu function
C. tanh function
D. Sigmoid function
Answer: B
79.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 lo be indented (Right Answers}
D. Return the result with return and the function ends.
Answer: A B C D
80.Recurrent neural networks can capture dynamic information in serialized data.

A. TRUE
B. FALSE
Answer: A
81.AI Chips are divided into business applications and can be divided into?
A. training
B. GPU
C. Model building
D. reasoning
Answer: A D
82.Which of the following statements about passport recognition services are correct?
A. country_code Country code representing the country where the passport was issued
14 / 61
B. nationality Nationality of representative holder

C. passport_number Representative passport number
D. confidence Confidence information of related fields. The greater the confidence, the higher the
reliability of the corresponding field identified this time. In a statistical sense, the greater the confidence,
the higher the accuracy
Answer: A B C D
83.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
84.Corvolutional neural networks are more suitable for dealing with speech recognition problems
A. True
B. False
Answer: B
85.Which of the following activation functions are prone to vanishing gradient problems?
A. ReLU
B. Softplus
C. Tanh
D. Sigmoid
Answer: C D
86.Which of the following products is related to artificial intelligence? (Multiple Choice)

A. Alpha Go
B. Self-driving
C. Voice input
D. Huawei Mate mobile phone
Answer: A B C D
87.In the process of deep learning model training, what are the common optimizers?
A. Adam
B. Adagrad
C. SGD
D. Momentum
Answer: A B C D
88.What model is not a cyclic neural network?

A. RNN
B. LSTM
C. GBOT
15 / 61
D. GRU
Answer: C
89.Huawei Ascend AI Chip is NPU (Neural network processor) One of the typical representatives.
A. TRUE
B. FALSE
Answer: A
90.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
B. 2
C. 3
D. 4
Answer: A
91.The model composed of machine learning algorithms cannot represent the true data distribution
function on a theoretical level. Just approach it.
A. TRUE
B. FALSE
Answer: A
92.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)=O
B. D(X+ Y)=D(X)+D(Y)
C. D(CX)=C'C'D()()
D. D(XY)=D(X)D{Y)
Answer: B D
93.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 hyper parameter optimization
Answer: B
94.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 d1slribulion function
D. Distribution function describes the value rule of random variables
Answer: C
16 / 61
95.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: A B C D
96.The following options do not belong to the scope of Huawei's full-stack solution are?
B. Edge computing
D. Chip enable
Answer: B
97.In which year was the first introduction of "artificial intelligence"?

A. 1946
B. 1960
C. 1916
D. 1956
Answer: D
98.Feature is the dimension that describes the characteristics of the sample.

Regarding its interpretability in traditional machine learning and deep learning, the following statement is
correct:
B. Features are weak in traditional machine learning, but strong in deep learning
C. Features are weak in interpretability in traditional machine learning and deep learning
D. Features are interpretable in traditional machine learning and deep learning
Answer: A
99.Which of the following is not a specific technology of artificial intelligence?

A. Knowledge map
B. Machine translation
C. Riemann geometry
D. Semantic understanding
Answer: C
100.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
17 / 61
Answer: B C D
101.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
102.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:
D. Hinge Loss function
Answer: B
103.Which of the following is an image label service URI?

A. /v1.0/image/recognition
B. /v1.0/image/celebrity-recognition
C. /v1.0/image/recapture-detect
D. /v1.0/image/tagging
Answer: D
104.Convolutional neural networks are more suitable for image recognition problems than cyclic neural
networks.
A. True
B. False
Answer: A
105.In neural networks, weights are often shareD.

Which of the following neural networks will share weights?
A. sensor
Answer: B D
106.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
B. False
Answer: A
18 / 61
107.Which of the following items are included in the results returned when the face search service is
successfully called?
A. Searched face similarity
B. Searched faces id
C. Searched face position
D. Searched face number
Answer: A B C
108.tf.keras.datasets Can view keras The built-in data set.

A. TRUE
B. FALSE
Answer: A
109.TensorFlow2.0 of keras.preprocessing The role is?

A. keras Data processing tools
B. keras Built-in optimizer
C. keras Model deployment tool
D. Keras Generative model tool
Answer: A
110.HiAI Mobile computing platform provides rich upper-level functional services API, Can run efficiently
on mobile devices.
A. TRUE
B. FALSE
Answer: A
111.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 combmat1on of weight and deviation until the best value 1s 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
112.What are the mainstream deep learning open source tools? (Multiple choice)
A. TensorFlow
B. Caffe
C. Torch
D. Theano
E. Scikit..Jeam
Answer: A B C D
19 / 61
113.Which of the following options belong to Tensorflow concept? (Multiple Choice)

A. Tensor
B. Variables
C. Placeholder
D. Operation
E. Conversation
Answer: A B C D E
114.The following evaluation indicators belonging to the regression algorithm are?

A. Recall rate
B. Confusion matrix
C. Mean square error
D. Accuracy
Answer: C
115.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
116.What is wrong description of the normal distribution?

A. In natural phenomena and social phenomena, many random variables obey or approximate a normal
distribution.
B. The normal distribution lakes the maximum value at the mean
C. The larger the standard deviation the steeper the normal d1stribuhon curve
D. The larger the standard deviation, the slower the normal distribution curve.
Answer: C
117.The Python dictionary is identified by “{}”, and the internal data consists of the key and its
corresponding value.
A. True
B. False
Answer: A
118.Which of the following are solutions for the Huawei Cloud El industry scenario? (Multiple Choice)
A. Intelligent Logistics
B. Intelligent Water
C. Intelligent transportation
D. Intelligent Finance
E. Intelligent manufacturing
Answer: A B C D E
119.Self-encoder is an unsupervised learning algorithm

A. True
20 / 61
B. False
Answer: A
120.Tensor Yes MindSpore The storage component of the data.

A. TRUE
B. FALSE
Answer: A
121.Which of the following options is not a tensor attribute?

A. name (Logo)
B. shape (Dimensionality)
C. type (Type)
D. node (Node)
Answer: D
122.On the premise of ensuring data privacy and security, federated learning utilizes different data
sources to cooperatively train models to improve breakthroughs in data bottlenecks.
A. TRUE
B. FALSE
Answer: A
123.The following description of machine learning is correct?

C. Deep learning and machine learning are both artificial intelligence but have no relationship to each
other
Answer: A
124.The following about the standard RNN Model, the correct statement is?
B. Do not consider the time direction when backpropagating
D. There will be a problem of attenuation of long-term transmission and memory information
Answer: D
125.None of the following options TensorFlow2.0. The supported operators are:

A. pow
B. @
C. ^
D. //
Answer: C
126.The face detection service will report an error when there is no face in the test image.
21 / 61
A. TRUE
B. FALSE
Answer: B
127.Which of the following steps in deep learning are automatically completed by the model?
A. Model training
Answer: B D
128.About the image content review service returned when the call is successful suggestion Field, the
correct statement is?
A. pass Representative does not contain sensitive information, passed
B. review The representative needs a manual review
C. suggestion The field represents whether the test passed
D. block Representative contains sensitive information and does not pass
Answer: A B C D
129.Which of the following are the elements of artificial intelligence? (Multiple choice)
A. Algorithm
B. Computing power
C. Data
D. Scene
Answer: A B C D
130.In the environment preparation for invoking HUAWEI CLOUD services, which of the following does
not require confirmation of installation?
A. requests
B. setuptools
C. websocket-client
D. Jieba
Answer: D
131.Which of the following are the activation functions of deep learning algorithms?
A. Sigmoid
B. ReLU
C. Tanh
D. Sin
Answer: A B C
132.What are the numeric types of Python? (Multiple Choice)

A. int (integer type)
B. long (long integer type)
22 / 61
C. float (floating point type)

D. complex (complex number type)
Answer: A B C D
133.ModelArts Service and (). The combination of services can easily deploy the model to "end"?
A. OBS
B. OCR
C. ECS
D. HiLens
Answer: D
134.Which of the following is not an artificial intelligence school?

A. Symbolism
B. StatJst1caltsm
C. Behaviorism
D. Connectionism
Answer: B
135.What is not the optimization method in deep learning?

A. Random gradient descent
B. Back propagation algorithm
C. Principal component analysis
D. Momentum
Answer: C
136.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
137.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 lo unsupervised learning
Answer: A
138.Loss function and model function are the same thing.

A. TRUE
B. FALSE
Answer: B
139.Gyclic neural networks are more suitable for dealing with image recognition problems.
23 / 61
A. True
B. False
Answer: B
140.Which of the following is not MindSpore Features of the core architecture?

A. Automatic differentiation
B. Automatic tuning
C. Automatic coding
D. Automatic parallel
Answer: C
141.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
Answer: B
142.As the following, what are the Python language design philosophy? (Multiple Choice)
A. Beautiful
B. Expensive
C. Explicit
D. Simple
Answer: A C D
143.Tensorflow Operations and Computation Graph are not - run in the Session
A. True
B. False
Answer: B
144.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
145.Generating a confrontation network is like a game system. The generator generates fake samples,
and the discriminator judges whether it is true or false.
Our ideal result is?
B. The discriminator efficiently distinguishes the true and false samples generated by the generator
24 / 61
D. The generator produces different samples

Answer: C
146.GPU Good at computationally intensive and easy to parallel programs.

A. TRUE
B. FALSE
Answer: A
147.The for loop statement in the Python language can iterate through the items in any sequence.
A. True
B. False
Answer: A
148.What are the algorithms supported by Tensorflow? (Multiple Choice}

A. GNN
B. ZNN
C. RNN (Rig ht Answers)
D. HUT
Answer: A C
149.Which of the following is the computing model of TensorFlow?

A. Calculation chart
B. Tensor
C. Conversation
D. Variable
Answer: A
150.HUAWEI HiAI Empower APP What value?

A. Safety
B. stable
C. real time
D. at any time
Answer: A B C D
151.When we describe the house, we usually use the attributes such as residential area, house type,
decoration type, etC. If we use plain With Bayes as a model, we assume that there is no relationship
between attributes.
A. TRUE
B. FALSE
Answer: A
152.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.:
25 / 61
A. Motherboard
B. RAM
C. GPU
D. CPU
Answer: C
153.Huawei AI The full scenarios include public cloud, private cloud, various edge computing, IoT industry
terminals, and consumer terminals and other end, edge, and cloud deployment environments.
A. TRUE
B. FALSE
Answer: A
154.Which of the following options are <Artificial Intelligence Comprehensive Experiment> May be used
in calling Huawei cloud services?
A. ak
B. region
C. project_id
D. sk
Answer: A B C D
155.TensorFlow2.0 not support tensor in GPU with CPU Transfer between each other.
A. TRUE
B. FALSE
Answer: B
156.About Bayesian formula- P(WlX)=P(XlW)*P(W)/P(X) What is the correct description?

A. P(WIX) is a prior probability
B. P(XIW) 1s a conditional probability
C. P(W) is the posterior probability
D. P(X) is the posterior probability
Answer: B
157.TensorFlow2.0 of keras.metrics
The methods that are not included in the built-in indicators are?
A. Accuracy
B. Recall
C. Sum
D. Mean
Answer: C
158.The commonly used functions for mathematical operations in Python are basically in lhe math
module and the cmath module.
A. True
B. False
26 / 61
Answer: A
159.What are the implementation modes ofTensorflow? (Multiple Choice)

A. Stand-alone mode
B. D1stnbuted mode
C. Reverse mode
D. Forward mode
Answer: A B
160.Use with Atlas300 (3000) What conditions need to be checked when the accelerator card server
compiles and runs the program?
A. carry out Atlas Driver Installation
B. It has been installed CUDA Software package
C. It has been installed Cmake Compilation tool
D. Complete environment configuration
Answer: A C D
161.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
162.The training error will continue to decrease as the model complexity increases.
A. TRUE
B. FALSE
Answer: A
163.Which of the following services belong to Huawei EI Service family?

A. Dialogue robot
B. EI Basic services
D. EI Big data service
Answer: A B C D
164.What is wrong description or backpropagation?

A. The learning process or 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. until the response of the network to the input reaches the predetermined target range.
Answer: C
165.The following about KNN Algorithm k The value description is correct?
27 / 61
A. K The larger the value, the easier the model is to overfit

B. K The larger the value, the smoother the segmentation surface of the classification
C. K Value is a hyperparameter
D. can k Value is set to 0
Answer: B C
166.Which of the following description about Python regular expression is not correct?
A. Python has added there 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
167.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
C. A+B
D. AB-BC
Answer: B
168.Which of the following description 1s 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
169.In random forest, what strategy is used to determine the outcome of the final ensemble model?
B. Find the average
C. Voting system
D. Cumulative system
Answer: B C
170.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
Answer: A
171.Which is not a deep learning algorithm?

A. Setr-encoder
28 / 61
B. Convolutional neural networks

C. Recurrent neural networks
D. Support vector machine
Answer: D
172.Computer vision is the study of how to make computers "see".

A. TRUE
B. FALSE
Answer: A
173.From the labeled historical data, we can predict that the retail sales of the next quarter will be in 20-30
Wan still 30-40 Ten thousand", What is the problem?
B. Rule problem
C. Classification problem
D. Clustering problem
Answer: C
174.Which of the following is not part of the deep learning development framework?
A. CNTK
B. Keras
C. SAFA
D. MXNet
Answer: C
175.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
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 wrth commas
D. When a tuple contains only one element you need to add a comma after the element
Answer: A B C D
176.Atlas 200 DK Yiteng 310 The main application of the chip is?
A. Model reasoning
B. Build model
C. Training model
Answer: A
177.With Atlas300 (3000)Install the accelerator card in the server DDK (Device Development Kit) Only
need to install when dependent on the environment Python3.
A. TRUE
B. FALSE
Answer: A
29 / 61
178.in TensorFlow2.0 in tf.contrib The method can continue to be used.

A. TRUE
B. FALSE
Answer: B
179.What are the regularizations in deep learning? (Multiple Choice)

A. L1 norm. L2 norm
B. Data set enhancement
C. Integration method
D. Dropout
Answer: A B C D
180.Which of the following statements about universal form recognition services are correct?
A. rows Represents the line information occupied by the text block, the number is from 0 Start, list form
B. colums Represents the column information occupied by the text block, the number is from 0 Start, list
form
C. The incoming image data needs to go through base64 coding
D. words Representative text block recognition result
Answer: A B C D
181.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
182.In a neural network based on connectionism, each node can express a specific meaning.
A. TRUE
B. FALSE
Answer: B
183.The Python language can use multiple statements on the same line, separated by commas ".".
A. True
B. False
Answer: B
184.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.
A. TRUE
B. FALSE
Answer: A
185.Huawei Cloud El builds enterprise intelligence services based on three-tier services.
30 / 61
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
186.What of the following does belong to convolutional neural network (CNN)? (Multiple Choice)
A. VGGNet
B. ResNet
C. AlexNet
D. GoogleNet
Answer: A B C D
187.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. TRUE
B. FALSE
Answer: A
188.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
189.In the gradient descent algorithm, which of the following algorithms is the most confusing algorithm
for the trajectory on the loss function surface?
A. SGD
B. BGD
C. MGD
D. MBGD
Answer: A
190.Which of the following are included in the application scenarios ofTensorflow? (Multiple choice}
A. Speech recognition
B. Face recognition
C. Image style changes
D. Autopilot
Answer: A B C D
191.Which of the following is not a face recognition technology?
31 / 61
A. Face search
B. Face comparison
C. Face Detection
Answer: D
192.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. Park Agent
B. Traffic Agent
C. Network agent
D. Industrial Agent
Answer: C
193.Regarding the convolutional neural network pooling layer, the following description is correct?
A. Pooling operation is realized by scanning window
C. Commonly used pooling methods include maximum pooling and average pooling
D. The pooled feature image becomes smaller
Answer: A B C D
194.If the deep learning neural network has the problem of gradient disappearance or gradient explosion,
our common solution is.
A. Gradient shear
C. use Relu Activation function
D. Regularization
Answer: A C D
195.Python script execution mode includes interactive mode and script mode
A. True
B. False
Answer: A
196.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
197.What are the common types of dirty data?

A. Malformed value
32 / 61
B. Duplicate value
C. Logically wrong value
D. Missing value
Answer: A B C D
198.Training error will reduce the accuracy of the model and produce under-fitting.
How to improve the model fit? (Multiple choice)
A. Increase the amount of data
C. Reduce regularization parameters
D. Add features
Answer: A B D
199.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
200.The following statement about recurrent neural networks is wrong?

A. Recurrent neural network can be unfolded according to the time axis
B. LSTM Unable to solve the problem of vanishing gradient
C. LSTM It is also a recurrent neural network
Answer: B
201.What are the service categories included in Huawei Cloud El Enterprise Intelligence? (Multiple
Choice)
A. El visual cognition
B. El online games
C. El speech semantics
D. El industry scene
Answer: A C D
202.Which of the following description of the validation set is wrong?

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 1s used for training and 20% 1s used for verification.
Answer: A
203.The following does not belong TensorFlow2.0 Is characterized by?

A. Multi-core CPU accelerate
33 / 61
B. distributed
C. multi-language
D. Multi-platform
Answer: A
204.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. Law of reciprocity
D. Rule of Normalization
Answer: A
205.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
A. True
B. False
Answer: A
206.Which of the following neural network structures will share weights? (Multiple choice)
A. Convolutional neural network
D. All of the above
Answer: A B
207.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
208.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 data
B. Industry wisdom
C. algorithm
34 / 61
D. Computing power
Answer: A B C D
209.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?
A. Backpropagation update parameters
B. Forward the dream and update
C. Backpropagation calculation result
D. Forward calculation results
Answer: A
210.Among the following properties TensorFlow2.0 Does not support creation tensor. The method is?
A. zeros
B. fill
C. create
D. constant
Answer: C
211.Which of the following are the cloud services provided by Huawei Cloud El Visual Cognition? (Multiple
choice)
A. Text recognition
B. Face recognition
D. Content detection
E. Image processing
Answer: A B C D E
212.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 011e possible outcome for each trial. and all possible outcomes of the trial
can be clarified m advance
C. All results of the lest cannot be clarified in advance
D. It Is not possible to determine which result will appear before conducting a test.
Answer: A B D
213.HUAWEI HiAI Which module does the face detection in the platform belong to?
A. HiAI Engine
B. HiAI Framework
C. HiAI Foundation
D. HiAI Service
Answer: A
214.According to the American Society of Automotive Engineers (SAE) The automatic driving is divided
35 / 61
into Which level?

A. L1~L4
B. L1~L5
C. L0~L4
D. L0~L5
Answer: D
215.The Python language does not allow embedding another loop body in a loop body
A. True
B. False
Answer: B
216.TensorFlow2.0 The methods that can be used for tensor merging are?
A. join
B. concat
C. split
D. unstack
Answer: B
217.What are the commonly used gradient descent optimization functions? (Multiple Choice)
A. Random gradient descent
B. Adadelta
C. Adagrad
D. momentum
E. RMSProp
Answer: A B C D E
218.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 1s 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
219.Bagging In integrated learning, the relationship between each base learner is?
B. relationship
C. The latter model must be built on top of the previous model
D. Independent
Answer: B
220.In deep learning tasks, when encountering data imbalance problems, which of the following methods
can we use to solve the problem?
A. batch deletion
36 / 61
Answer: B C D
221.MSIR It is a simple, efficient and flexible graph-based function IR, A functional semantic type that can
be represented Have?
A. Free variable
B. Higher order function
C. Interrupt
D. Recursion
Answer: A B D
222.Linear regression in 3 The fitted surface in the dimension above dimension is?
A. Curved surface
B. flat
C. Hyperplane
D. Hypersurface
Answer: C
223.Which of the following features does Python support? (Multiple choice)

A. Inheritance
B. Heavy load
C. Derived
D. Multiple inheritance
Answer: A B C D
224.Tensorflow supports multi-TPU cluster computing

A. True
B. False
Answer: A
225.What quotation marks can the Python language use? (Multiple Choice)
A. Single quotes
B. Double quotes
C. Three quotes
D. Four quotes
Answer: A B C
226.What are the core concepts in TensorFlow? (Multiple Choice)

A. Gridding
B. Calculation Chart
C. Dot Product
D. Tensor
37 / 61
Answer: B D
227.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
C. Manifold learning
D. Feature Engineering
Answer: D
228.From the perspective of technical architecture AI The classification of chips includes?

A. FPGA
B. CPU
C. GPU
D. ASIC
Answer: A B C D
229.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
230.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. Feature engineering
C. Data cleaning
D. Model building
Answer: B
231.Which of the following descriptions about Python program are correct? (Multiple choice)
A. The program consists of modules
B. Module contains statements
C. statement contains an expression
D. Expressions create and process objects
Answer: ABCD
232.Which is not a Python operator?

A. arithmetic operator
B. Reasoning operator
38 / 61
C. Logical operator
D. Comparison operator
Answer: B
233.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
234.In polynomial regression, there is a square term in the formula of the model, so it is not linear.
A. TRUE
B. FALSE
Answer: B
235.HUAWEI HiAI What are the supported models?

A. P30
B. Mate 20
C. glory V20
D. iPhone10
Answer: A B C
236.Which description is wrong about the hyperparameter?

A. 1-typerparameters 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
237.In supervised learning," Those who are near Zhu are red and those who are near Mo are black"Is
used to describe which of the following models?
A. K-Means
B. SVM
C. KNN
D. Neural Networks
Answer: C
238.What are the commonly used loss functions? (Multiple choice)

A. Mean variance
B. Sigmoid cross entropy
C. Soft max cross enl10py
D. Sparse cross entropy
E. Weighted S1gmo1d Cross Entropy
Answer: A B C D E
39 / 61
239.There are many commercial applications for machine learning services.

What are the main business scenarios covered? (Multiple Choice)
A. Flnancial product recommendation
B. Predictive maintenance
C. Telecom customer retention
D. Retailer grouping
Answer: A B C D
240.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. TRUE
B. FALSE
Answer: A
241.The test error will keep getting smaller as the complexity of the model increases.
A. TRUE
B. FALSE
Answer: B
242.Voice recognition refers to the recognition of audio data as text data.

A. TRUE
B. FALSE
Answer: A
243.The pooling layer in the convolutional neural network can reduce the size of the lower layer input.
Common pooling is:
A. Minimum strata
C. Maximum pooling layer
D. Average pooling layer
Answer: C D
244.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
245.Which of the following is HUAWEI HiAI Foundation Function of the module?

A. App integrated
B. Let the service actively find users
C. Quickly convert and migrate existing models
D. According to user needs, push services at the right time and at the right time
40 / 61
Answer: C
246.TensorFlow in keras The built-in optimizer of the module has?

A. Adam
B. SGD
C. Adaboost
D. Adadelta
Answer: A B C D
247.Artificial intelligence at this stage is still in the weak artificial intelligence stage.
A. TRUE
B. FALSE
Answer: A
248.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. The global gradient algorithm can find the minimum value of the loss function
B. Batch gradient algorithm can solve the local minimum problem
C. Stochastic gradient algorithm can find the minimum value of the loss function
D. The convergence process of the global gradient algorithm is time-consuming
Answer: C
249.In the deep learning neural network, the perceptron is the simplest neural network. The correct
statement about its structure is:
A. There are only two hidden layers
B. Only one hidden layer
C. Its network uses Sigmoid Activation function
D. Its network uses Relu Activation function
Answer: B
250.Which of the following data types does Tensorflow not support?

A. int8
B. ftoat32
C. double64
D. bool
Answer: C
251.View Atlas300 (3000) Which command should be used if the accelerator card driver is installed
successfully?
A. npu-sim info
B. npu info
C. atlas-driver info
D. atlas info
41 / 61
Answer: A
252.K Folding cross-validation refers to dividing the test data set into K Sub-data sets.
A. TRUE
B. FALSE
Answer: B
253.It is known that the total parameter of a certain layer of the fully connected neural network is 330, The
number of neurons in the previous layer and this layer May be?
A. 32 with 10
B. 10 with 33
C. 33 with 10
D. 9 with 33
Answer: B C
254.Add to the loss function of linear regression L1 Regular term, this time the regression is called Lasso
return.
A. TRUE
B. FALSE
Answer: A
255.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 d1stnbutJon 1s an approx1mallon of the bmom1al d1stnbubon when n 1s very large and p 1s
small (
C. Poisson distribution has no relationship with binomial distribution
D. Poisson distribution can replace binomial distribution
Answer: B
256.Ce11 Provides basic modules for defining and performing calculations, Ce11 The object can be
executed directly, the following statement is wrong?
A. __init__,Initialization parameters(Parameter), Submodule(Ce11),operator(Primitive)Equal group
Software for initial verification
B. Construct, Define the execution process. In graph mode, it will be compiled into graphs for execution,
and there is no syntax restriction
C. There's some left optim Commonly used optimizers,wrap Pre-defined commonly used network
packaging functions Ce11
D. bprop (Optional), The reverse of the custom module
Answer: A C D
257.Where is the main foothold of symbolism?

A. The foothold is in neuron networks and deep learning.
42 / 61
B. The foothold is reasoning, symbolic reasoning and machine reasoning.

D. The foothold is behavior control, adaptive and evolutionary computing.
Answer: B
258.Which of the following schools does the neural network study belong to?
A. Symbolism
B. Connectionism
C. Behaviorism
Answer: B
259.What is wrong about the image content review service?

A. politics Test results for sensitive persons involved in politics
B. terrorism Test results for political violence
C. confidence Represents confidence, range 0-100
D. labe1 Label representing each test result
Answer: C
260.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: A B C
261.Which of the following can improve the computational efficiency of the neural network model?
(Multjple choice)
A. TPU
B. GPU
C. FPGA
D. Large-scale distributed cluster
Answer: A B C D
262.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:
B. Its input node has computing power
D. Information only along-Pass through
Answer: D
43 / 61
263.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 Regularization can do feature selection
B. L1 with L2 Regularization can be used for feature selection
C. L2 Regularization can do feature selection
D. L1 with L2 Regularization cannot be used for feature selection
Answer: A
264.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
265.How many spaces does PEP 8 stipulate that Python's first fine indentation needs to indent?
A. 1
B. 2
C. 4
D. 8
Answer: C
266.Which of the following environments does not support to install the T11nsorFlow?
A. Linux
B. Mac OS
C. Docker
D. OpenStack
Answer: D
267.Ascend among the following options 310 And Shengteng 910

The attributes are the same?
A. Maximum power
B. Architecture
C. Integer precision
D. 7nm Process
Answer: B
268.The time interval in the Python language is a Hoating-point fraction in seconds

A. True
B. False
Answer: A
269.What are the common clustering algorithms?
44 / 61
A. Density clustering
D. Kmeans
Answer: A B C D
270.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: A B C D
271.When dealing with actual problems, when should machine learning be used in the following
situations?
A. The data distribution itself changes over time, requiring continuous re-adaptation of the program, such
as predicting the trend of merchandise sales
B. The complexity of the rules is low, and the problem is small
C. Task rules will change over time, such as defect detection on the production line
D. The rules are very complicated or cannot be described, such as face recognition and voice recognition
Answer: A C D
272.In neural networks, which of the following methods are used to update the parameters when training
the network to minimize the loss function?
Answer: D
273.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: B C
274.Which of the following are the characteristics of TensorFlow? (Multiple choice)

A. Open source
B. Convenience
C. Mature
D. Flexible
45 / 61
Answer: A B C D
275.TensorFlow 2.0 The way to view tensor dimensions in is?

A. dimens
B. dtype
C. ndim
D. device
Answer: C
276.When the universal text recognition service is successfully called, the returned text recognition result
is stored in which of the following fields?
A. text
B. result
C. content
D. words
Answer: B
277.Grid search is a method of parameter adjustment.

A. TRUE
B. FALSE
Answer: B
278.As shown in the figure below, what is the value of the determinant A?
A. 24
B. 18
C. -24
D. 0
Answer: D
279.During the two classification process, we can set any category as a positive example.
A. TRUE
B. FALSE
Answer: A
280.CPU Mainly through increasing instructions, increasing the number of cores, etc. to improve AI
performance.
A. TRUE
B. FALSE
Answer: A
46 / 61
281.PyTorch All existing versions support Python2 with Python3?

A. TRUE
B. FALSE
Answer: B
282.Atlas accelerate AI What processor is used for inference?

A. Different 910 processor
B. Different 310 processor
C. GPU
D. FPGA
Answer: B
283.Label when the image remake detection service is called successfully sugestion
What's the situation?
A. certainty
B. FALSE
C. uncertainty
D. TRUE
Answer: B C D
284.HiAI What are the advantages of mobile computing platforms?

A. Quick source code
B. Diverse tool chain
D. plentiful API
Answer: A B C D
285.The following applications that are not part of the adversarial generation network are?
A. Text generation
B. Image generation
C. Image Identification
D. Data enhancement
Answer: C
286.What is the most important difference between batch gradient descent, mini-batch gradient descent,
and stochastic gradient descent?
A. Gradient size
C. Learning rate
D. Number of samples used
Answer: D
287.What are the advantages of Python? (Multiple choice)

A. Simple
47 / 61
B. Free
C. High-level language
D. Rich library
Answer: A B C D
288.TensorFlow 2.0 stand by GPU The acceleration is mainly manifested in the parallel operation of the
addition, subtraction, multiplication and division of the matrix accelerate.
A. TRUE
B. FALSE
Answer: A
289.Generative confrontation networks are currently widely used.

The following scenarios can use this network?
A. Data enhancement
D. Image generation
Answer: A B C D
290.Among the machine learning algorithms, the following is not an integrated learning strategy?
A. Boosting
B. Stacking
C. Bagging
D. Marking
Answer: D
291.TensorF1ow2.0 Can be used to check if it is tensor The methods are?

A. dtype
B. isinstance
C. is_tensor
D. device
Answer: B C
292.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
293.Deep learning is different from machine learning and there are no unsupervised algorithms
A. True
B. False
48 / 61
Answer: B
294.What is the incorrect relationship between neural networks and deep learning?
A. The concept of deep learning sterns from the study or 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
295.Regarding the face search service, which of the following statements are correct?
A. When there is no face set, you need to create a face set first, then add face data, and then search
B. The size of a face set cannot exceed 10000 Pictures
C. There is a dedicated interface to delete the specified face set
D. There is a dedicated interface to delete the face data in a certain face set
Answer: A C D
296.Deep learning is a branch of machine learning

A. True
B. False
Answer: A
297.The main computing resources included in the Da Vinci architecture computing unit are?
B. Scalar Computing Unit
C. Tensor computing unit
D. Matrix calculation unit
Answer: A B D
298.Which of the following functions can numerically stabilize overflow and underflow?
A. Softminus function
B. Softplus function
C. Soft max function
D. Softmin func!Jon
Answer: C
299.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.:
A. The number of hidden layers is appropriately reduced, and the resolution ability of the neural network
remains unchanged
B. Increase the number of hidden layers appropriately, the stronger the resolution ability of the neural
network
C. The number of hidden layers is appropriately reduced, the stronger the resolution ability of the neural
network is
49 / 61
D. The number of hidden layers increases appropriately, the weaker the resolution ability of the neural
network B
300." print” in Python 3 must be used with “()”

A. True
B. False
Answer: A
301.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
302.During application development, which of the following operations is not a typical network definition?
A. Network definition
D. Weight initialization
Answer: C
303.The determinant of square matrix A is a scalar.

A. True
B. False
Answer: A
304.In the classic convolutional neural network model, SoftMax

What hidden layer does the function follow?
B. Pooling layer
D. All of the above
Answer: C
305.When using TensorFlow2. 0 of keras When building a neural network with an interface, the network
needs to be compiled.
Which of the following methods should be used to work?
A. compile
B. write
C. join
D. fit
Answer: A
50 / 61
306.Which of the following descriptions of the depth feedforward network is correct?

A. Deep feedforward network 1s 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 1s used to deal with linear problems.
Answer: A
307.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+ 8) T = AT +BT
C. (!EA) T = !EAT
D. (AB} T = A T +BT
Answer: A B C
308.TensorFlow is Google's first generation of proprietary machine learning systems

A. True
B. False
Answer: B
309.Image label service returned tag There can be more than one.
A. TRUE
B. FALSE
Answer: A
310.The following options belong to keras under estimator

The encapsulation methods are?
A. Assessment
B. training
C. prediction
D. Output model
Answer: A B C
311.With a lot of sales data but no labels, companies want to identify VIP Customer, the following model
Suitable?
B. SVM
C. K-Means
Answer: C D
312.What operating systems does Python support? (Multiple Choice)

A. DOS
B. Linux
51 / 61
C. Mac OSX
D. Windows
Answer: A B C D
313.The perceptron can be displayed in space as?

A. line
B. flat
C. Hyperplane
D. point
Answer: C
314.Regarding backpropagation, the following statement is wrong?

A. Backpropagation can only be used in feedforward neural networks
B. Backpropagation can be combined with gradient descent algorithm to update network weights
C. Backpropagation passes through the activation function
Answer: A
315.HUAWEI HiAI Which tool does the platform support to integrate with?
A. Jupyter Notebook
B. MyEclipse
C. Android Studio
D. Spider
Answer: C
316.The following belongs to TensorFlow2.0 The characteristic is?

D. Continue to be compatible Tensorflowl.x Related modules
Answer: A C D
317.HUAWEI CLOUD ModelArts Is for AI Which of the following functions are in the developed one-stop
development platform Mode1Arts Can have?
A. Data governance
B. AI market
C. Visual workflow
Answer: A B C D
318.Vector is a number.
A. True
B. False
Answer: B
52 / 61
319.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 1s fast and reliable [Right Answers}
D. Random gradient descent is one of the commonly used optimization algorithms in deep learning
algorithms.
Answer: C
320.TensorFlow It is one of the most popular deep learning frameworks.

A. TRUE
B. FALSE
Answer: A
321.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
322.Regularization is an important and effective technique for reducing generalization errors in traditional
machine learning. The following techniques are regular.
The technology is:
A. L1 Regularization
B. L2 Regularization
C. Dropout
D. Momentum optimizer
Answer: A B C
323.What is the English abbreviation for AI?

A. Automatic Intelligence
B. Artifical Intelligence
C. Automatic Information
D. Artifical Information
Answer: B
324.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: A B D
53 / 61
325.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
326.What research areas does artificial intelligence have? (Multiple choice)

B. Computer Vision
C. Machine learning
D. Speech recognition
Answer: A B C D
327.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. Minsky
B. Turing
C. Simon
D. Shannon
Answer: D
328.On-Device Execution, that is, the entire image is offloaded and executed, and the computing power
of the Yiteng chip can be fully utilized, which can greatly reduce the interaction overhead, thereby
increasing the accelerator occupancy rate.
On-Device The following description is wrong?
A. MindSpore Realize decentralized autonomy through adaptive graph optimization driven by gradient
data A11 Reduce, Gradient aggregation is in step, and calculation and communication are fully
streamlined
B. Challenges of model execution under super chip computing power: Memory wall problems, high
interaction overhead, and difficulty in data supply. Partly in Host Executed, partly in Device Execution,
interaction overhead is even much greater than execution overhead, resulting in low accelerator
occupancy
C. MindSpore Through the chip-oriented depth map optimization technology, the synchronization wait is
less, and the "data computing communication" is maximized. The parallelism of “trust”, compared with
training performance Host Side view scheduling method is flat
D. The challenge of distributed gradient aggregation under super chip computing power:ReslNet50 Single
iteration 20ms Time will be generated The synchronization overhead of heart control and the
communication overhead of frequent synchronization. Traditional methods require 3 Synchronization
completed A11 Reduce, Data-driven method autonomy A11 Reduce, No control overhead
Answer: C
329.Where should the labeled data be placed in the confrontation generation network?
A. As the output value of the generated model
54 / 61

D. As input value for generative model
Answer: B
330.Which of the following aspects belong to Huawei's full stack AI solution?

A. Ascend
B. CANN
C. ModelArts
D. MindSpore
Answer: A B C D
331.Code model.fit(mnist.train.inmage,mnist.train.labels.epochs=5)in of epochs Parameter

representative?
A. The entire training set will be trained 5 Times
B. The entire test set will be tested 5 Times
C. The entire training set will be divided into 6 Share
D. The entire training set will be divided into 5 Share
Answer: A
332.HUAWEI CLOUD EI There are many types of agents according to different industries and application
scenarios. Currently, Huawei Cloud EI The agent has?
A. Industrial Agent
B. Traffic Agent
C. Park Agent
D. Car intelligence
Answer: A B C D
333.Which of the following libraries are installed Atlas300 (3000) Install the accelerator card in the server
DDK (Device Development Kit) What needs to be installed before?
A. decorator
B. setuptools
C. numpy
D. tensorflow
Answer: A B C
334.The Python language can use the "#" at the beginning of a single line of code for code comments .
A. True
B. False
Answer: A
335.Which of the following description of the number of rows and columns of the matrix is correct?
A. The number of rows 1s greater than the number of columns
B. The number of rows 1s equal to the number of columns
55 / 61
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
336.Which of the following are AI Subfield?

A. Machine learning
B. Computer vision
C. Speech Recognition
Answer: A B C D
337.TensorFlow2.0 The mechanism of graphs and conversations has been cancelled in.
A. TRUE
B. FALSE
Answer: B
338.TensorFlow It is an end-to-end open source platform for machine learning and deep learning.
A. TRUE
B. FALSE
Answer: A
339.Which of the following are AI Application areas?

A. Wisdom education
B. Smart City
C. Smart home
D. Smart medical
Answer: A B C D
340.All convolution kernels of the same convolution layer in the convolutional neural network are
weight-sharing.
A. TRUE
B. FALSE
Answer: A
341.Which of the following statements about supervised learning is correct?

A. Decision tree is a supervised learning
B. Supervised learning cannot use cross-validation for training
C. Supervised learning is a rule-based algorithm
Answer: A
342.Which of the following is not an application of image recognition services?

A. Target Detection
56 / 61
C. Scene analysis
D. Speech synthesis
Answer: D
343.Which of the following are the topics of speech processing research?

A. Speech processing
B. Voiceprint recognition
C. Speech Recognition
D. Wake up
Answer: B C D
344.In polynomial regression, there is a square term in the formula of the model, so it is not linear.
A. TRUE
B. FALSE
Answer: B
345.As shown below, which are the following matrix A characteristic value? (Multiple choice)
A. 2
B. -2
C. 4
D. 4
Answer: A D
346.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 lo left defaults from -1
D. Index from right to left defaults from 0
Answer: A C
347.The three stages of artificial intelligence include computational intelligence, () , Cognitive intelligence.
B. Perceptual intelligence
C. Behavioral intelligence
Answer: B
348.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 d1cl1onary are unique and the data type 1s uniform
57 / 61
Answer: A B C D
349.Which of the following statements about Python are correct? (Multiple choice)
A. Invented m 1989 by the Dutch Guido van Rossum, the first public release was issued in 1991.
B. Python 1s 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 1ndentat1on [Right Answers)
D. Python 1s often nicknamed glue language which can easily connect various modules made 1n other
languages
Answer: A B C D
350.Which of the following capabilities does the genetic knowledge map have?
A. Auxiliary case diagnosis
B. Disease prediction and diagnosis
C. Gene test report generation
D. Entity query
Answer: A C D
351.The commonly used loss functions in deep learning are?

A. L1 Loss function
D. Self-declining loss function
Answer: B C
352.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. Relu function
B. Sigmoid function
C. tanh function
D. Softsign function
Answer: A
353.There are many types of neural networks in deep learning. The following neural network information
is one-way propagation:
A. LSTM
C. Recurrent neural network
D. GRU
Answer: B
354.Which of the following about the description of the number of rows and columns for the determinant is
58 / 61
correct?
A. The number of rows 1s greater than the number of columns
B. The number of rows 1s equal lo the number of columns
D. The number of rows has no relationship with the number of columns.
Answer: B
355.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 1s rounded to zero
C. Overflow occurs when a large number of levels are approximated as 1P or - 1P
D. Underflow occurs when a large number of levels are approximated as 1P or - 1P
Answer: A C
356.For the image classification problem, which of the following neural networks is more suitable to solve
this problem?
A. sensor
C. Convolutional Neural Network
D. Fully connected neural network
Answer: C
357.Deep learning algorithms can be divided into supervised learning and unsupervised learning.
A. True
B. False
Answer: A
358.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
359.The constituent elements of the Python identifier include: numbers, letters, and underscores.
A. True
B. False
Answer: A
360.Python tuples are identified by "()'' and internal elements are separated by ":".
A. True
B. False
59 / 61
Answer: B
361.A scalar k is multiplied by matrix A equal to k and each of the numbers in A is multiplied.
A. True
B. False
Answer: A
362.The label predicted by the regression algorithm is?

A. Self-deformation
B. Discrete
C. Strain type
D. Continuous
Answer: D
363.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
A. True
B. False
Answer: A
364.On Huawei Cloud EI Which of the following can be AI Technology is integrated into the application
scenarios of all walks of life AI The advantages of technology to improve efficiency and improve
experience.
A. EI Agent
B. OBS Object storage service
C. Cloud database
D. EI Big data service
Answer: A
365.Which command can be checked Atlas 300 (3000) Whether the accelerator card is in place?
A. 1spci | grep'npu'
B. 1spci | grep'd100'
C. 1spci grep'atlas'
D. atlas info
Answer: B
366.Not supported in the following options TensorFLow2.0 The attributes for dimensional transformation
are.
A. squeeze
B. reshape
C. gather
D. transpose
60 / 61
Answer: C
367.What is the performance of artificial intelligence in the stage of perceptual intelligence?

B. Machines begin to calculate and transmit information just like humans
C. The machine starts to understand and understand, make judgments, and take some simple actions
Answer: C
368.Recurrent neural network is different from convolutional neural network, it is better at solving the
following problems?
A. Sequence related issues
B. Image classification
C. Image detection
D. Recommended question
Answer: A
369.Cut off 2019 At the end of the year, TensorFlow

The released versions are?
A. 1
B. 3
C. 2
D. 0.1
Answer: A C D
61 / 61
Exam-Killer
http://www.exam-killer.com/
The best certification study material provider. 100% pass guarantee.
Instant Download - Best Exam Practice Material - 100% Money Back Guarantee!
IT Certification Guaranteed, The Easy Way!
Exam : H13-311_V3.0
Title : HCIA-AI V3.0 Exam
Vendor : Huawei
Version : DEMO
1
Get Latest & Valid h13-311_v3.0 Exam's Question and Answers from Exam-killer.com. 1
https://www.exam-killer.com/H13-311_V3.0-valid-questions.html
NO.1 TensorF1ow2.0 Can be used to check if it is tensor The methods are?

A. isinstance
B. device
C. is_tensor
D. dtype
Answer: A,C
NO.2 According to Huawei Cloud EI Intelligent platform, which of the following solutions can be
provided?
A. Vehicle identification scheme
B. Policy query scheme based on knowledge graph
C. Intrusion recognition scheme
D. Crowd statistics program
Answer: A,B,C,D
NO.3 Which of the following is not a face recognition technology?

A. Remake recognition
B. Face search
C. Face comparison
D. Face Detection
Answer: A
NO.4 Which of the following options is not a tensor attribute?

A. node (Node)
B. type (Type)
C. name (Logo)
D. shape (Dimensionality)
Answer: A
NO.5 If a model has a large deviation on the test set and a small variance, it means that the model?
A. Underfitting
B. May be overfitting may be underfitting
C. Just fit
D. Overfitting
Answer: A
NO.6 Feature is the dimension that describes the characteristics of the sample. Regarding its
interpretability in traditional machine learning and deep learning, the following statement is correct:
B. Features are weak in traditional machine learning, but strong in deep learning
C. Features are interpretable in traditional machine learning and deep learning
D. Features are weak in interpretability in traditional machine learning and deep learning
Answer: A
2
NO.7 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. Simon
B. Turing
C. Shannon
D. Minsky
Answer: C
NO.8 According to the definition of information entropy what is the bit entropy of throwing a
uniform coin?
A. 0
B. -1
C. 0.5
D. 1
Answer: D
NO.9 The Python dictionary is identified by "{}", and the internal data consists of the key and its
corresponding value.
A. True
B. False
Answer: A
NO.10 In random forest, what strategy is used to determine the outcome of the final ensemble
model?
B. Voting system
C. Cumulative system
D. Find the average
Answer: B,D
NO.11 Which of the following are the topics of speech processing research?
A. Speech processing
B. Wake up
C. Voiceprint recognition
D. Speech Recognition
Answer: B,C,D
NO.12 Which of the following description of the validation set is wrong?

A. Typically 80% of the training data 1s used for training and 20% 1s used for verification.
B. The verification set can coincide with the test set.
C. The test set can coincide with the training set
D. The subset used to pick hyperparameters is called a validation set
3
Answer: B
NO.13 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
B. TRUE
Answer: B
NO.14 PyTorch Which of the following functions does not have?

A. Automatic derivative
B. GPU accelerate
C. Support dynamic graph
D. Inline keras
Answer: D
NO.15 What are the commonly used activation functions? (Multiple Choice)
A. sigmoid
B. tanh
C. relu
D. danish
Answer: A,B,C
NO.16 Which of the following about the description of the number of rows and columns for the
determinant is correct?
A. The number of rows has no relationship with the number of columns.
B. The number of rows 1s equal lo the number of columns
D. The number of rows 1s greater than the number of columns
Answer: B
NO.17 Artificial intelligence at this stage is still in the weak artificial intelligence stage.
A. TRUE
B. FALSE
Answer: A
NO.18 About Bayesian formula- P(WlX)=P(XlW)*P(W)/P(X) What is the correct description?

A. P(X) is the posterior probability
B. P(WIX) is a prior probability
C. P(XIW) 1s a conditional probability
D. P(W) is the posterior probability
Answer: C
4
Huawei - H13-311_V3.0
Question #:1
Python language can use "#" at the beginning of a single line of code for code comments.
A. True
B. False
Answer: A
Question #:2
What are the steps that do not belong to the operation of Python file objects?
A. open
B. delete
C. read
D. write
Answer: B
Question #:3
Information theory is a branch of applied mathematics, the main research is to quantify how much information
a signal contains. Who first proposed the information entropy?
A. Minsky
B. Turing
C. Simon
D. Shannon
Answer: D
Question #:4
What is the following description about the positive definiteness of matrices?
A. a matrix whose all eigenvalues are positive is called positive definite
B. the matrix with all eigenvalues being non-negative is called semi-positive definite
C.
1 of 48
Huawei - H13-311_V3.0
C. the matrix with all eigenvalues being negative is called negative definite
D. and the matrix with all eigenvalues being non-positive is called semi-negative definite
Answer: A B C D
Question #:5
Which of the following options is not how the TensorFlow program reads data?
A. preload data (Preload)
B. feed data (Feeding)
C. read from the file
D. write a file format reader
Answer: D
Question #:6
Recurrent neural networks are more suitable for processing image recognition problems.
A. True
B. False
Answer: B
Question #:7
Which of the following options are concepts in TensorFlow?
A. tensor
B. variable
C. placeholder
D. operation
E. conversation
Answer: A B C D E
Question #:8
2 of 48
Huawei - H13-311_V3.0
A scalar k multiplied by matrix A is equal to k multiplied by each number in A.
A. True
B. False
Answer: A
Question #:9
Print in Python 3 must be used with ().
A. True
B. False
Answer: A
Question #:10
Artificial intelligence is a new technical science that studies and develops theories, methods, and application
systems for simulating, extending, and expanding human intelligence. It is one of the core research areas of
machine
learning.
A. True
B. False
Answer: A
Question #:11
Which of the following options is not the reason why traditional machine learning algorithms promote the
development of deep learning?
A. dimension disaster
B. local invariance and smooth regularization
C. manifold learning
D. feature engineering
Answer: D
3 of 48
Huawei - H13-311_V3.0
Question #:12
The number or matrix stored in the tensor.
A. True
B. False
Answer: B
Question #:13
The Python script execution model includes interactive mode and script mode.
A. True
B. False
Answer: A
Question #:14
What is the following description about unsupervised learning?
A. Unsupervised learning only deals with "features", not "tags"
B. Dimensionality reduction algorithms are not classified as unsupervised learning
C. K-means algorithm and SVM algorithm are all unsupervised learning
D. none of the above
Answer: D
Question #:15
X and Y are random variables and C is a constant. What is incorrect about the nature of mathematical
expectations in the following options?
A. E(C)=C
B. E(X+Y)=E(X)+E(Y)
C. E(CX)=CE(X)
D. E(XY)=E(X)E(Y )
4 of 48
Huawei - H13-311_V3.0
Answer: D
Question #:16
According to the definition of information entropy, what is the information entropy of tossing a uniform coin?
A. 0
B. 0.5
C. 1
D. -1
Answer: C
Question #:17
What are the advantages of Python?
A. simple
B. free
C. high-level language
D. rich library
Answer: A B C D
Question #:18
Functions are organized, non-reusable, and are used to implement a single, or related function code segment.
A. True
B. False
Answer: B
Question #:19
Which of the following options is not a tensor attribute?
A. name (identification)
B. shape
5 of 48
Huawei - H13-311_V3.0
C. type
D. node
Answer: D
Question #:20
What is wrong about hyperparameter description?
A. hyperparameters are the parameters set before the algorithm starts learning
B. most machine learning algorithms have hyperparameters
C. hyperparameters cannot be modified
D. a lot of values of hyperparameters are learned through the algorithm itself
Answer: C
Question #:21
The neural network training process will generate a lot of data. What mechanism does TensorFlow use to
avoid the problem of too much input data?
A. Client
B. feed
C. placeholder
D. fetch
Answer: C
Question #:22
Python tuples are identified by "()", and internal elements are separated by ";".
A. True
B. False
Answer: B
Question #:23
6 of 48
Huawei - H13-311_V3.0
Deep learning is different from machine learning, there is no unsupervised algorithm.
A. True
B. False
Answer: B
Question #:24
There are many business applications of machine learning services. What are the common businesses covered?
A. financial product recommendation
B. predictive maintenance
C. retention of telecommunications customers
D. retailer grouping
Answer: A B C D
Question #:25
TensorFlow only supports Python and R languages.
A. True
B. False
Answer: B
Question #:26
What are the regularization in deep learning?
A. L1 norm L2 norm
B. data set enhancement
C. integration method
D. Dropout
Answer: A C D
7 of 48
Huawei - H13-311_V3.0
Question #:27
What are the characteristics of Python code?
A. legibility
B. conciseness
C. fastness
D. scalability
Answer: A B C D
Question #:28
What is not a deep learning algorithm?
A. self-encoder
B. convolutional neural network
C. recurrent neural network
D. support vector machine
Answer: D
Question #:29
What is the correct description of the relationship between frequency and probability?
A. frequency equals probability
B. frequency is less than probability
C. frequency is greater than probability
D. when the number of repeated tests is gradually increased, the frequency shows stability, close to the
probability.
Answer: D
Question #:30
What is wrong about backpropagation?
A. The learning process of the backpropagation algorithm consists of the forward propagation process and
8 of 48
Huawei - H13-311_V3.0
A.
the back propagation process
B. Back propagation algorithm is suitable for a learning algorithm of multi-layer neural network, which is
based on the gradient descent method
C. Back-propagation stage sends training input to the network to obtain stimulus response
D. Backpropagation algorithm mainly consists of two links (incentive propagation, weight update)
iteratively iterate until the target range of the network's response to the input
Answer: C
Question #:31
What is correct about Python?
A. invented by the Dutch Guido van Rossum in 1989, the first public release was released in 1991
B. Python is purely free software, the source code follows the GPL (GNU General Public License)
agreement
C. Python syntax is concise and clear, One of the features is to force the use of white space characters as
indentation
D. Python is often nicknamed the glue language, which can easily connect various modules made in other
languages.
Answer: A B C D
Question #:32
TensorFlow is Google's first generation of proprietary machine learning system.
A. True
B. False
Answer: B
Question #:33
HUAWEI CLOUD EI builds an enterprise intelligent business based on the three-tier service, which of the
following options does not belong to the three-tier service.
A. basic platform services
B. general domain services
C.
9 of 48
Huawei - H13-311_V3.0
C. industry domain services
D. integration services
Answer: D
Question #:34
the vector is a number.
A. True
B. False
Answer: B
Question #:35
Which of the following options are characteristic of TensorFlow?
A. open source
B. convenient
C. mature
D. flexible
Answer: A B C D
Question #:36
What is wrong with the description of the lambda function in Python?
A. lambda is just an expression, the function body is much simpler than def
B. The body of lambda can be an expression or a code block
C. lambda function can access parameters in the global namespace
D. lambda function only accepts one parameter value
Answer: B
Question #:37
The optimization method that does not belong to deep learning is .
10 of 48
Huawei - H13-311_V3.0
A. stochastic gradient descent
B. back propagation algorithm
C. principal component analysis
D. momentum
Answer: C
Question #:38
the for loop statement in the Python language can traverse the items in any sequence.
A. True
B. False
Answer: A
Question #:39
Which of the following functions can numerically stabilize overflow and underflow?
A. Softminux function
B. Softplus function
C. Softmax function
D. Softmin function
Answer: C
Question #:40
Which of the following does not belong to the AI school?
A. symbolism
B. statism
C. behaviorism
D. connectionism
Answer: B
11 of 48
Huawei - H13-311_V3.0
Question #:41
Deep learning is a branch of machine learning.
A. True
B. False
Answer: A
Question #:42
The Python dictionary is identified by "{}", and the internal data consists of the key and its corresponding
value.
A. True
B. False
Answer: A
Question #:150
Deep learning algorithms can be divided into supervised learning and unsupervised learning.
A. True
B. False
Answer: B
Question #:44
What does not belong to supervised learning below?
A. logistic regression
B. Support Vector Machine
C. Decision Tree
D. Principal component analysis
Answer: D
12 of 48
Huawei - H13-311_V3.0
Question #:143
The following statements about overflow and underflow in numerical calculations are correct?
A. underflow occurs when numbers close to zero are rounded to zero
B. An overflow occurs when the number close to zero is rounded to zero
C. Overflow occurs when a large number of levels are approximated to positive infinity or negative infinity
D. Underflow occurs when a large number of levels are approximated to positive infinity or negative
infinity
Answer: A C
Question #:46
What is the correct description about the Python creation function?
A. the created function starts with the def keyword, followed by the function name and parentheses
B. parameters need to be placed in parentheses
C. the content of the function starts with a colon and needs to be indented by
D. use return to return the result, and the function ends
Answer: A B C
Question #:47
What is correct about the dictionary description in Python?
A. each key and its corresponding value must be separated by ";"
B. and different key-value pairs are separated by ","
C. the entire dictionary is included in "{}"
D. the keys of the dictionary are unique and the data types are unified
Answer: A B C
Question #:48
Which of the following is not a specific technology of artificial intelligence?
A. knowledge graph
13 of 48
Huawei - H13-311_V3.0
B. Machine translation
C. Riemannian geometry
D. Semantic understanding
Answer: C
Question #:49
What is not the difference between Python 2 and Python 3?
A. print
B. Unicode
C. import
D. xrange
Answer: C
Question #:50
What are the following conditions that random trials need to meet?
A. can be executed repeatedly under the same conditions
B. There is more than one possible result for each test, and all possible results of the test 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 an experiment
Answer: A B D
Question #:51
What operating systems does Python support?
A. DOS
B. Linux
C. Mac OSX
D.
14 of 48
Huawei - H13-311_V3.0
D. Windows
Answer: A B C D
Question #:52
In TensorFlow, data is expressed in the form of tensors and computation graphs.
A. True
B. False
Answer: A
Question #:53
GAN is a deep learning model and one of the most promising methods for unsupervised learning on complex
distributions in recent years.
A. True
B. False
Answer: A
Question #:54
As shown in the figure below, find the value of determinant A.
|1 2 3|
A = |4 5 6|
|7 8 9|
A. 24
B. 18
C. -24
D. 0
Answer: D
Question #:55
15 of 48
Huawei - H13-311_V3.0
What is not a deep learning development framework?
A. CNTK
B. Keras
C. BAFA
D. MXNet
Answer: C
Question #:56
X and Y are random variables, and C is a constant. What is wrong about the nature of variance in the
following options?
A. D(C)=0
B. D(X+Y)=D(X)+D(Y)
C. D(CX)=C*C*D(X)
D. D(XY)=D( X)D(Y)
Answer: B D
Question #:57
what is not a Python operator?
A. arithmetic operator
B. inference operator
C. logical operator
D. comparison operator
Answer: B
Question #:58
Timestamps in the Python language are expressed in terms of how long (seconds) elapsed from midnight
(epoch) on January 1, 1970.
A. True
B.
16 of 48
Huawei - H13-311_V3.0
B. False
Answer: A
Question #:59
What algorithms does TensorFlow support?
A. CNN
B. ZNN
C. RNN
D. HUT
Answer: A B C D
Question #:60
Which of the following are cloud services provided by Huawei Cloud EI Visual Cognition?
A. text recognition
B. face recognition
C. image recognition
D. content detection
E. image processing
Answer: A B C D E
Question #:61
Which option does the convolutional neural network structure not include?
A. convolutional layer
B. pooling layer
C. circulation layer
D. fully connected layer
Answer: C
17 of 48
Huawei - H13-311_V3.0
Question #:62
Which of the following are solutions for Huawei Cloud EI industry scenarios?
A. Intelligent Logistics
B. Intelligent Water Affairs
C. Intelligent Transportation
D. Intelligent Finance
E. Intelligent Manufacturing
Answer: A B C D E
Question #:63
Through deep learning, it is easy to derive simple mathematical functions from a large amount of
high-dimensional data to describe complex decision interfaces.
A. True
B. False
Answer: A
Question #:64
What is the English abbreviation for AI?
A. Automatic Intelligence
B. Artifical Intelligence
C. Automatic Information
D. Artifical Information
Answer: B
Question #:65
Neural network research belongs to which of the following schools?
A. symbolism
B.
18 of 48
Huawei - H13-311_V3.0
B. connectionism
C. behaviorism
D. neither
Answer: C
Question #:66
What are the commonly used loss functions?
A. mean square error
B. Sigmoid cross entropy
C. Softmax cross entropy
D. Sparse cross entropy
E. weighted Sigmoid cross entropy
Answer: A B C D E
Question #:67
Deep learning is different from machine learning, there is no unsupervised algorithm.
A. True
B. False
Answer: B
Question #:68
Training error will reduce the accuracy of the model, resulting in under-fitting, how to improve the model fit at
this time?
A. increase the amount of data
B. feature engineering
C. reduce regularization parameter
D. increase features
Answer: A B D
19 of 48
Huawei - H13-311_V3.0
Question #:69
What is correct about the index description of Python?
A. left-to-right index starts from 0 by default
B. left-to-right index starts from 1 by default
C. right-to-left index starts from -1 by default
D. right-to-left index starts from 0 by default
Answer: A C
Question #:70
What are the scenarios or industries suitable for Python?
A. Artificial Intelligence
B. Web development
C. game development
D. Hardware development
Answer: A B C
Question #:71
The correlation coefficient is also called the linear correlation coefficient, which is used to measure the linear
relationship between two variables, and its value is a real number greater than 0.
A. True
B. False
Answer: B
Question #:72
In a large-scale relationship analysis scenario, which of the following options does not belong to the three high
demands of massive relationship processing?
A. need for efficient relationship discovery of massive data
B.
20 of 48
Huawei - H13-311_V3.0
B. The demand for efficient sharing of massive data
C. Requirements for efficient storage and access to massive amounts of data
D. High scalability and high availability for the relationship analysis platform
Answer: A
Question #:73
What is the correct description of the row number and column number of the determinant?
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 is not related to the number of columns
Answer: B
Question #:74
Which method is not supported to define variables in TensorFlow?
A. random number
B. constant
C. is calculated from the initial values of other variables
D. null
Answer: D
Question #:75
the trace operation returns the sum of the diagonal elements of the matrix. Therefore, the traces of matrix A
and its transposed matrix are equal.
A. True
B. False
Answer: A
21 of 48
Huawei - H13-311_V3.0
Question #:76
What are the return results of if conditional statements in Python?
A. TRUE
B. 0
C. FALSE
D. null
Answer: A C
Question #:77
What is wrong with the following gradient descent statement?
A. Stochastic gradient descent is a common type of gradient descent
B. Gradient descent includes random gradient descent and batch gradient descent
C. The gradient descent algorithm is fast and reliable
D. Stochastic gradient descent is one of the optimization algorithms commonly used in deep learning
algorithms
Answer: C
Question #:78
Among the following options, what is the philosophy of Python language design?
A. Beautiful
B. Expensive
C. Explicit
D. Simple
Answer: A C D
Question #:79
The history of robots is not long. In 1959, the United States Engelberg and Dvor made the first generation of
industrial robots in the world. The history of robots really began. According to the development process of
robots, it is
22 of 48
Huawei - H13-311_V3.0
usually divided into three generations. What are they?
A. teaching reproducible robot
B. robot
C. with feeling, robot
D. who thinks, intelligent robot
Answer: A B D
Question #:80
Linear algebra is a branch of algebra that deals mainly with linear problems. The linear relationship is that the
relationship between the exponential objects is expressed in the form of times.
A. 1
B. 2
C. 3
D. 4
Answer: A
Question #:81
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. This
research branch is called?
A. Expert system
B. Machine learning
C. Neural network
Answer: B
Question #:82
23 of 48
Huawei - H13-311_V3.0
What are the quotation marks that can be used in the Python language?
A. single quote
B. double quote
C. triple quote
D. quad quote
Answer: A B C
Question #:83
What is wrong about the 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. and a single-layer neural network is also a kind of deep learning
D. convolutional neural network is a type of deep learning
Answer: C
Question #:84
Which of the following functions does Python support?
A. inherits
B. overloads
C. derived
D. Multiple inheritance
Answer: A C D
Question #:85
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
24 of 48
Huawei - H13-311_V3.0
Answer: A
Question #:86
Python lists can be identified by "[]", and the default index of the first element from left to right is 1.
A. True
B. False
Answer: B
Question #:87
What is incorrect about the following description of the distribution function, distribution law, and density
function of random variables?
A. The distribution law can only describe the law of the value of discrete random variables
B. the density function can only describe the value of continuous random variables
C. discrete random variables have no distribution function
D. distribution function describes the law of random variables
Answer: C
Question #:88
When was the first time that "artificial intelligence" was proposed?
A. 1946
B. 1960
C. 1916
D. 1956
Answer: D
Question #:89
What is wrong about expectations and variances in the following options?
A. expect to reflect the average level of random variable values
25 of 48
Huawei - H13-311_V3.0
B. Variance reflects the high degree of random variable and its mathematical expectation
C. Expectation and variance are both numerical characteristics of random variable
D. the greater the expectation, the smaller the variance
Answer: D
Question #:90
What is not a function of the re regular expression module in the Python language?
A. search search
B. match match
C. update update
D. compile complie
Answer: C
Question #:91
Numerical calculation: Refers to the method and process of using digital computers to find approximate
solutions to mathematical problems, as well as disciplines composed of related theories. Which of the
following processes are
involved in solving practical problems with computers?
A. Recognize practical problems
B. Mathematical model
C. numerical calculation method
D. program design
E. computer calculation results
Answer: A B C D E
Question #:92
Which of the following neural network structures will share weights? (Multiple Choice)
A. Convolutional Neural Network
B.
26 of 48
Huawei - H13-311_V3.0
B. recurrent neural network
C. fully connected neural network
D. all of the above
Answer: A B
Question #:93
What is not an automatic hyperparameter optimization algorithm?
A. grid search
B. stochastic gradient descent
C. random search
D. model-based hyperparameter optimization
Answer: C
Question #:94
PEP 8 stipulates how many spaces are needed to indent the first line of Python?
A. 1
B. 2
C. 4
D. 8
Answer: C
Question #:95
the time interval in the Python language is a floating-point decimal in seconds.
A. True
B. False
Answer: A
Question #:96
27 of 48
Huawei - H13-311_V3.0
The Python language can use multiple statements on the same line, separated by commas ",".
A. True
B. False
Answer: B
Question #:97
As shown in the figure below, which of the following options are the eigenvalues of matrix A?
A. 2
B. -2
C. -4
D. 4
Answer: A D
Question #:98
TensorFlow is the second-generation artificial intelligence learning system developed by Google based on .
A. DistBelief
B. PaleyFunction
C. ConvexOne
D. Infinity
Answer: A
Question #:99
Huawei firmly believes that the value of AI is ultimately reflected in solving the actual business problems of
enterprises. Therefore, in the development of AI technology, Huawei firmly relies on domain solutions as the
traction and focuses.
A.
28 of 48
Huawei - H13-311_V3.0
A. True
B. False
Answer: A
Question #:100
The components of the Python identifier include: numbers, letters, and underscores.
A. True
B. False
Answer: A
Question #:101
Does the cyclic neural network use scenario not include?
A. machine translation
B. speech recognition
C. image style transfer
D. text generation
Answer: C
Question #:102
Which of the following options can improve the calculation efficiency of the neural network model?
A. TPU
B. GPU
C. FPGA
D. large-scale distributed cluster
Answer: A B C D
Question #:103
Which of the following options is not TensorFlow's build process?
29 of 48
Huawei - H13-311_V3.0
A. construct a calculation graph
B. input a tensor
C. generate a session
D. and update the weights
Answer: D
Question #:104
Machine learning service is a data mining and analysis platform service that helps users quickly discover data
laws and build predictive models through machine learning technology, and deploy them as predictive analysis
solutions.
A. True
B. False
Answer: A
Question #:105
Which of the following options correctly describes the relationship between the Poisson distribution and the
binomial distribution?
A. The mathematical model of Poisson distribution and Binomial distribution are both Lagrangian
probability
B. Poisson distribution and binomial distribution when n is very large and p is very small
C. Poisson distribution is not related to the binomial distribution
D. Poisson distribution can replace binomial distribution
Answer: B
Question #:106
Which of the following options is TensorFlow's calculation model?
A. calculation graph
B. tensor
C. session
30 of 48
Huawei - H13-311_V3.0
D. variable
Answer: A
Question #:107
what are the Python language data types?
A. number (number)
B. string (string)
C. list (list)
D. tuple (tuple)
E. dictionary (dictionary)
Answer: A B C D E
Question #:108
What is the model that does not belong to the recurrent neural network?
A. RNN
B. LSTM
C. GBDT
D. GRU
Answer: C
Question #:109
TensorFlow supports multi-TPU cluster computing. Correct 5. What is the description about the deep
feedforward network?
A. deep feedforward network is a kind of neural network
B. The deep feedforward network has only one hidden layer.
C. there are generally countless units on the hidden layer of the deep feedforward network
D. the deep feedforward network is used to deal with linear problems
Answer: A
31 of 48
Huawei - H13-311_V3.0
Question #:110
What is the correct description about Python tuples?
A. the difference between lists and tuples is that the elements of tuples cannot be modified
B. square brackets are used for lists, and parentheses are used for tuples
C. the creation of the tuple is very simple, just add the element in the brackets, and use commas to separate
D. when the tuple contains only one element, you need to add a comma after the element
Answer: A B C D
Question #:111
Artificial intelligence is currently more suitable for scenarios with "known environment, clear goals, and
predictable actions." Deep learning In the fields of image recognition, speech recognition, translation, etc.,
artificial intelligence
basically has human recognition capabilities, and even surpasses humans. Based on these capabilities, it has
been applied to many scenarios, such as medical treatment and public safety. But it is still lacking in reasoning
and
cognition.
A. True
B. False
Answer: A
Question #:112
Deep learning is a branch of machine learning.
A. True
B. False
Answer: A
Question #:113
What are the implementation modes of TensorFlow?
A.
32 of 48
Huawei - H13-311_V3.0
A. stand-alone mode
B. distributed mode
C. reverse mode
D. Forward mode
Answer: C D
Question #:114
TensorFlow framework does not support Windows systems.
A. True
B. False
Answer: B
Question #:115
What is incorrect about the following matrix?
A. Multiplying any matrix with the identity matrix will not change
B. The symmetric matrix A is transposed or A itself
C. The orthogonal matrix A's transpose is equal to the inverse of A
D. diagonal matrix does not exist inverse matrix
Answer: D
Question #:116
What are the core concepts in TensorFlow?
A. grid
B. calculation graph
C. dot product
D. tensor
Answer: B D
33 of 48
Huawei - H13-311_V3.0
Question #:117
The matrix generated by exchanging rows and columns of matrix A is called transpose of A. What are the
correct properties of matrix transpose? (ABC)
A. matrix A The transpose of the transpose is still A itself
B. The transpose of the sum of matrix A and matrix B is equal to the transpose of A plus the transpose of B
C. The transpose of a number times the product of A equals the transpose of this number times A
D. The transpose of the product of matrix A and matrix B is equal to the transpose of matrix A times the
transpose of matrix B
Answer: A B C
Question #:118
In a neural network, knowing the weight and deviation 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. randomly assigned values, pray that they are correct
B. Search the combination of weight and bias until the best value is obtained
C. gives an initial value, by checking the difference with the best value, and then iteratively update the
weight
D. is not correct
Answer: C
Question #:119
In which year did Huawei formally provide services as a cloud service, and cooperate with more partners to
provide richer artificial intelligence practices?
A. 2002
B. 2013
C. 2015
D. 2017
34 of 48
Huawei - H13-311_V3.0
Answer: D
Question #:120
The matrix generated by exchanging rows and columns of matrix A is called transpose of A. What are the
correct properties of matrix transpose?
A. B. (A+B) T = A T + B T
B. C. (AA) T - AA T
C. D. (AB) T = A T B T
Answer: A B C
Question #:159
Which of the following environments does not support the installation of TensorFlow?
A. Linux
B. Mac OS
C. Docker
D. OpenStack
Answer: D
Question #:122
What is the correct description about recurrent neural network (RNN)?
A. can be used to process sequence data
B. can not handle variable length sequence data
C. unlike the convolutional audit network, the parameters of the RNN cannot be shared
D. the units above the hidden layer are not related to each other
Answer: A
Question #:123
TensorFlow's Operations and Conmputation Graph do not necessarily run in Session.
A.
35 of 48
Huawei - H13-311_V3.0
A. True
B. False
Answer: B
Question #:124
The author of Python deliberately designed a very restrictive syntax, so that bad programming habits (such as
if the next line of the if statement is not indented to the right) cannot be compiled.
A. True
B. False
Answer: B
Question #:125
the description of the verification set is wrong.
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 select hyperparameters is called the validation set
D. Usually 80% of the training data is used for training and 20% is used for verification
Answer: C
Question #:126
What are the research areas of artificial intelligence?
A. natural language processing
B. computer vision
C. machine learning
D. speech recognition
Answer: A B C D
Question #:127
36 of 48
Huawei - H13-311_V3.0
The meaning of artificial intelligence was first proposed by a scientist in 1950, and at the same time proposed
a test model of machine intelligence. What is the scientist?
A. Minsky
B. Zade
C. Turing
D. Von Neumann
Answer: C
Question #:128
What is not a long-short term memory (LSTM) network structure?
A. memory gate
B. forget gate
C. input gate
D. output gate
Answer: A
Question #:129
Narrowing the gap between training error and test error will produce overfitting, how to prevent overfitting?
A. cross-validation
B. Integration method
C. Increase regularization
D. Feature engineering
Answer: A B
Question #:130
What is wrong with the introduction of Python module time?
A. import time
B. from time import*
37 of 48
Huawei - H13-311_V3.0
C. import time as t
D. from time
Answer: D
Question #:131
Vector group a1=(1,1,1),a2=(0,2,5),a3=(1,3,6) Which of the following options is correct?
A. linearly related
B. linearly independent
C. a1+a2+a3=0
D. 2a1+a2+a3=0
Answer: C D
Question #:132
What are the commonly used gradient descent optimization functions?
A. stochastic gradient descent
B. Adadelta
C. Adagrad
D. momentum
E. RMSProp
Answer: A B C D E
Question #:133
What are the commonly used activation functions?
A. sigmod
B. tainh
C. relu
D. danish
38 of 48
Huawei - H13-311_V3.0
Answer: A B C
Question #:134
What is incorrect about the description of Python regular expressions?
A. Python has added the re regular expression module since version 1.5
B. re regular expression module makes the Python language have all the regular expression functions
C. re regular expression is a powerful tool for processing strings
D. re regular expression can only handle string data, not numeric data
Answer: D
Question #:135
Principal Component Analysis (PCA) is a statistical method. Through orthogonal transformation, a set of
variables that may have correlation is converted into a set of linearly related variables. The converted set of
variables is called
the principal component.
A. True
B. False
Answer: B
Question #:136
What conditions must be met for m rows and n columns of matrix A and p rows and q columns of matrix B to
be multiplied?
A. m=p, n=q
B. n=p
C. m=n
D. p=q
Answer: B
Question #:137
39 of 48
Huawei - H13-311_V3.0
The correlation coefficient is also called the linear correlation coefficient, which is used to measure the linear
relationship between two variables, and its value is a real number greater than 0.
A. True
B. False
Answer: B
Question #:138
What are the application scenarios of break statement in Python language?
A. any Python statement
B. while loop statement
C. for loop statement
D. nested loop statement
Answer: B C D
Question #:139
Python is a completely object-oriented language. Which of the following options are Python objects?
A. function
B. module
C. number
D. string
Answer: A B C D
Question #:140
In May 1997, the famous "Human Machine Battle" final computer defeated the world chess king Kasparov
with a total score of 3.5 to 2.5. This computer is called?
A. dark blue
B. dark green
C. deep thinking
D.
40 of 48
Huawei - H13-311_V3.0
D. blue sky
Answer: A
Question #:141
The functions commonly used in mathematical operations in Python are basically in the math and cmath
modules.
A. True
B. False
Answer: A
Question #:142
What are the service categories included in Huawei Cloud EI Enterprise Intelligence?
A. EI visual cognition
B. EI online game
C. EI speech semantic
D. EI industry scenario
Answer: A C D
Question #:143
The following statements about overflow and underflow in numerical calculations are correct?
A. underflow occurs when numbers close to zero are rounded to zero
B. An overflow occurs when the number close to zero is rounded to zero
C. Overflow occurs when a large number of levels are approximated to positive infinity or negative infinity
D. Underflow occurs when a large number of levels are approximated to positive infinity or negative
infinity
Answer: A C
Question #:144
compared to recurrent neural networks, convolutional neural networks are more suitable for image recognition
41 of 48
Huawei - H13-311_V3.0
problems.
A. True
B. False
Answer: A
Question #:145
TensorFlow supports multi-TPU cluster computing.
A. True
B. False
Answer: A
Question #:146
What is the correct description about the Python program? (Multiple Choice)
A. the program is composed of modules
B. the module contains the statement
C. and the statement contains the expression
D. the expression builds and processes the object
Answer: A B C D
Question #:147
Regarding Bayesian formula: P(W|X)=P(X|W)*P(W)/P(X) What is the correct description?
A. P(W|X) is the prior probability
B. P(X|W) is the conditional probability
C. P(W) is the posterior probability
D. P(X) is the posterior probability
Answer: B
Question #:148
42 of 48
Huawei - H13-311_V3.0
Convolutional neural networks are more suitable for handling speech recognition problems.
A. True
B. False
Answer: B
Question #:149
The number of rows and columns of the three matrices A, B, and C are 3 rows, 2 columns, 2 rows, 3 columns,
and 3 rows, 3 columns. Which of the following operations makes sense?
A. AC
B. BC
C. A+B
D. AB=BC
Answer: B
Question #:150
Deep learning algorithms can be divided into supervised learning and unsupervised learning.
A. True
B. False
Answer: B
Question #:151
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
43 of 48
Huawei - H13-311_V3.0
Question #:152
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. linearly independent
C. a1+a2+a3=0
D. 2a1+a2+a3=0
Answer: C
Question #:153
What are the application scenarios of TensorFlow?
A. voice recognition
B. face recognition
C. image style change
D. automatic driving
Answer: A B C D
Question #:154
What are convolutional neural networks (CNN)?
A. VGGNet
B. ResNet
C. AlexNet
D. GoogleNet
Answer: A B C D
Question #:155
The factors that promote the development of artificial intelligence do not include?
A. big data
B.
44 of 48
Huawei - H13-311_V3.0
B. computing power
C. algorithm theory
D. blockchain
Answer: D
Question #:156
What is correct about the description of Python tuples?
A. the difference between lists and tuples is that the elements of tuples cannot be modified
B. square brackets are used for lists, and parentheses are used for tuples
C. the creation of the tuple is very simple, just add the element in the brackets, and use commas to separate
D. when the tuple contains only one element, you need to add a comma after the element
Answer: A B C D
Question #:157
Python regular expressions are a special sequence of characters that can easily check whether a string matches
a certain pattern.
A. True
B. False
Answer: A
Question #:158
What are the elements of artificial intelligence?
A. algorithm
B. computing power
C. data
D. scene
Answer: A B C D
45 of 48
Huawei - H13-311_V3.0
Question #:159
Which of the following environments does not support the installation of TensorFlow?
A. Linux
B. Mac OS
C. Docker
D. OpenStack
Answer: D
Question #:160
A and B can only be added if the number of rows and columns of matrix A and matrix B are the same.
A. True
B. False
Answer: A
Question #:161
The autoencoder is an unsupervised learning algorithm.
A. True
B. False
Answer: A
Question #:162
Which of the following conditions is not a condition to be satisfied by the n-fold Bernoulli test?
A. each test is repeated under the same conditions
B. There are only two possible outcomes per trial, that is, event A occurs and event A does not occur
C. each test itself must follow a normal distribution
D. the results of each test are independent of each other
Answer: C
46 of 48
Huawei - H13-311_V3.0
Question #:163
Which of the following options is not the central content of linear algebra?
A. probability theory
B. linear transformation
C. matrix theory
D. vector space
Answer: A
Question #:164
What is wrong about the image recognition cloud server in the following options?
A. service can identify specific objects such as skyscrapers, buildings and ships
B. service can also identify scenes such as sea, city and port
C. service can perform semantic analysis on the content in the picture to extract specific abstract tags
D. service can provide object and scene labels, and can also provide more concept labels
Answer: D
Question #:165
Which of the following data types does TensorFlow not support?
A. int8
B. float32
C. double64
D. bool
Answer: C
Question #:166
What are the characteristics of the Python language?
A. explanatory
47 of 48
Huawei - H13-311_V3.0
B. process-oriented
C. object-oriented
D. dynamic data type
Answer: A C D
Question #:167
the determinant of square matrix A is a scalar.
A. True
B. False
Answer: A
48 of 48
1. Which of the following are sub-fields of AI? ----------------------------------- ABCD A. Ascend
A. Machine learning B. CANN
B. Computer Vision C. ModelArts
C. Speech recognition D. MindSpore
D. Natural language processing (Single choice) 7. According to the Society of Automotive Engineers (SAE), autonomous driving is divided
into
(Judgment) 2. Voice recognition refers to recognizing audio data as text data. ------------------------- A
A.L1~L4
A.TRUE
B.L1~L5
B.FALSE
C.L0~L4
(Single choice) 3. The three stages of artificial intelligence include computational intelligence, () and
cognitive intelligence. ----------------- B D.L0~L5
B. Perceptual Intelligence (Judgment) 8. Computer vision is the science of studying how to make computers "see". ---------------------
----A
C. Behavioral Intelligence
(Judgment) 10. Jobs that are highly repetitive and require weak social skills are the jobs most easily
replaced by AI. ------------ A
(Single choice) 4. Which of the following options does not belong to the scope of Huawei's full-stack
A.TRUE
solution? ------------------------ B
B.FALSE
(Judgment) 11.Huawei's AI scenarios include public cloud, private cloud, various edge computing, IoT
B. Edge computing
industry terminals, and consumer
The deployment environment of terminal, edge, and cloud such as fee-based terminals. ----------------------
D. Chip enable ----------- A
(Multiple choice) 5. What are the elements of artificial intelligence? -------------------------------------- - ABCD A.TRUE
A. Algorithm B.FALSE
B. Scene (Single choice) 12. What is the performance of artificial intelligence in the perceptual intelligence stage?
--------------------------- C – capable listening, thinkin
C. Computing power
D. Data
B. Machines begin to calculate and transmit information like humans
(Judgment) 6. In the neural network based on connectionism, each node can express a specific meaning. C. The machine starts to understand and understand, make judgments, and take some
--------- B à A simple actions
A.TRUE
(Judgment) 13. Under the premise of ensuring data privacy and security, federated learning uses
B.FALSE
different data sources to cooperatively train models to make progress and break through the bottleneck
9. Which of the following aspects belong to Huawei's full-stack AI solution?----------------------- ABCD of data.
------------------------------------------------------------A
(Single choice) 14. Where is the foothold of symbolism? --------------------------------B
(Judgment) 20. Huawei's AI full scenarios include public cloud, private cloud, various edge computing,
A. The foothold is in neuron networks and deep learning. IoT industry terminals, and consumer terminals and other end, edge, and cloud deployment
environments.
B. The foothold is in reasoning, symbolic reasoning and machine reasoning.
------------------------------------------------------------A
D. The foothold is behavior control, adaptive and evolutionary computing. 2. Overview of machine learning
(Judgment) 2. A model composed of machine learning algorithms cannot represent the true data
(Judgment) 15. The current stage of artificial intelligence is still in the weak artificial intelligence stage. -- distribution function on a theoretical level. ------------------------------------------------ A
------------------------- A (Single choice) 3. Among the machine learning algorithms, which of the following are not integrated
16. Which of the following are the application areas of AI? -------------------------------ABCD learning strategies? ---------------------D
A. Wisdom Education A.Boosting
B. Smart City B. Stacking
C. Smart Home C. Bagging
D. Smart medical D. Marking
(Judgment) 17. Artificial intelligence is the research and development of theories, methods and (Judgment) 4. The loss function of logistic regression is the cross-entropy loss function. -----------------------
application systems for simulating, extending and expanding human intelligence Of a new technological ------- A
science. --------------------------------------------- A A.TRUE
B.FALSE
(Single choice) 18. Which of the following schools does neural network research belong to? ----------------- (Judgment) 5. Regular term can also be added to logistic regression to avoid overfitting. ----------------------
-------------------------------------------B A - regularization
A. Symbolism A.TRUE
B. Connectionism B.FALSE
C. Behaviorism 6. What is the fitting surface of linear regression in more than 3 dimensions? ----------------------------- C
D. None of the above A. Surface
B. Plane
19. What are the main technical applications of artificial intelligence? ---------------------------------------------- C. Hyperplane
------------ACD
D. Hypersurface
(Judgment) 7. K-fold cross-validation refers to dividing the test data set into K sub-data sets. ----------------
B. Control system ---- B
C. Computer Vision A.TRUE
D. Speech recognition B.FALSE
(Judgment) 8. Grid search is a method of parameter adjustment. ---------------------------------- B (Single choice) 13. In Bagging integrated learning, what is the relationship between each base
learner? -------------------- B
A.TRUE
B.FALSE
B. Correlation
(Single choice) 9. Which of the following statements about supervised learning is correct? -------------------
--------- A C. The latter model must be built on top of the previous model
A. Decision tree is a supervised learning D. Independent
B. Supervised learning cannot use cross-validation for training (Single choice) 14. After the data has completed the feature engineering operation, in the process
C. Supervised learning is a rule-based algorithm of constructing the model, which of the following options is not a step in the decision tree
construction process?
-------------------------------------------------- --------- C
(Multiple choice) 10. When dealing with actual problems, when should machine learning be used? ACD
A. Pruning
A. The data distribution itself changes over time and requires continuous re-adaptation of the
program, such as predicting the trend of merchandise sales. B. Feature selection
B. The complexity of the rules is low and the problem is small. C. Data cleaning
C. The rules of the task will change over time, such as the production line. Defect detection
on D. Decision tree generation
D. The rules are very complicated or cannot be described, such as face recognition and 15. What are the commonly used kernel functions in SVM? --------------------------------- ABCD
speech recognition
11. Regularization is an important and effective technique for reducing generalization errors in
traditional machine learning. The following techniques are regular The technology is: ------------- B. Polynomial kernel function
-------------------------------- ---ABC C. Sigmiod kernel function
A. L1 regularization
D. Linear kernel function
B. L2 regularization
(Single choice) 16. In the gradient descent algorithm, which algorithm is the most confusing
C.Dropout trajectory on the loss function surface? -----A
D. Momentum optimizer A. SGD

B.BGD
12. Which of the following is correct about the difference between machine learning algorithms C.MGD
and traditional rule-based methods? -------- ABC
D.MBGD
(Judgment) 17. In the two-classification process, we can set any category as a positive example. -
B. Traditional rule-based methods use explicit programming to solve problems ------------------------ A
C. The mapping relationship of the model in machine learning is automatically learned A.TRUE
D. The mapping relationship of the model in the machine learning institute must be implicit B.FALSE
(Judgment) 18. When we describe housing, we often use the attributes of residential area, house (Single choice) 31. Which of the following descriptions of machine learning is correct? -----------
type, decoration type, etc. If we use plain ---------------------- A
With Bayes as a model, we assume that there is no relationship between attributes. ----------------- A. Deep learning is a branch of machine learning
-A
(Judgment) 19. Loss function and model function are the same thing. ---------------------------------
C. Deep learning and machine learning are both artificial intelligence but have no relationship
---B
with each other
(Single choice) 20. "From historical data with labels, predict whether the retail sales of the next
quarter will be 200,000-300,000 or 300,000-400,000",
(Multiple choice) 32. Which of the following descriptions of the k value in the KNN algorithm is
correct?--------------------------- BC
A. The larger the K value, the easier the model is to overfit
is not linear. ----------- B
B. The larger the K value, the smoother the segmentation surface of the classification
A.TRUE
C. K value is a hyperparameter
B.FALSE
D. The value of k can be set to 0
28. What are the common types of dirty data? ---------------------------------- ABCD
A. Malformed value
is not linear.
B. Duplicate value
-------------------------------------------------- ---------- B
C. logically wrong value
A.TRUE
D. Missing values
B.FALSE
(Single choice) 29. What is the most important difference between batch gradient descent, small
(Judgment) 34. Add the L1 regular term to the loss function of linear regression. At this time, the
batch gradient descent, and stochastic gradient descent? ------ D
regression is called Lasso regression.
A. Gradient size
-------------------------------------------------- ----------- A
(Single choice) 35. What is the label predicted by the regression algorithm? -------------------------
C. Learning rate ------------ ------D
D. Number of samples used A. Self-deformation
(Single choice) 30. Which of the following evaluation indicators belong to the regression B. Discrete
algorithm? ---------------------------------- C
C. Strain type
A. Recall rate
D. Continuous
B. Confusion matrix
(Single choice) 36. Global gradient descent algorithm, stochastic gradient descent algorithm and
C. Mean square error batch gradient descent algorithm are all gradient descent algorithms. The following statements
about their advantages and disadvantages are wrong:
D. Accuracy
-------------------------------------------------- -----------C
A. The global gradient algorithm can find the minimum value of the loss function A.TRUE
B. Batch gradient algorithm can solve the local minimum problem B.FALSE
C. Stochastic gradient algorithm can find the minimum value of the loss function Three, deep learning overview
D. The convergence process of the global gradient algorithm is time-consuming 1. In a convolutional neural network, different layers have different functions. Which of the
following layers can play a dimensionality reduction effect?
37. In the random forest, what strategy does the final ensemble model use to determine the model
result?
-------------------------------------------------- ----------BC -------------------------------------------------- ----------------- BCD à CD
A. Cumulative system A. Input layer
B. Find the average B. Fully connected layer
C. Voting system C. Convolutional layer
D. Multiplying system D. Pooling layer
(Judgment) 38. The test error will keep getting smaller as the complexity of the model increases. 2. Which of the following activation functions are prone to vanishing gradient problems? ---------
---------------- A
-------------------------------------------------- --------- B
A. ReLU
A.TRUE
B.Softplus
B.FALSE
C. Tanh
39. What are the common clustering algorithms?--------------------------------------- ABCD à
ABD D. Sigmoid
A. Density clustering (Single choice) 3. Which of the following statements about the standard RNN model is correct? -
---------------------------- D
B. The time direction is not considered when backpropagating
D.Kmeans
(Single choice) 40. In supervised learning, which of the following models is used to describe "the
one who is close to the red and the one who is red to the black"? D. There will be a problem of attenuation of long-term transmission and memory information
-------------------------------------------------- ------------ C (Single choice) 4. In the process of training the neural network, we use the gradient descent
method to continuously update which value, and then make the loss
A.K-Means
Minimize function? ---------------------------------------------- -----D
B. SVM
C.KNN
B. Characteristic value
D. Neural network
C. Hyperparameters
(Judgment) 41. The naive Bayes algorithm does not require independent and identical
distribution among sample features. ------------------- B D. Parameters
(Single choice) 9. The loss function reflects the error between the target output of the neural
network and the actual output. The loss commonly used in deep learning
(Single choice) 5. 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 The loss function is: ---------------------------------------------- -------B
neural networks. Which of the following statements about feedforward neural networks are
correct:
-------------------------------------------------- ------------D
D. Hinge loss function
B. Its input nodes have computing capabilities
(Single choice) 10. There are many types of neural networks in deep learning. The following
neural network information is one-way propagation: ---------B
D. Information between layers is only transmitted in one direction
A. LSTM
(Single choice) 6. In the neural network, which of the following methods are used to update the B. Convolutional Neural Network
parameters when training the network to minimize C. Recurrent Neural Network
D. GRU
The loss function? --------------------------------------------- ------D
(Single choice) 11. In the deep learning neural network, the perceptron is the simplest neural
network. The correct statement about its structure is:
-------------------------------------------------- --------------------B
A. Its hidden layer has only two layers
B. Its hidden layer has only one layer
(Single choice) 7. The optimizer is an important part of training neural networks. The purpose of
C. The Sigmoid activation function is used in the network
using the optimizer does not include which of the following:
D. Relu activation function is used in its network
-------------------------------------------------- --------------------C
12. Generative confrontation networks are currently widely used. Which of the following
scenarios can use this network?
B. Reduce the difficulty of setting manual parameters
-------------------------------------------------- -----------------ABCD
C. Avoid overfitting à normalizer
A. Data enhancement
(Single choice) 8. Can the perceptron be displayed in space? ----------------------------- C à
A. Line
D. Image generation
B. Plane
(Single choice) 13. Recurrent neural network is different from convolutional neural network.
C. Hyperplane Which of the following problems is it better at solving?
D. Point -------------------------------------------------- ---------A

A. Sequence Related Issues
B. Image classification D. Self-decreasing loss function
C. Image detection
(Single choice) 18. The following applications that are not part of the adversarial generation
D. Recommendation question
network? -------------------------------- C
14. Weight sharing is common in neural networks. Which of the following neural networks will
A. Text generation
share weights?
B. Image generation
-------------------------------------------------- ---------BD
A. Perceptron
D. Data enhancement
(Judgment) 19. All convolution kernels of the same convolution layer in the convolutional neural
network are weight-sharing. --------------- A
TRUE
FALSE
15. In deep learning tasks, when encountering data imbalance problems, which of the following
20. Knowing that the total number of parameters of a layer of the fully connected neural network
methods can we use to solve the problem?
is 330, then the number of neurons in the previous layer and this layer
May be? ----------------------------------------------- -----BC
-------------------------------------------------- --------- BCD
A.32 and 10
A. Batch delete
B.10 and 33
C.33 and 10
D. 9 and 33
(Single choice) 21. God will often encounter many problems when training the network. For the
(Single choice) 16. Regarding backpropagation, is the following statement wrong? ----------------- problem of gradient disappearance, we can choose to use
---------------- A
Which of the following functions should be used to alleviate this problem? --------------------------
A. Backpropagation can only be used in feedforward neural networks -------------A
B. Backpropagation can be combined with gradient descent algorithm to update network weights A. Relu function
C. Backpropagation will go through the activation function B. Sigmoid function
D. Back propagation refers to the back propagation of errors through the network C.tanh function
17. What are the commonly used loss functions in deep learning? ------------------------------------ D. Softsign function
BC
(Single choice) 22. Thnaive bayese number of hidden layers of the deep learning neural network
A. L1 loss function has a certain impact on the performance of the network. Which of the following statements about
its impact is correct: ------------------ ------------------------------------B
A. The number of hidden layers is appropriately reduced, and the resolution ability of the neural
network remains unchanged
B. Increase the number of hidden layers appropriately, the stronger the resolution ability B. Recurrent neural network
of the neural network
C. Convolutional neural network
C. The number of hidden layers is appropriately reduced, the stronger the resolution ability of the
D .Fully connected neural network
neural network is
(Single choice) 27. The activation function plays an important role in learning and understanding
D. The number of hidden layers increases appropriately, the weaker the resolution ability of the
very complex problems of neural network models. The following statement about the activation
neural network
function is correct.
-------------------------------------------------- ----------B
23. In the process of deep learning model training, what are the common optimizers?
-------------------------------------------------- -------ABCD
A. Adam
C. The activation function is partly a non-linear function and part is a linear function
B. Adagrad
C. SGD
D. Momentum
(Single choice) 29. 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:
(Single choice) 24. In the classic convolutional neural network model, what hidden layer does
the Softmax function follow? -------------------------------------------------- -----------C
-------------------------------------------------- ----------C A. Motherboard
A. Convolutional layer B. RAM
B. Pooling layer C. GPU
C. Fully connected layer D. CPU
D. All of the above 30. Which of the following steps in deep learning are automatically completed by the model?
(Single choice) 25. What is wrong about the following statement about cyclic neural network? --- -------------------------------------------------- ------BD
---------------------------B
A. Model training
A. Cycle The neural network can be expanded according to the time axis
B. LSTM cannot solve the problem of gradient disappearance
C. LSTM is also a recurrent neural network
31. Is the following description correct about the convolutional neural network pooling layer?
(Single choice) 26. For the image classification problem, which of the following neural networks
-------------------------------------------------- ------ABCD
is more suitable to solve this problem? ------------C
A. The pooling operation uses a scanning window to achieve
A. Perceptron
C. Commonly used pooling methods include maximum pooling and average pooling C. Maximum pooling layer
D. After the characteristics of pooling Image becomes smaller D. Average pooling layer
(Single choice) 32. In the deep learning network, the backpropagation algorithm is used to find (Single choice) 36. In deep learning, if the neural network has a large number of layers, the
the optimal parameters. What rules are used in the backpropagation algorithm to obtain the problem of gradient disappearance is more likely to occur. Strictly speaking, in which of the
derivation layer by layer? following links does the sample disappearance problem occur?
-------------------------------------------------- ---------A -------------------------------------------------- ------------A
A. Chain rule A. Back propagation update parameters
B. The rule of accumulation B. Forward propagation update dreams
C. Law of Equivalence C. Back propagation calculation results
D. Rule of Normalization D. Forward propagation calculation results
(Single choice) 33. L1 and L2 regularization are methods commonly used in traditional machine (Judgment) 37. The cyclic neural network can capture the dynamic information in the serialized
learning to reduce generalization errors. The following is about the two data.
The correct method is: --------------------------------------------- --------A -------------------------------------------------- -----------A
A. L1 regularization can be used for feature selection (Single choice) 38. Where should the labeled data be placed in the adversarial generative
network?
B. Both L1 and L2 regularization can be used for feature selection
-------------------------------------------------- ----------B
C. L2 regularization can do feature selection
A. As the output value of the generative model B. As the input value of the discriminant model
D. Neither L1 nor L2 regularization can be used for feature selection
C. As the output value of the discriminant model D. As the input value of the generative model
(Single choice) 34. Traditional machine learning and deep learning are the core technologies of
39. If the deep learning neural network has the problem of gradient disappearance or gradient
artificial intelligence, and there is a slight difference in the engineering process. The following
explosion, our common solution is.
steps
-------------------------------------------------- --------ACD
What you don't need to do in deep learning is -----------------------------------B
A. Gradient shear
A. Model evaluation
C. Use Relu activation function
C. Data cleaning
D. Regularization
D. Model construction
35. The pooling layer in the convolutional neural network can reduce the size of the lower layer
input. Common pooling are: (Single choice) 40. Input a 32*32 image, use a 5*5 convolution kernel to do the convolution
calculation with a step size of 1, and output the image
-------------------------------------------------- ----------CD
The size is ------------------------------------------------ ------ B
A. Minimal stratum
A.28*23
B.28*28
C.29*29 B.concat
D.23*23 C.split
(Single choice) 41. There are often many problems in the training of the gods network. For the D.unstack
problem of gradient disappearance, we can choose which of the following functions to alleviate
the problem?
(Judgment) 2. The three main advantages of the Keras interface in TensorFlow 2.0 are user-
-------------------------------------------------- -----------B
friendly, modular, and
42. Which of the following are the activation functions of the deep learning algorithm?
-------------------------------------------------- --------ABC
Combination, easy to expand. -------------------------------------------------- A
A. Sigmoid
A.TRUE
B. ReLU
B.FALSE
C. Tanh
(Judgment) 3.tf.keras.datasets You can view the built-in data sets in keras. ------------------ A
D. Sin
A.TRUE
B.FALSE
(Single choice) 43. In the process of training the neural network, our goal is to keep the loss
4. As of the end of 2019, what version of TensorFlow has released? ---------------------- ACD
function reduced. Which of the following methods do we usually use to minimize the loss
function? A.1
-------------------------------------------------- ----------A B.3
A. Gradient descent C.2
B. Dropout
C. Cross validation D.0.1
D. Regularization (Judgment) 5. TensorFlow is an end-to-end open source platform for machine learning and deep
(Single choice) 44. What is wrong with what people say about neural networks?-------------------- learning. ---------- A
------------- -A A.TRUE
A. With the increase in the number of hidden layers of the neural network, the classification B.FALSE
ability of the model gradually weakens
(Judgment) 6. TensorFlow is one of the most popular deep learning frameworks. -------------------
B. The limitation of the single-layer perceptron is that it cannot solve the XOR problem ---- A
C. The feedforward neural network can be represented by a directed acyclic graph A.TRUE
D. Front There is no connection between neurons in the same layer of the feed neural network B.FALSE
Fourth, the industry's mainstream development framework 7. The following code was used when compiling the model:
(Single choice) 1. What methods can be used for tensor merging in TensorFlow 2.0? -------------- model.compile(optimizer='Adam,loss='categorical.crossentropy',
-----------B metrics=[tf.keras.metrics.accuracy]),
A.join
When using the evaluate method to evaluate the model, which of the following indicators will be (Single choice) 12. What data type does TensorFlow use? ------------------------------------C
output?
A. Scalar
-------------------------------------------------- -----------AC
B.Vector
A.accuracy
C.Tensor
B.categorical_ 1oss
D.Matrix
C.loss
D.categorical accuracy
(Judgment) 13.TensorFlow 2.0 supports GPU acceleration mainly in the parallel operation of
(Single choice) 8. What methods are not included in the built-in indicators of keras.metrics of matrix addition, subtraction, multiplication and division
TensorFlow 2.0? ----------C
accelerate. -------------------------------------------------- ------- A
A. Accuracy
A.TRUE
B. Recall
B.FALSE
C. Sum
14. What are the following features of TensorFlow 2.0? ---------------------------- ACD
D. Mean
(Judgment) 9. The graph and session mechanism is cancelled in TensorFlow 2.0. -------------------
--------B
D. Continue to be compatible with related modules of Tensorflowl.x
(Single choice) 10. When building a neural network using the keras interface of TensorFlow 2.0,
the network needs to be compiled (Judgment) 15. PyTorch's current version supports Python2 and Python3? ---------------------- B
Which of the following methods should be used to work? ---------------------------------------A A.TRUE
A.compile B.FALSE
B.write (Judgment) 16. TensorFlow2.0 does not support tensor transfer between GPU and CPU. ----------
-- B
C.join
A.TRUE
D.fit
B.FALSE
(Single choice) 11. The operators that are not supported by TensorFlow 2.0 among the following
options are ----------------------C (Single choice) 17. What are the following features that are not part of TensorFlow 2.0? ----------
------------------- A
A.pow
A. Multi-core CPU acceleration
B.@
B. Distributed
C.^
C. Multilingual
D.//
D. Multi-platform
(Single choice) 18. What is the way to view the tensor dimension in TensorFlow 2.0? ------------- D. GPU acceleration
------------C
23. What is the built-in optimizer of the keras module in TensorFlow?----------------------ABCD
A.dimens
A. Adam
B.dtype
B. SGD
C.ndim
C. Adaboost
D.device
D. Adadelta
(Single choice) 19.Code model.fit(mnist.train.inmage,mnist.train.labels.epochs=5)
What does the epochs parameter represent? --------------------------------------------A
24. Is there a way to check whether it is a tensor in TensorF1ow2.0?
A. The entire training set will be trained 5 times
-------------------------------------------------- ---------BC
B. The entire test set will be tested 5 times
A.dtype
C. The entire training set will be divided into 6 parts
B.isinstance
D. The entire training set will be divided into 5 parts
C.is_tensor
D.device
20. Which of the following options belong to the estimator package method under keras? ---------
--------ABC
(Single choice) 25. Which of the following companies developed TensorF1ow first?
A. Evaluation
-------------------------------------------------- ----------D
B. Training
A. Oracle
C. Forecast
B. Facebook
D. Output model
C. Nvidia
(Single choice) 21. Among the following attributes, TensorFlow 2.0 does not support the method
to create tensor? ---------------C D. Google
A.zeros
B.fill (Single choice) 26.Which company did Pytorch first launch?
C.create -------------------------------------------------- ----------C
D.constant A. Baidu
(Single choice) 22. Which of the following functions does PyTorch have? --------------------------- B. Google
-------A
C. Facebook
A. Embedded keras
D. Huawei
-------------------------------------------------- -------------C
(Single choice) 27.What is the role of keras.preprocessing of TensorFlow 2.0? A.MindSpore uses gradient data-driven adaptive graph optimization to achieve decentralized
autonomous A11 Reduce, Gradient aggregation is in the same step, and calculation and
-------------------------------------------------- ---------A
communication are fully streamlined.
A.keras data processing tool
B. Challenges of model execution under super chip computing power:
B.keras built-in optimizer
Memory wall problems, high interaction overhead, and difficulty in data supply. Partly executed
C.keras model deployment tool on the Host and partly executed on the Device, the interaction overhead is even much greater
than the execution overhead, resulting in low accelerator occupancy
D.Keras generating model tool
C. MindSpore uses chip-oriented depth map optimization technology to minimize
(Judgment) 28. The tf.contrib method can continue to be used in TensorFlow 2.0. synchronization waiting and maximize "data computing The parallelism of “trust”, the training
-------------------------------------------------- ---------B performance is the same as that of the Host side graph scheduling method.
D. The challenge of distributed gradient aggregation under super-powered chip computing
power: ReslNet50 will be generated in a single iteration of 20ms.
(Single choice) 29. The following options do not support TensorFLow2.0 for dimension
transformation. The synchronization overhead of heart control and the communication overhead of frequent
synchronization. The traditional method requires 3 simultaneous completion of A11 Reduce, and
-------------------------------------------------- ---------C the data-driven method is autonomous A11 Reduce, without control overhead
A.squeeze
B.reshape (Single choice) 4. Which of the following is not a feature of MindSpore's core architecture?
C.gather -------------------------------------------------- -------------C
D.transpose A. Automatic differentiation
B. Automatic tuning
5. Huawei's AI development framework MindSpore C. Automatic coding
(Judgment) 1. Tensor is the data storage component in MindSpore. ----------------------------A D. Automatic parallel
(Single choice) 2. Which of the following is not a commonly used Operation in MindSpore? -----
---------------------A
5. MSIR is a concise, efficient and flexible graph-based functional IR, a functional semantic type
A.signal that can be expressed
B.math Have?------------------------------------------------ --------ABD
D.nn A. Free variables
C.array B. Higher-order functions
(Single choice) 3. On-Device execution, that is, the entire image is offloaded and executed, and C. Interrupt
the computing power of the Yiteng chip can be fully utilized, which can greatly reduce the
interaction overhead, thereby increasing the accelerator occupancy rate. What is wrong about the D. Recursion
implementation of the following description of on-Device?
(Single choice) 6. Faced with the challenge of implementing efficient distributed training for D. Weight initialization
ultra-large-scale models, how does MindSpore handle it?
-------------------------------------------------- -------------A
Six, Atlas artificial intelligence computing platform
1. What conditions need to be checked when compiling and running programs using a server
B. Serial
equipped with Atlas300 (3000) accelerator card?
C. Manual parallel
-------------------------------------------------- -----------------ACD
(Single choice) 7. Ce11 provides basic modules for defining and executing calculations. The
objects of Ce11 can be executed directly. What is wrong with the following statement? ----------- A. Complete Atlas driver installation
--------- ---------------------------------ACD
B. CUDA software package installed
A.__init__, initialization parameter (Parameter), submodule (Ce11), Operators (Primitive)
and other groups C. Cmake compilation tool is installed
B.Construct, which performs initialization verification, defines the execution process. In the D. Complete environment configuration
graph mode, it will be compiled into a graph for execution, and there is no grammatical
restriction.
2. From the perspective of technical architecture, the classification of AI chips includes? ----------
C. There are also some predefined optim commonly used optimizers, wrap commonly used -------------------ABCD
network packaging functions, etc. Cel1
A. FPGA
D.bprop (optional), the reverse of custom module
B. CPU
C. GPU
(Single choice) 8. Facing the huge gap between industry research and full-scenario AI
applications, which of the following is not included in the technological innovation of D. ASIC
MindSpore to help inclusive AI across the application gap? (Judgment) 3.Root users cannot install DDK (Device Device) on the server with Atlas300 (3000)
-------------------------------------------------- ------------B accelerator card.
A. New collaboration method Development Kit), can only be installed by ordinary users. ------------------------A
B. New programming language (Single choice) 4. What is the main application of the Yiteng 310 chip based on Atlas 200 DK? -
-------------------A
A. Model reasoning
B. Build the model
(Single choice) 9. During application development, which of the following operations is not a
typical network definition? C. Training model
-------------------------------------------------- -------------C
A. Network definition 5. What are the main computing resources contained in the Da Vinci architecture computing
unit? ------------------------ABD
B. Scalar calculation unit
C. Tensor calculation unit C. Model building
D. Matrix calculation unit D. Reasoning
(Single choice) 11.Which command can check whether the Atlas 300 (3000) accelerator card is
in place? ---------------- B
(Judgment) 6. AI chips are also called AI accelerators, which are specially designed to handle a
large number of computing tasks in artificial intelligence applications A.1spci | grep'npu'
functional module. -------------------------------------------------- --- A B.1spci | grep'd100'
A.TRUE C.1spci grep'atlas'
B.FALSE D.atlas info
(Single choice) 7. Which of the following options is not included in the logical architecture of (Judgment) 12. GPU is good at computationally intensive and easy to parallel programs. ----------
the flatten AI processor? ---------------------- B -------------------- A
A. DVPP A.TRUE
B. GPU B.FALSE
C. AI calculation engine (Multiple choice) 13. Atlas 800 AI server has multiple models, which is based on the Kunpeng
processor platform? ------- B
A. Atlas 800 Model: 9000
(Single choice) 8. Check if the Atlas300 (3000) accelerator card driver is installed successfully,
B. Atlas 800 Model: 3000
which command should be used? ------- A
C. Atlas 800 Model: 3010
A.npu-sim info
(Single choice) 14. Which of the following options is the same as Shengteng 310 and Shengteng
B.npu info 910? -----------------------B
C.atlas-driver info A. Maximum power
D.atlas info B. Architecture
(Judgment) 9. Install DDK (Device Development Kit) in the server with Atlas300 (3000) C. Integer precision
accelerator card
D. 7nm process
Only Python3 needs to be installed when the environment is dependent. ------------------------------
(Judgment) 15. Huawei Shengteng AI chip is one of the typical representatives of NPU (Neural
------ A
Network Processor). ------------A
A.TRUE
(Single choice) 16. What processor does Atlas use to accelerate AI inference?
B.FALSE
-------------------------------------------------- -----------B
10. AI chips are divided into business applications and can be divided into? -------------------------
A. Isotron 910 processor
---- AD
B. Yiteng 310 processor
A. Training
C. GPU
B. GPU
D. FPGA
(Single choice) 5. Which tool does the HUAWEI HiAI platform support integration with? --------
---------------------C
17. Which of the following libraries are installed in the server with Atlas300 (3000) accelerator
card DDK (Device Development Kit) which needs to be installed before? A. Jupyter Notebook
-------------------------------------------------- -------ABC B. MyEclipse
A.decorator C. Android Studio
B.setuptools D. Spider
C.numpy (Single choice) 6. Which HUAWEI HiAI module does Huawei chip support? -----------------------
--B
D.tensorflow
A. HiAI Engine
B.HiAI Foundation
(Judgment) 18. The CPU mainly improves AI performance by increasing instructions and
increasing the number of cores. C.HiAI Framework
-------------------------------------------------- ------------A D.HiAI Service
7. Huawei Smart Terminal AI Open Platform
(Single choice) 1. Which module does the face detection in the HUAWEI HiAI platform belong 7. What is the value of HUAWEI HiAI empowering APP? -------------------------------ABCD
to? -----------------------A
A. Security
A. HiAI Engine
B. Stable
B.HiAI Framework
C. Real time
C.HiAI Foundation
D. Anytime
D.HiAI Service
(Judgment) 2. HUAWEI HiAI Engine can easily integrate multiple AI capabilities with App. ----
8. What are the advantages of the HiAI mobile computing platform?---------------------------------
---------A
ABCD
(Judgment) 3. The HiAI mobile computing platform provides a wealth of upper-level functional
A. Quick start source code
business APIs, which can run efficiently on mobile devices.
B. Diverse tool chains
-------------------------------------------------- --------------------A
4. What models does HUAWEI HiAI support? --------------------------------ABC
D. Rich API
A. P30
B. Mate 20
(Single choice) 9. Which of the following is a function of the HUAWEI HiAI Foundation
C. Honor V20
module?
D. iPhone10
-------------------------------------------------- ------------C
A. App integration
B. Let the service actively find users C.content
C. Quickly convert and migrate existing models D.words
D. Push the service at the right time and according to the needs of users
5. According to different industries and application scenarios, there are many kinds of HUAWEI
CLOUD EI agents. What are the currently available HUAWEI CLOUD EI agents?
-------------------------------------------------- ----------ABCD
8. Huawei Cloud Enterprise Intelligent Application Platform
A. Industrial intelligence
1. Which statement about the general form recognition service is correct? ------------------------
ABCD B. Traffic intelligence
A.rows represents the row information occupied by the text block, the number starts from C. Park intelligence
0, in the form of a list
D. Automobile intelligence
B.colums represents the column information occupied by the text block, the number starts
6. Which of the following items are included in the results returned when the face search service
from 0, in the form of a list
is successfully invoked?
C. The incoming image data needs to be base64 encoded
-------------------------------------------------- ---------ABC
D.words represents the result of text block recognition
A. The similarity of the searched face
B. The searched face id
(Judgment) 2. There can be multiple tags returned by the image tag service. -------------------------
C. The searched face position
------A
D. The searched face number
(Single choice) 3. When the voice recognition service is successfully called, which field is the
recognition result stored in? (Single choice) 7. The combination of ModelArts service and () service can easily deploy the
model to the "end"?
-------------------------------------------------- ---------A
-------------------------------------------------- ------------D
A.result
A.OBS
B.content
B.OCR
C.data
C.ECS
D.text
D.HiLens
(Single choice) 4. When the universal text recognition service is successfully invoked, the
returned text recognition result is stored in which of the following fields? (Single choice) 8. Which of the following is not an application of image recognition service? ----
-------------------------D
-------------------------------------------------- ---------B
A. Target detection
A.text
B.result
C. Scene analysis
D. Speech synthesis A. Dialogue robot
B.EI basic services
9. HUAWEI CLOUD ModelArts is a one-stop development platform for AI development. C. Natural language processing
Which of the following functions are in
Is it available on Mode1Arts? --------------------------------------- ABCD
(Multiple choice) 14. Which of the following capabilities does the gene knowledge map have? --
A. Data governance ---------------------------- ACD
B. AI market A. Auxiliary case diagnosis
C. Visual workflow B. Disease prediction and diagnosis
D. Automatic learning C. Gene test report generation
(Multiple choice) 10. Which of the following statements about the type field returned by the D. Entity query
universal form recognition service is correct?
15. Is the following statement about passport identification service correct? -------------------------
-------------------------------------------------- -------- BD -ABCD
A.type represents the type of text recognition area A.country_code represents the country code of the country where the passport was issued
B.nationality represents the nationality of the holder
B. When type is text, it represents the text recognition area
C.passport_number represents the passport number
C.type represents the table type
D.confidence The confidence information of related fields. The greater the confidence, the
D.type is table, it represents the table recognition area
higher the reliability of the corresponding field identified this time. In a statistical sense, the
(Judgment) 11. Only one image can be passed in when calling the face comparison service. ------ greater the confidence, the higher the accuracy.
------------------- B
(Judgment) 16. In the face search service, if we want to delete a face set, we can use this code:
A.TRUE
firs_cliet.get_v2().get_face_set_service().delete_face_set("**"),
B.FALSE
Where "**" is to fill in the actual face set name. ------- ----------------------A
(Multiple choice) 12. HUAWEI CLOUD EI allows more enterprise boundaries to use AI and big
A. Target detection
data services to accelerate business development and benefit society.
In which of the following aspects can Huawei Cloud EI service serve enterprises? ------------------
------ ABCD C. Scene analysis
A. Industry data D. Speech synthesis
B. Industry wisdom 17. What happens to the label sugestion when the image remake detection service is called
successfully? -----------BCD
C. Algorithm
A.certainty
D. Computing power
B.FALSE
(Multiple choice) 13. Which of the following services belong to the Huawei EI service family? --
-------------------------- ABCD C.uncertainty
D.TRUE B. Policy query plan based on knowledge graph
C. Vehicle identification plan
18. Which of the following are the topics of speech processing research? ----------------------------- D. Intrusion identification plan
---BCD
(Single choice) 24. In the environment preparation for calling HUAWEI CLOUD services,
A. Voice processing which of the following does not require confirmation of installation?
B. Voiceprint recognition -------------------------------------------------- ----------D
C. Speech recognition A.requests
D. Voice wake up B.setuptools
C.websocket-client
(Judgment) 19. The face detection service will report an error when there is no face in the test D.Jieba
image. ------------------------B
(Judgment) 25. The word recognition in the speech recognition service refers to the synchronous
(Judgment) 20. The image data of the passport recognition service does not need to be base64 recognition of short speech. Upload the entire audio at once, and the recognition result will be
encoded. -------------------B returned in the response.
(Single choice) 21. What is wrong about the image content review service? -------------------------- -------------------------------------------------- -----------A
---C
26. Which of the following options are need to be setting in Huawei Cloud
A.politics is the detection result of politically sensitive persons
-------------------------------------------------- -------ABCD
B.terrorism is the test result of political violence and terrorism
A.ak
C.confidence represents the degree of confidence, ranging from 0-100
B.region
D.labe1 represents the label of each test result
C.project_id
22. Regarding the suggestion field returned when the image content review service is called
D.sk
successfully, which statement is correct?
27. Regarding the face search service, which of the following statements are correct?
-------------------------------------------------- ---------ABCD
-------------------------------------------------- -------ACD
A.pass means no sensitive information, pass
A. When there is no face set, you need to create a face set first, and then add face data before you
B.review represents the need for manual review
can search B. The size of a face set cannot exceed 10,000 pictures C. There is a special interface
C.suggestion field represents whether the test passed to delete a specified face set D. There is a dedicated interface to delete the face data in a certain
face set
D.block means that it contains sensitive information and does not pass
(Single choice) 28. In HUAWEI CLOUD EI, which of the following can integrate AI technology
23. According to Huawei Cloud EI Intelligent Platform, which of the following solutions can be
into the application scenarios of various industries and take advantage of the advantages of AI
provided?
technology, thereby improving efficiency and enhancing experience.
-------------------------------------------------- -------ABCD
-------------------------------------------------- ----------A
A. People flow statistics plan
A. EI agent
B. OBS Object Storage Service A./v1.0/image/recognition
C. Cloud database B./v1.0/image/celebrity-recognition
D. EI big data service C./v1.0/image/recapture-detect
D./v1.0/image/tagging
(Single choice) 29. Can the traffic agent realize () hour all-time/global traffic perception? (Single choice) 1. Feature is the dimension that describes the characteristics of the sample.
Regarding its interpretability in traditional machine learning and deep learning, which of the
-------------------------------------------------- ----------A
following statements is correct: --------------- -------------------------------A
A.7×24
B.5×24
B. Features are weak in traditional machine learning, but strong in deep learning.
C.5×12
C. Features are weak in traditional machine learning and deep learning.
D.7×12
D. Features are strong in traditional machine learning and deep learning.
What is the problem? -------------------------------------------- C
(Single choice) 30. An e-commerce company has always had problems with its official website
stalling, poor utilization of network resources, and difficulty in operation and maintenance. Then
which of the following EI services can the company use? B. Rule issues
-------------------------------------------------- -----------C C. Classification issues
A. Park agent D. Clustering problem
B. Traffic Agent (Single choice) 21. Compared with the random forest algorithm, which of the following
statements is wrong with the GBDT algorithm?
C. Network Agent
-------------------------------------------------- ---------- C
D. Industrial Agent
A. GBDT algorithm is easier to underfit than random forest
B. Random forest is calculated in parallel, while GBDT cannot
(Single choice) 31. Which of the following is not a face recognition technology?--------------------
-------------- D C. GBDT algorithm is easier to overfit than random forest
A. Face search D. GBDT and random forest are both based on the CART tree
B. Face comparison (Single choice) 22. Which of the following machine learning algorithms is not unsupervised
learning? ----------------------- B
C. Face detection
A. GMM
B.Xgboost
C. Clustering
(Single choice) 32. Which of the following is the URI of the image label service?
-------------------------------------------------- ---------D
(Judgment) 23. The training error will continue to decrease as the model complexity increases.
-------------------------------------------------- --------- A D. The generator produces different samples
A.TRUE
B.FALSE
(Single choice) 24. If a model has a large deviation on the test set and a small variance, then it
means the model?
-------------------------------------------------- --------- D
A. Overfitting
B. It may be overfitting or underfitting
C. Just fit
D. Underfitting
25. If there is a large amount of sales data but no labels, the company wants to identify VIP
customers. The following model is combined
Suitable? ---------------------------------------------- ------ CD
B. SVM
C.K-Means
(Single choice) 26. In machine learning, what does the model need to input to train itself and
predict unknown? ---------------- D
A. Manual procedures
B. Neural Network
D. Historical data
(Single choice) 28. Generating a confrontation network is like a game system. The generator
generates fake samples, and the discriminator judges whether it is true or false. Our ideal result
is?-------------- ------------------------C
B. The discriminator efficiently distinguishes the true and false of the samples generated by the
generator
Write a custom catalog title here Q4. Which of the following is not a ModelArts development type?
Artificial Intelligence HCIAv2.0 Question Bank A. Zero-based modeling
B. Agile development
Update date: 2020-6-23 Update content: Correct some wrong questions, the question bank score range C. Quick modeling
is 800-900
D. Standard model development
The question bank is expected to expire on October 1, 2020, please note
Answer: B
A. Symbolism
Q5. Which of the following are ModelArts development models?
B. Connectionism
A. Custom development
C. Behaviorism
B. Iterative learning
C. Data annotation
Answer: B
Answer: ACD
Q2. Smart speakers are one of the typical application scenarios for voice processing.
A.True
Q6. Which of the following engines cannot be provided by HUAWEI HiAI Engine?
B.False
A. NLU engine
B. CV engine
Answer: A
C. ASR engine
D. DSP engine
Q3. What are the main technical applications of artificial intelligence?
A. Control system
Answer: D
C. Computer Vision
Q7. When HiAI Foundation offline model calculation, the offline model is loaded from the file, and the
D. Natural language processing user’s input
Data (such as pictures) is copied and calculated on HiAI's NPU. When calculating, each inference only
Answer: BCD Need to import and export user data from DDR to NPU once.
A.True
B.False
Q11. Which of the following can generate a dictionary?
Answer: A A.{{"C",3}:"c"}
Q8. HiAI Foundation does not support sparse model acceleration. B.{{"D":4}:"d"}
A.True C.{[1, 2,3]:"a"}
B.False D.{"b":["B’"]}
Answer: B Answer: D
Q9. Which of the following features does Python have? Q12. Which of the following output results are wrong?
A. Object-oriented A.sorted([(‘b’ ,2),(‘a’ ,1),(‘c’,3),(‘d’,4)], key=lambda x:x[1])
B. Open source Output result: [( a’,1), (3),(‘d’,4)]
C. Simple B.print (*filter (lambda x: x%2==1, [1, 2, 3]))
D. Explanatory Output result: 1 3
C.print (map(lambda x:xx, [1,2,3]))
Answer:ABCD Output result: 1 4 9
D.print (*zip([1,2,3],["b"], ["A", "B","C"]))‘C’)
Q10. Output result: (1,’a’,""A";,"(2,b,‘B’)
a=1
b=2 Answer: ACD
c = a if a>b else 0
What is the value of c? Q13. What is the result of 5/2 in Python3?
A.2 A.3
B.0 B.2
C.None C.2.5
D.1 D.1
Answer: B Answer: C
Answer: A
The data in Q14.json format is a dictionary in Python.
A.True Q18. What are the following steps for Python3 operation MySQL database?
B.False A. Establish a database connection
B. Execute SQL statement
Answer: B C. Install PyMysq1
D. Close the database connection
A. Website development Answer: CABD
B. Artificial Intelligence
C.APP Q19. Which of the following options describes the relationship between the Poisson distribution and the
binomial distribution correctly?
D. Data Science
A. Poisson distribution can replace binomial distribution
B. Poisson distribution is an approximate calculation of binomial distribution when n is large and p is

Answer:ABCD
small
C. The mathematical models of Poisson distribution and binomial distribution are both Lagrangian
Q16. Which of the following identifiers are named incorrectly? probabilities
A.print D. Poisson distribution has no relationship with binomial distribution
B.If
C._0a Answer: B
D.a0* Q20. Which of the following options describes the numerical characteristics of random variables
correctly?
A. The greater the expectation, the smaller the variance.

Answer: AD
B. Variance reflects the degree of deviation between random variables and their mathematical
expectations.
Q17. Concurrency refers to the simultaneous execution of multiple tasks, and concurrent operations can C. Expectation reflects the average level of random variables.
be completed by using multiple processes in Python.
D. The correlation coefficient is used to measure the linear relationship between two variables.
A.True
B.False
Answer: BCD
B.False
Q21. The first-order tensor is a scalar.
A.True Answer: B
B.False
Q25. After multiplying a vector by an orthogonal matrix. The matrix only produces expansion and
contraction effects on the vector, without rotation and
Answer: B
Space mapping function.
A.True
Q22. Which of the following descriptions of the distribution function, distribution law, and density
function of random variables are incorrect? B.False
A. Discrete random variables have no distribution function
B. The density function can only describe the value law of continuous random variables Answer: B
C. The distribution function describes the value law of random variables
Q26. The direction of the gradient is the direction of the maximum directional derivative.
D. Distribution law can only describe the value law of discrete random variables A.True
B.False
Answer: A
Answer: A
Q23. Regarding matrix decomposition, which of the following descriptions is correct?
A. Only square matrices can perform singular value decomposition. Q27. Which of the following statements about tensors are correct?
B. Carry out singular value decomposition on matrix A, the left singular matrix and right singular matrix A. Tensors of order 4 and above, like vectors of 4 dimensions and above, have no direct geometric
are both orthogonal matrices. meaning
C. After a matrix is multiplied by its eigenvector, the vector only undergoes scaling transformation. B. Array is a tensor
D. The larger the singular value, the greater the amount of information contained in the corresponding C. The operation and model optimization process in TensorF1ow are all based on tensors
singular vector.
D. A matrix is a second-order tensor
Answer: BD
Answer: DABC à BCD
Q24. Perform singular value decomposition on a matrix. The singular value matrix must be a square
matrix and the singular values are larger
Q28. What services does TensorFlow 2.0 support?
Arrange in order of smallest.
A.Lite
A.True
B. TensorFlow2.0 core library C. Binary cross entropy loss
C. JavaScript D. Error squared loss
D.Extended
Answer: ACD
Answer:ABCD
Q33.TensorFlow is an end-to-end open source platform for machine learning and deep learning.
Q29. What kind of network layer is the pooling layer generally connected to? A.True
A. Input layer B.False
B. Output layer
C. Convolutional layer Answer: A
Q34. In Keras, model.summary() is a method used to print the network overview.
Answer: C A.True
B.False
Q30.TensorF1ow is one of the most popular deep learning frameworks.
A.True Answer: A
B.False
Q35. What is the role of tf.keras.preprocessing?
Answer: A
A.keras model deployment tool
Q31. In TensorFlow 2.0, gather is used to increase the dimension. B.keras data processing tool
A.True C. Keras generative model tool
B.False D.keras built-in optimizer
Answer: B Answer: B
What are the built-in loss functions under Q32.tf.keras.losses? Q36. The fully connected layer is generally located in the middle of the entire network to reduce the
amount of parameters.
A. Absolute percentage error
A.True
B. Similarity
B.False
Answer: B
Answer: B
Q41. The higher the validity of the point estimate, the higher the degree of dispersion of the estimator.
Q37.tf.keras.datasets can view the built-in data sets in keras. A.True
A.True B.False
B.False
Answer: B
Answer: B
Q42. What are the common clustering algorithms?
Q38. Unbiased estimation in point estimation means that the mathematical expectation of the sampling A.K-means
distribution of the estimator is equal to the true value of the overall parameter.
B. Spectral clustering
A.True
C. Density clustering
B.False
Answer: A
Answer: A
Q43. Which of the following activation functions are prone to vanishing gradient problems?
A.ID3
A. ReLU
B. Gini coefficient
B. Tanh
C. Information Entropy
C.Softplus
D.CART
D. Sigmoid
Answer: ABCD
Answer: A
Q40. The naive Bayes algorithm does not require independent and identical distribution among sample
Q44. Which of the following descriptions about the convolutional neural network pooling layer is
features.
correct?
A.True
A. The pooled feature image becomes smaller
B.False
B. Pooling operation is realized by scanning window
C. The pooling layer can play a role in dimensionality reduction
D. Commonly used pooling methods include maximum pooling and average pooling Q48. Which of the following structures belong to LSTM?
A. Input gate
Answer: ABCD B. Forgotten Door
C. Memory gate
Q45. What are the strategies to avoid overfitting in model training? D. Output gate
A. Stop training early
B.Dropout Answer: ABD
C. Delay to stop training
D. L2, L1 regular Q49. In the process of deep learning model training, what are the common optimizers?
A. Momenturm
Answer: ABCD à ABD B. Adagrad
C. SGD
Q46. Can the perceptron be displayed in space? D. Adam
A. Point
B. Line Answer: ABCD
C. Plane
D. Hyperplane Q50. Which of the following statements about L1 regularization and L2 regularization are correct?
A. L2 regularization produces a sparser model than L1 regularization
Answer: A B. L1 regular term is conducive to enhancing the generalization ability of the model
C. After adding the L2 regular term, the gradient descent algorithm cannot be used to iterate the
parameter values
Q47. What is the mini-batch in mini-batch gradient descent?
D. L1, L2 regular terms cannot be applied to the loss function.
A. Part of the error is extracted from the error for each parameter iteration
B. Extract a part of the hyperparameters for training at each parameter iteration

Answer: B
C. Extract a part of the feature for training at each parameter iteration

Q51. The Softmax function is often used as the output layer of multi-classification tasks.
D. At each parameter iteration, a part of the data set is extracted for training
A.True
B.False
Answer: D
A.True
Answer: A B.False
Q52. The training method used by the recurrent neural network is called back-propagation, or BPTT for Answer: A
short.
A.True
Q56. Which of the following applications are not part of the adversarial generation network?
B.False
A. Text generation
B. Data enhancement
Answer: A
C. Image generation
D. Image recognition
Q53. Regarding backpropagation, is the following statement wrong?
A. Backpropagation will go through the activation function

Answer: D
B. Backpropagation can only be used in feedforward neural networks
Q57. Recurrent neural network can capture dynamic information in serialized data.
C. Backpropagation can be combined with gradient descent algorithm to update network weights
A.True
B.False
Answer: B
Answer: A
Q54. What is wrong about the following statement about convolutional neural networks?
Q58. What are the elements of artificial intelligence?
A. Common pooling layers include maximum pooling and average pooling.
A. Algorithm
B. The convolution kernel cannot be used to extract global features of the picture.
B. Data
C. When processing the image, the image is convolved by scanning the window.
C. Scene
D. Convolutional neural networks can include convolutional layers, pooling layers and fully connected
D. Computing power
layers.
Answer: ABCD
Answer: B
Q59. The current stage of artificial intelligence is still in the stage of weak artificial intelligence.
Q55. All convolution kernels of the same convolution layer in the convolutional neural network are
weight-sharing. A.True
B.False D. Data Science
Answer: A Answer: ABCD
Q60. Jobs that are highly repetitive and require weak social skills are the jobs most easily replaced by AI. Q65.Python supports coexistence of multiple versions.
A.True A.True
B.False B.False
Answer: A Answer: A
Q61. Which of the following platforms does HiAI Service include? Q66. Which is correct about Python?
A. Quick service smart platform A. Exceptions in Python can only be thrown by the system
B. Xiaoyi Dialogue Development Platform B. Use the try statement to catch exceptions
C. IOS application development platform C. Users can customize exceptions
D. Android application development D. Abnormal will not terminate the execution of the program
Answer: ABD Answer:BC
Q63. Which of the following statements about iterators and generators are correct?
A. Only the _iter_ method is implemented in the iterator object Q67. Which of the following expressions can match 110 in "My phone number is 110, please find me"?
B. yield is the same as return A.re.match("\d+", str1)
C. List is an iterator B.re.search("\d+", strl)
D. The generator will save the current state C.re.split("\d+", str1)
D.re.match("[0-9]+", str1)
Answer: CD
Answer: B
A. Artificial Intelligence Q68. Is the statement about *args in function parameters correct?
B.APP A. *args receives parameters in the form of a dictionary
C. Website development B.*args can be defined before other parameters of the function
C.*args must appear at the end of all parameters C. The normal distribution takes the maximum value at the mean.
D.*args can receive multiple parameters D. The larger the standard deviation, the slower the normal distribution curve.
Answer: CD Answer: A
Q69. Methods that start with a double underscore in Python classes may be private methods. Q73. Optimization problems can be divided into the following categories according to the constraints?
A.True A. Unconstrained problem
B.False B. Equality constraint problem
C. Inequality constraint problem
Answer: A D. Semi-constrained problem
Q70. Which statement is correct? Answer:ABC
A. If the function has a prescription guide number at a certain point, its directional derivative is unique.
B. The extreme point of the function must be a stationary point Q74. A matrix corresponds to a linear transformation, which is realized by matrix multiplication.
C. The stagnation point of the function must be the extreme point A.True
D. If the function has a gradient at a certain point, its gradient is unique. B.False
Answer: AD Answer: A
Q71. The covariance matrix is a symmetric matrix, and the elements on the diagonal are the covariances Q75. The probability of occurrence of a solution with a known cause is usually called conditional
that change randomly. probability or prior probability.
A.True A.True
B.False B.False
Answer: A Answer: A
Q72. What is wrong about the description of the normal distribution? Q76. What is the built-in optimizer in keras?
A. The larger the standard deviation, the steeper the normal distribution curve. A. SGD
B. A random variable t that obeys the standard normal distribution, its mean and variance are both 0. B. Adaboost
C. Adadelta B. Extract image features
D. Adam C. Act as a classifier
D. Increase nonlinearity
Answer: ACD
Answer: A
Q77. The following are activation functions
A. Adam Q81. TensorFlow has released versions
B.sigmoid A.1
C. Relu B.2
D.softmax C.0.1
D.3
Answer: BCD
Answer:ABC
Q78. model. sumnary in Keras (the method used to print the network overview)
A.True Q82. For the products of an automobile company, factory A accounts for 40%, factory B accounts for
60%, the defective product rate of A is 1%, and the defective product of B
B.False
The rate is 2%. Now a car is a defective product. What is the probability that A will be produced?
A.0.25
Answer: A
B.0.5
C.0.2
Q79. Which of the following options are not supported by TensorFlow 2.0?
D.0.3
A.//
B.@
Answer: A
C.^
D.POW
Q83. The Naive Bayes algorithm does not require independent and identical distribution among sample
features.
Answer: C
A.True
Q80. When creating a network layer, what is the main purpose of adding a dropout layer?
B.False
A. Prevent overfitting
C.LSTM is also a recurrent neural network
Answer: B D. Recurrent neural network can be expanded according to the time axis
Q84. Which of the following options is not an evaluation indicator for parameter estimation? Answer: B
A. Consistency
B. Unbiased Q88. Which of the following are the activation functions of deep learning algorithms?
C. Adaptive A. ReLU
D. Effectiveness B.Sin
C. Tanh
Answer: C D. Sigmoid
Q85. Which of the following are machine learning algorithms? Answer: ACD
A. Supervised learning
B. Unsupervised learning Q89. Which of the following steps in deep learning are automatically completed by the model?
C. Reinforcement learning A. Model training
D. Semi-supervised learning
Answer: ABCD C. Analysis and positioning tasks
Q86. The higher the validity of the point estimation, the higher the degree of dispersion of the
estimated t.
Answer: AD à ACD
A.True
B.False
Q90. Which statement is wrong about the adversarial generative network?
A. The discriminant model needs to input labeled real samples

Answer: B
B. Adversarial generative network includes generative model and discriminant model
C. The adversarial generation network cannot be used to generate new pictures

Q87. What is wrong about the following statement about recurrent neural network?
D. The input to the generated model can be random numbers
A. Recurrent neural network can be abbreviated as RNN
B. LSTM cannot solve the problem of vanishing gradient

Answer: C
Q91. Which of the following activation functions can solve the vanishing gradient problem well? D. As the output value of the generative model
A. Relu
B. Tanh Answer: C
C. Softsign
D. Sigmoid Q95. In what year was "artificial intelligence" first proposed?
A.1956
Answer: A B.1946
C.1950
Q92. In which of the following fields is deep learning widely used? D.1959
A. Decision-making assistance
B. Voice processing Answer: A
D. Image processing Q96. In a neural network based on connectionism, each node can express a specific meaning.
A.True
Answer: ABCD B.False
Q93. Which statement about the optimizer is correct? Answer: A
A. Adagrad optimizer learning rate is not updated automatically.
B. The common feature of SGD, MBGD and momentum optimizer is that each iteration is updated with Q97. Which school does neural network research belong to?
the same learning rate
A. Symbolism
C. There is no connection between the Adam optimizer and the Adagrad optimizer.
B. Connectionism
D. The learning rate of the momentum optimizer does not need to be manually set
C. Behaviorism

Answer: B
Answer: B
Q94. In the adversarial generative network, where should the labeled data be placed?
A. As the input value of the generative model

Q98. Which of the following is not an application of image recognition service?
A. Smart photo album
B. Scene analysis
C. Target detection Q102. How to get the help information of the function?
D. Speech synthesis A.dir()
B.help()
Answer: D C.del()
D.id()
Q99. Which of the following is not the core of odelArt s?
A. Cloud Answer: B
B. End
C.API Q103. Methods that start with double underscores in Python classes may be private methods.
D. Side A.True
B.False
Answer: C
Answer: A
Q100. Which of the following functions does the data management module of modelArts include?
A. Data version management Q104. What variables can only act on the current function?
B. Upload data A. Local variables
B. Anything
C. Create a data set C. Global variables
D. Annotate data D. Module variables
Answer: ABCD Answer: A
Q101. The sys module is responsible for the interaction between the program and the operating system, Q105. Which statement about the data structure in Python is correct?
and provides an interface for accessing the bottom layer of the operating system.
A. float is an immutable data type.
A.True
B. Any tuple can be defined using ().
B.False
C. Neither sets nor dictionaries can use subscripts to get elements.
D. The strings defined using single, double and triple quotation marks are exactly the same.
Answer: B
Answer: BD
Answer: C
Q106. The role of the decorator is to add new functions to the function without changing any of the
function's properties.
Q110. What is the way to view tensor dimensions in TensorFlow 2.0?
A.True
A.dtype
B.False
B.ndim
C.device
Answer: A
D.dimens
Q107. Which of the following statements about tensors is correct?

Answer: B
A. A matrix is a second-order tensor
Q111. The mechanism of graphs and sessions is eliminated in TensorFlow 2.0
C. Tensors of order 4 and above, like vectors of 4 dimensions and above, have no direct geometric
A.True
meaning
B.False
D. The operations and model optimization processes in TensorFlow are all based on tensors
Answer: A
Answer: ABD
Q112. Which of the following attributes does TensorFlow 2.0 do not support creating tensors?
Q108. The determinant of square matrix A is a scalar.
A.constant
A.True
B.create
B.False
C.fill
D.zeros
Answer: A
Answer: B
Q109. When a function is at a certain point in the variable space, which direction has the greatest rate of
change?
A. Minimum directional derivative Q113. Why is TensorFlow becoming the mainstream framework for deep learning?
B. Derivative A. Automatic derivative
C. Gradient B. Back-end development
D. Directional derivative C. Deep Learning API
D. GPU parallel acceleration B.False
Answer: ABCD Answer: A
Q114. In what year was TensorFlow2.0 released? Q118. If the accuracy of the model on the training set is very high, but the accuracy on the validation set
is very low. can
A.2017
What methods are taken to improve the model?
B.2018
A. Add a regularization term after the loss function
C. 2019
B. Dropout method
D.2015
C. Increase the amount of training set data
D. Early stopping
Answer: C
Answer: ABCD
Q115. TensorFlow2.0 version is the second version of TensorFlow iteration.
A.True
Q119. Which of the following models can be used for classification?
B.False
A. Decision tree
B. Linear regression
Answer: B
C. Logistic regression
D. Support Vector Machine

Q116. TensorFlow2.0 supports GPU acceleration mainly in the parallel operation of matrix addition,
subtraction, multiplication, and division.
Calculated acceleration. Answer: ACD
A.True
B.False Q120. Which of the following statements about supervised learning is correct?
A. Supervised learning is a rule-based algorithm
Answer: A B. Decision tree is a supervised learning
C. Supervised learning can be trained without labels
Q117. The three main advantages of Keras interface are user-friendly, modular and combinable, and D. Supervised learning cannot use cross-validation for training
easy to expand.
A.True
Answer: A à B
A large number of optimized neural network algorithm models are built-in to allow users to easily use
deep learning technology.
Q121. What is wrong about the SVM algorithm
Live scheduling provides model training, evaluation and prediction in an on-demand service-oriented
A. SVM can be used to solve classification problems
manner.
B. Support vector is the vector point closest to the hyperplane
A. ModelArts
C. SVM algorithm can use multiple kernel functions
B.GES
D. The SVM algorithm cannot be used to process nonlinear data sets
C.Hilens
D.DLS
Answer: D
Answer: A
Q122. In the classic convolutional neural network model, what hidden layer does the softmax function
follow?
A.DIS
B.MRS
B. Pooling layer
C.HiAI
D. ModelArts
D. All of the above
Answer: B à AB
Answer: C
Q126. Which of the following conditions must be met to run a Python program?
Q123. Which of the following is the offline mode in HiAI?
A. Windows system
A. HiAI Service
B. Compliance with PEP8
B.HiAI Foundation
C. Comply with grammatical format
C.HiAI Engine
D. Python interpreter
D.HiAI IDE
Answer: DC
Answer: B
Q127. Is the Python IO operation correct?

Q124. Which service is based on the one-stop deep learning platform service provided by Huawei
Cloud’s powerful high-performance computing? A. After the file operation is completed, use the close method to close it.
B. You can use the context manager to open the file for operation.
C. In r mode, you can use the write method to write data to the file. A.
D. Use the flush method to write data into memory 7
Answer: B 4
Q128. Which of the following is correct? 1
A. The two object usage judgment results generated by shallow copy and deep copy are both True B.
B. The two objects of shallow copy use is to judge the result as True 7
C. The result of using the two objects of deep copy is False 6
D. The two objects of deep copy use is to judge the result as Irue 4
Answer: AB 2
C.
Q129. 7
i=8 5
while i>0: 4
i -= 1 2
if i == 6: D.
continue 6
print (i) 3
if i == 2: 1
break Answer: C
What is the output? Q130. A vector is just a number.
A.True
B.False A. Training
B. Export
Answer: B C. Evaluation
D. Forecast
Q131. What methods are not included in the built-in metrics of tf. keras. metrics?
A. Accuracy Answer: ABCD
B. Recall
Q135. K-fold cross-validation refers to dividing the test data set into K sub-data sets.
C.mean A.True
D.Sum B.False
Answer: D Answer: A
Q132. What are the main functions of the matplotlib library? Q136. Which statement about gradient descent is correct?
A. Drawing images A. When using gradient descent, the learning rate should be consistent.
B. Build a network model B. Global gradient descent is faster than stochastic gradient descent.
C. Used to process matrix operations C. Stochastic gradient descent is sensitive to noise points.
D. Data analysis D. Adam is not a gradient descent algorithm.
Answer: A Answer: C
Q133. When creating a network, another convolutional layer cannot be directly connected after the Q137. Which statement about the standard RNN model is correct?
convolutional layer.
A. There is no many-to-many model structure
A.True
B. There will be a problem of attenuation of long-term transmission and memory information
B.False
C. There is no one-to-one model structure

Answer: B
D. Do not consider the time direction when backpropagating
Q134. Which of the following options belong to the estimator package method under keras?
Answer: B
Q138. What are the commonly used loss functions in deep learning? A.list (set(list))
A. Mean square error loss function B.int (float)
B. L1 loss function C.str(int)
C. Self-declining loss function D.list(tuple(list))
D. Cross entropy error loss function
Answer: DC
Answer: ABD
Q142. Which statement about the matrix is incorrect?
Q139. Artificial intelligence is the research and development of theories, methods and applications for A. Any matrix multiplied by the identity matrix equals itself
simulating, extending and expanding human intelligence.
B. The transpose of the orthogonal matrix A is equal to the inverse of A
A new technological science using the system.
C. Non-square matrices also have inverse matrices
A.True
D. The transpose of symmetric matrix A is still A itself
B.False
Answer: C
Answer: A
Q143. Singular value decomposition can be performed on non-square matrices.

Q140. Which of the following statements about HiAI Foundation is incorrect?
A.True
A. Provides an acceleration API that has nothing to do with processor hardware. Application vendors and
B.False
developers can
Accelerate model calculation and operator calculation on the acceleration system.

Answer: A
B. Provide commonly used artificial intelligence business function API, which can run efficiently on
mobile devices.
C. Supports basic operator calculation interfaces, but does not include interfaces such as convolution, Q144. Which of the following evaluation indicators belong to the regression algorithm?
pooling, and full connection.
A. Recall rate
D. Support AI model management interfaces such as model compilation, model loading, model running,
and model destruction. B. Accuracy
C. Confusion matrix
Answer: A D. Mean square error
Q141. What is the operation that will not change the data after the following transformation? Answer: D
Q145. Which of the following options are the basis for the decision tree to divide nodes? D.HiAI
A.ID3
B. Information Entropy Answer: AC
C.CART
D. Gini coefficient Q149. What will the following code output?
If -1:
Answer: ABCD à BD
a=True+1
Q146. Is the following statement about neural networks wrong? b=False+True
A. There is no connection between neurons in the same layer of the feedforward neural network
B. As the number of hidden layers of the neural network increases, the classification ability of the model else:
gradually weakens

a=True-False
D. The limitation of a single-layer perceptron is that it cannot solve the XOR problem
b=False+I
Answer:BC
print(a,b)
Q147. The data annotation function in ModelArts only supports the modification of a single image tag,
A.True
and does not support batch modification.
B.False
Change the picture label.
C.2 1
A.True
D.1 1
B.False
Answer: C
Answer: B
Q150. Which of the following options does not support TensorFlow2.0 for dimension transformation is
the attribute.
A. MRS
A.squeeze
B. ModelArts
B.transpose
C.DIS
C.gathed
D.reshape
Answer: ABD
Q151. The methods that can be used to check whether it is a tensor in TensorFlow 2.0 are:
A.dtype
B.is_ tensor
C.isinst ance
D.device
Answer:BC
Q152. What data type does TensorFlow use?
A.Vector
B.Tensor
C.Scalar
D.Matrix
Answer: ABCD
Q153. Which of the following descriptions of machine learning is correct?
C. Deep learning and machine learning are both artificial intelligence but have no relationship with each
other
Answer: A
SHW – KOM IPB

Banksoal Digabungkan

Uploaded by

Copyright:

Available Formats

Banksoal Digabungkan

Uploaded by

Document Information

Original Description:

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

Banksoal Digabungkan

Uploaded by

Copyright:

Available Formats

1. Which of the following are sub-fields of AI?

D. Natural language processing

A. Weak artificial intelligence

D. Strong artificial intelligence

C. Open source framework

9. Which of the following aspects belong to Huawei's full-stack AI solution?----------------------- ABCD

A. Machines begin to understand, think and make decisions like humans

A. The foothold is in neuron networks and deep learning.

A. Natural language processing

2. Overview of machine learning

A. Decision tree is a supervised learning

B. Supervised learning cannot use cross-validation for training

C. Supervised learning is a rule-based algorithm

D. Supervised learning can be trained without labels

-------------------------------------------------- ----------------- BCD

-------------------------------------------------- --------- BCD

Combination, easy to expand. -------------------------------------------------- A

24. Is there a way to check whether it is a tensor in TensorF1ow2.0?

(Single choice) 26.Which company did Pytorch first launch?

5. Huawei's AI development framework MindSpore

Six, Atlas artificial intelligence computing platform

7. What is the value of HUAWEI HiAI empowering APP? -------------------------------ABCD

8. What are the advantages of the HiAI mobile computing platform?---------------------------------

8. Huawei Cloud Enterprise Intelligent Application Platform

9. HUAWEI CLOUD ModelArts is a one-stop development platform for AI development.

Artificial Intelligence HCIAv2.0 Question Bank

The question bank is expected to expire on October 1, 2020, please note

Q1. Which school does neural network research belong to?

D. None of the above

Q3. What are the main technical applications of artificial intelligence?

D. Natural language processing

Q4. Which of the following is not a ModelArts development type?

D. Standard model development

Q5. Which of the following are ModelArts development models?

Q8. HiAI Foundation does not support sparse model acceleration.

Q9. Which of the following features does Python have?

What is the value of c?

Q11. Which of the following can generate a dictionary?

Q12. Which of the following output results are wrong?

A.sorted([(‘b’ ,2),(‘a’ ,1),(‘c’,3),(‘d’,4)], key=lambda x:x[1])

Output result: [( a’,1), (3),(‘d’,4)]

B.print (*filter (lambda x: x%2==1, [1, 2, 3]))

D.print (*zip([1,2,3],["b"], ["A", "B","C"]))‘C’)

Output result: (1,’a’,""A";,"(2,b,‘B’)

Q13. What is the result of 5/2 in Python3?

The data in Q14.json format is a dictionary in Python.

Q15. Which of the following fields can Python be used in?

Q16. Which of the following identifiers are named incorrectly?

A. Establish a database connection

B. Execute SQL statement

D. Close the database connection

A. Poisson distribution can replace binomial distribution

B. Poisson distribution is an approximate calculation of binomial distribution when n is large and p is

D. Poisson distribution has no relationship with binomial distribution

A. The greater the expectation, the smaller the variance.

C. Expectation reflects the average level of random variables.

Q21. The first-order tensor is a scalar.

A. Discrete random variables have no distribution function