You are on page 1of 28

The speed of grid search must be faster than random search.

Answer: False

The biggest difference between logic regression and perceptron is that logistic regression uses the
Sigmoid activation function at the output layer, whereas perceptron uses the sign function. In addition,
neither of the two models can resolve the XOR problem.

Answer: True

HiLens can help small and micro companies realize smart access control.

Answer: True

Polynomial regression cannot resolve the nonlinear problem.

Answer: False

Softmax can’t be used for multiple classification problems.

Answer: False

The Adam optimizer in TensorFlow 2.0 is the built-in optimizer of the Keras module.

Answer: True

The HiAI mobile computing platform architecture supports various mainstream front-end frameworks,
such as TensorFlow and Caffe.

Answer: True

To install TensorFlow of the GPU version, you need to install CUDA and cuDNN.

Answer: True
In a feedforward network, neurons in the same layer are connected to each other.

Answer: False

Backpropagation makes gradient descent more efficient by means of dynamic programming.

Answer: True

In a medical image recognition task, we usually are more concerned about the recall ratio than the
precision ratio when evaluating model performance.

Answer: True

After the image label service is called, only one corresponding label value is returned.

Answer: False

Similar to hidden layer, the quantity of neurons at both an input layer and an output layer of a neural
network is a hyperparameter that can be adjusted.

Answer: False

Cifar10 is a built-in data set of Keras.

Answer: True

The generalization error will continue to decrease as the complexity of the model increases.

Answer: False

MindSpore supports GPU acceleration.

Answer: True
Voice recognition refers to the recognition of text data as audio data.

Answer: False

PyTorch supports GPU acceleration.

Answer: True

In the KNN model, a small K value may cause model underfitting.

Answer: False

Malformed data is not dirty data.

Answer: False

Factory A produces 60% of an automotive company’s products, and factory B accounts for 40%. The
defective rate of factory A is 1%, and that of factory B is 3%. If a vehicle is defective, what is the
possibility that the vehicle is produced by factory A?

Answer: B. 1/3

Which of the following statements about dropout is incorrect?

Answer: D. dropout can’t be added during inference.

Which of the following reasons is the main reason why deep learning developed rapidly after 2010?

Answer: D. Improvement of computing power

Which of the following methods cannot resolve overfitting in a decision tree?

Answer: D. Use the integration algorithm to set up an AdaBoost model.


Which of the following network layers does not need to be included in a convolutional neural network?

Answer: C. Pooling layer

Which of the following is not a feature of the Atlas 500 AI edge stations?

Answer: C. High energy consumption

What determines the number of neurons in the input layer of the neural network?

Answer: Data dimension

What operation does “isinstance” represent in TensorFlow 2.0?

Answer: C. Determine whether it is Tensor

Which of the following schools does the expert system belong to?

Answer: A. Symbolism

Input_shape = (4, 28, 28, 1)

X = tf.random.normal(input_shape)

Y = tf.keras.layers.Conv2D(filters=2, kernel_size=3, strides=(1, 1), padding=’valid’, activation=’relu’,

Input_shape=input_shape[1:])(x)

print(y.shape)

What is the output of the preceding code?

Answer: B. (4,26,26,2)
Without Padding, use a 3X3 convolution kernel to convolve a 5X5 image with stride 1. Which of the
following options is the size of the output image?

Answer: A. 3X3

When was the Turing Test proposed?

Answer: A. 1950

Which of the following options is the architecture adopted by Huawei Atlas series chips?

Answer: D. Da Vinci

In the process of training the neural network, which of the following methods is used to update the
parameters so as to minimize the loss function?

Answer: C. Gradient descent

Which of the following options is not the feature of ModelArts?

Answer: C. Support automatic tuning

In machine learning, how many support vectors can the SVM algorithm have?

Answer: D. Several

import numpy as np

n = np.arange(24).reshape(2,-1,2,2)

Which of the following is the return result of n.shape after running the preceding code?

Answer: A. (2,3,2,2)
Which of the following options is the artificial intelligence development framework launched by
Huawei?

Answer: C. MindSpore

Which of the following options is a commonly used loss function for classification problems?

Answer: B. Cross Entropy Loss Function

Which of the following problems can’t be handled by the perceptron?

Answer: C. XOR

Which of the following models is not suitable for the scenario where the covariance between data
features is large?

Answer: C. Naive Bayes

Which of the following options is Huawei’s artificial intelligence computing platform?

Answer: D. Atlas

A company wants to use the HiAI platform to develop its own APP. Which of the following IDEs can be
used in conjunction with the HiAI platform?

Answer: B. Android Studio

Which of the following options is the classification of the AI chip business from the application?

Answer: C. Training and Inference

Which of the following options belong to the activation function of the neural network?

Answer: A,B,C,D (All of the Above) (Check all letters: ABCD)


Which of the following statements about LSTM are correct?

Answer: A,B,C,D (All of the Above) (Check all letters: ABCD)

Which of the following options are the main advantages of the Keras interface in TensorFlow 2.0?

Answer: A,B,C

A. Convenient for users


B. Modular and combinable
C. Easy to expand

Which of the following options belong to the application scenarios of natural language processing?

Answer: B & D

B. Machine translation

D. Public opinion monitoring

Which of the following may cause overfitting?

Answer: B,C,D

B. The model capacity is large

C. The training set includes a large amount of dirty data

D. The number of training sets is small

Which of the following are common Python computer vision libraries?

Answer: C & D

C. matplotlib

D. cv2
Which of the following options belongs to Huawei HiAI platform content?

Answer: A, B, D

A. HiAI Service
B. HiAI Foundation

D. HiAI Engine

What are the elements of AI?

Answer: A,B,C,D

Which of the following are major functions of Mind Expression in MindSpore?

Answer: A,B,C,D

Which of the following are the problems you may encounter when training a generative adversarial
network?

Answer: A,B,C,D

Which of the following options are Huawei products?

Answer: A,B,C

Which of the following are AI applications?

Answer: A,B,C,D

Which of the following statements about TensorFlow are correct?

Answer: A,B,C
Compared with the Sigmoid activation function, which of the following options are the advantages of
ReLU?

Answer: A & D

Which of the following are deep learning frameworks?

Answer: A,B,C,D

Backpropagation makes gradient descent more efficient by means of dynamic programming.

Answer: True

In some image and voice fields, AI performs better than humans, so artificial intelligence has entered
the stage of strong artificial intelligence.

Answer: False

Automatic graph segmentation is one of the key technologies of MindSpore.

Answer: True

One of the disadvantages of deep learning is that the interpretability is very weak, but it is possible
to visualize the model using methods such as Local Interpretable Model-Agnostic Explanations
(LIME).

Answer: True

Atlas series chips need the support of CUDA software package like other GPU accelerator cards.

Answer: False

The speed of grid search must faster than random search.


Answer: False

In a medical image recognition task, we usually are more concerned about the recall ratio than the
precision ratio when evaluating model performance.

Answer: True

Voice recognition refers to the recognition of text data as audio data.

Answer: False

In convolutional neural networks, the convolutional layer can reduce the dimension of feature maps.

Answer: True

Laplace smoothing can be used to address the “zero probability” problem in the naïve Bayes model.

Answer: True

Knowledge graph can be applied to social networking, gene sequencing, and risk control.

Answer: True

Softmax can’t be used for multiple classification problems.

Answer: False

Da Vinci architecture is the core of the Ascend AI computing engine and AI processor.

Answer: True

In a feedforward network, neurons in the same layer are connected to each other.

Answer: False
The generalization error will continue to decrease as the complexity of the model increases.

Answer: False

If the metrics parameter is not specified in Model.compile, no less value will be output.

Answer: False

Polynomial regression cannot resolve the nonlinear problem.

Answer: False

The biggest difference between logistic regression and perceptron is that logistic regression uses the
Sigmoid activation function at the output layer, whereas perceptron uses the sign function. In
addition, neither of the two models can resolve the XOR problem.

Answer: True

The HiAI mobile computing platform architecture supports various mainstream front-end
frameworks, such as TensorFlow and Caffe.

Answer: True

To install TensorFlow of the GPU version, you need to install CUDA and cuDNN.

Answer: True

Without Padding, use a 3X3 convolution kernel to convolve a 5X5 image with stride 1. Which of the
following options is the size of the output image.?

Answer: A. 3X3

Which of the following statements about PyTorch is incorrect?

Answer: D. PyTorch uses the Lua programming language.


Which of the following learning methods is most suitable for training machine learning models of
chessboard games?

Answer: D. Reinforcement learning

Which of the following is the output range of the ReLU function?

Answer: A. [0,+∞]

Which of the following is not an integrated learning policy in machine learning algorithms?

Answer: D. Marking

In machine learning, how many support vectors can the SVM algorithm have?

Answer: D. Several

Input_shape = (4,28,28,1)

x = tf.random.normal(input_shape)

y = tf.keras.layers.Conv2D(filters=2, kernel_size=3, strides=(1,1), padding=’valid’,


activation=’relu’, input_shape=input_shape[1:])

(x)

print(y.shape)

What is the output of the preceding code?

Answer: B. (4,26,26,2)

In the process of training the neural network, which of the following methods is used to update the
parameters so as to minimize the loss function?
Answer: C. Gradient descent

Which of the following is a function of HiAI Foundation?

Answer: B. Quickly convert and migrate existing models

When was the Turing Test proposed?

Answer: A. 1950

Which of the following options is the artificial intelligence development framework launched by
Huawei?

Answer: C. MindSpore

Which of the following is not a feature of the Atlas 500 AI edge stations?

Answer: C. High energy consumption

Which of the following is not the model deployment and application feature of the ModelArts
platform?

Answer: B. Low throughput and low latency

adadelt=tf.keras.optimizers.Adadelta()

model.compile(loss=’binary_crossentropy’,

optimizer=’adam’,

metrics=[‘accuracy’])

Which optimizer is used in the above code?

Answer: A. Adam
Which of the following options is the architecture adopted by Hauwei Atlas series chips?

Answer: D. Da Vinci

Which of the following is a generative model?

Answer: C. Naive Bayes

What kind of operation does “concat” represent in TensorFlow 2.0?

Answer: A. Merge tensor

When tf.keras is used to build a network model, which of the following statements about
tf.keras.Model and tf.keras.Sequential is incorrect?

Answer: D. If a sharing layer exists on the network, tf.keras.Sequential is recommended for


building a model.

Which of the following models is not suitable for the scenario where the covariance between data
features is large?

Answer: C. Naive Bayes

How many images are included in the MNIST handwritten digit database in the lab guide?

Answer: C. 7,000

Which of the following options makes the GPU suitable for the acceleration of neural networks?

Answer: B. Parallel computing ability

Which of the following is the representative of behaviorism?

Answer: D. Reinforcement learning


Which of the following statements about ModelArts is incorrect?

Answer: A. Supports massive data preprocessing and automatic labeling.

Which of the following statements about dropout is incorrect?

Answer: D. dropout can’t be added during inference

What types of services can deployed on ModelArts?

Answer: A,B,C

Which of the following are AI applications?

Answer: A,B,C,D

Compared with the batch gradient descent method, which of the following options are the
advantages of the Mini batch gradient descent method?

Answer: A & B

Suppose the following scenario: Now we need to classify the data, and the data features are all
continuous values. Which of the following models are suitable for this scenario?

Answer: B & C

What are the elements of AI?

Answer: A,B,C,D

Which of the following are major functions of Mind Expression in MindSpore?

Answer: A,B,C,D
Which of the following are the problems you may encounter when training a generative adversarial
network?

Answer: A,B,C,D

Which of the following are deep learning frameworks?

Answer: A,B,C,D

Which of the following options belong to data preprocessing?

Answer: A,B,C,D

Which of the following are the major modules of the Ascend AI processor?

Answer: A,B,C,D

Regarding the standard RNN model statement, which of the following options are correct?

Answer: A & C

Which of the following statements about TensorFlow are correct?

Answer: A,B,C

Which of the following options belong to the Huawei Cloud EI service?

Answer: A,B,C,D

Which of the following options are Huawei products?

Answer: A,B,C

Which of the following statements about LSTM are correct?


Answer: A,B,C,D

Which of the following are applications scenarios of EI Intelligent Twins?

Answer: A,B,C,D

Operations and computation graphs of TensorFlow can be run without a session.

Answer: False

A matrix A can be decomposed into a product of three matrices using singular value decomposition.
Answer: True

Deep learning algorithms can be both supervised and unsupervised.

Answer: True

The print function in Python 3 is enclosed in parentheses (()).

Answer: True

In TensorFlow, a graph contains nodes and edges. The former represent data operations and the later
represent data and its dependency.

Answer: True

The perception is equivalent to a classifier and it could perform multi-task classification.

Answer: False

Data in TensorFlow is represented by tensors.


Answer: True

Functions are organized code segments which cannot be reused to implement single or related
operations.

Answer: False

Python tuples are enclosed in parentheses (()) and the internal elements are separated by semicolons
(;).

Answer: False

Matrix A and Matrix B can be added only if they have the same number of rows and columns.

Answer: True

The determinant of a N*N matrix is a scalar.

Answer: True

Deep learning is a class of machine learning algorithms that uses neural networks to extract features
from the raw input.

Answer: True

Huawei believes that the value of AI is reflected in addressing enterprise business problems. Regarding
the development of AI, guided by industry solutions, Huawei focuses on basic technologies and the
enabling platform, and works with industry leaders to develop industry solutions.

Answer: True

The Python author designed the restrictive syntax to filter out bad programming habits (like failing to
add an Indentation on the right of the next line of the if statement).

Answer: True
Convolutional neural networks (CNNs) can extract features from images automatically.

Answer: True

Continue statements in Python are used to skip the rest of the statements in the body of loop for the
current iteration and jump to the beginning of the loop for next iteration.

Answer: True

Lists in Python can be delimited by square brackets ([ ]), with the first element at index 0.

Answer: True

time.time() function can measure the number of seconds since the epoch as a floating point value.

Answer: True

Principal Component Analysis (PCA) is a statistical method. It uses orthogonal transformation to convert
a set of possibly correlated variables into a set of values of linear variables called principal components.

Answer: False

Recurrent neural networks are more suitable for image recognition problems.

Answer: False

Which of the following can be the input of a fully connected layer?

Answer: B. A fixed-length vector


If the input of a softmax function is [1,2,4,2,1], which of the following option may be the output?

Answer: C. [0.04, 0.10, 0.72, 0.10, 0.04]

Decision tree is one of the most popular machine learning algorithms and can be used for both
classification and regression problems. How do we choose the best attribute in ID3?

Answer: C. The attribute with the highest information gain

Information theory is a branch of applied mathematics. It studies the quantification of information in a


signal. Who was the first to put forward the concept of information entropy?

Answer: A. Claude Elwood Shannon

Which is not an optimization method in deep learning?

Answer: C. Principle component analysis (PCA)

Which of the following Python identifier is invalid?

Answer: D. 1var

What is the purpose of adding a pooling layer when creating a neural network?

Answer: D. Reduce the size of feature maps

Which of the following format can define a dictionary in Python?

Answer: C. (“Jane”:London, ”John”:London)

Which of the following type of RNN can be used for machine translation?

Answer: D. Many to many


Which of the following activation functions cannot effectively solve the vanishing gradient problem?

Answer: A. Tanh

Which of the following statements about normal distribution is not true?

Answer: C. The larger the standard deviation, the steeper the normal distribution curve.

Which is not an automatic hyperparameter search algorithm?

Answer: B. Stochastic gradient descent

Which of the following about matrix is not true?

Answer: C. The diagonal matrix does not have an inverse.

Which of the following is not a supervised learning algorithm?

Answer: B. Principle component analysis (PCA)

Which of the following can avoid overfitting during model training?

Answer: D. Early stoping

Which of the following is unchangeable in Python?

Answer: A. Tuple

Which of the following is not a deep learning development framework?

Answer: D. BAFA
Which of the following is not a Python re regular expression module function?

Answer: B. update

Which of the following methods cannot get data into the TensorFlow program?

Answer: C. Editing a file format reader

Which of the following is not a TensorFlow library module?

Answer: C. tf.boost

With regard to the relationship between frequency and probability, which of the following statements is
true?

Answer: A. Frequency stabilizes as a value close to probability as the number of repeated experiments
increases.

Which of the following statement about convolutional neural networks(CNNs) is not true?

Answer: B. A pooling layer contains a set of filters whose parameters need to be learned.

Which of the following is used as the gate function is LSTM?

Answer: C. Sigmoid function

Which of the following can be used to solve regression problems?

Answer: A. Linear regression

For object detection problems, which of the following neural network model should be considered first?

Answer: A. CNN
Which of the following school does reinforcement learning belong to?

Answer: D. Behaviorism

Which of the following is the computational model of TensorFlow?

Answer: D. Computational graph

Which of the following is not a TensorFlow session?

Answer: A. Multiple POST queries

Which of the following statement about the fetch operation in TensorFlow is true?

Answer: A. Only when the graph execution is triggered, you can fetch the tensor values you want.

When was “AI” first proposed?

Answer: B. 1956

Which of the following cloud services are provided by visual recognition of Huawei Cloud EI?

Answer: ABCD

A. Image recognition
B. Content moderation
C. Optical character recognition (OCR)
D. Image processing

What are the commonly used activation functions?

Answer: ACD

A. relu

C. sigmoid
D. tanh

According to constraints, what are the types of the optimization problem?

Answer: ABC

A. Equality constraint
B. No constraint
C. Inequality constraint

Which of the following statements about Python index are true?

Answer: BD

B. Default index from right to left is numbered from -1

D. Default index from left to right is numbered from 0

Which of the following statements about Python are true?

Answer: ABCD

A. Python is often nicknamed “glue language” because it can easily glue together various
modules developed in other languages.
B. Python is a kind of free and open-source software.
C. The Python syntax is simple and clear, characterized by using blank characters for indentation
D. It was created by Guido van Rossum, the Dutchman, in 1989, and it was first released in 1991.

What are the research areas in AI?

Answer: ABCD

A. Computer vision
B. Speech recognition
C. Natural language processing (NLP)
D. Machine learning
Python is an object-oriented programming language. Which of the following are Python objects?

Answer: ABCD

A. Character string
B. Number
C. Module
D. Function

Which of the neural network architectures can implement weight sharing?

Answer: BD

B. Recurrent neural network (RNN)

D. Convolutional neural network (CNN)

What are the implementation modes of TensorFlow?

Answer: BD

B. Distributed mode

D. Single-Node mode

Suppose X and Y are random variables and C is a constant, which of the following about the variance
properties is not true?

Answer: BD

B. D(XY) = D(X)D(Y)

D. D(X+Y) = D(X) + D(Y)

AI Overview
1. (True or False) By "all AI scenarios", Huawei means different deployment scenarios for AI, including
public clouds, private clouds, edge computing in all forms, industrial IoT devices, and consumer devices.
Answer: True

2. (Multiple-answer question) Which of the following are AI application fields? Answer: Smart
household, Smart healthcare, Smart City, Smart Education

Python

1. (True or false) Python2 and Python3 are frequently used. Answer: True

2. (Multiple-answer question) In which fields can Python be used? Answer: Image recognition, Website
development, Data mining, App development

3. (True or false) Use indentation to distinguish code blocks in Python. Answer: True

4. (True or false) Variables in Python can be classified into global variables and module variables based
on their application scope. Answer: False

5. (True or false) The data of the character string type cannot be changed. Answer: True

6. (Single-answer question) Which of the following statements are correct? Answer: Character strings
and values in Python cannot be modified

7. (Single-answer question) Which of the following operations are correct? Answer: List sort: list.sort()

8. (Multiple-answer question) Use slices to obtain all even numbers from the tuple t=(1, 2, 3, 4, 5, 6, 7).
Answer: t[1:6:2] & t[-2:0:-2]

9. (True or false) Both == and is can be used to determine whether the values are equal. Answer: True

10. (Single-answer question) Which of the following statements are correct? Answer: Duplicate
elements in the collection are deleted.

11. (Single-answer question) Which of the following code execution results are displayed? Answer: B

12. (Single-answer question) Which of the following codes output? Answer: 7 4 1 -2

13. (Multiple-answer question) Which of the following statements about functions and objects are
correct? Answer: In python, a function is also an object & Functions can be created using def and
lambda in Python

14. (Single-answer question) Which of the following statements about classes are correct? Answer: In
Python, you can quickly create classes through inheritance.

15. In Python, objects are subclasses of all classes by default.


(Multiple-answer question) Which of the following operation modes are available for file operations?
Answer: a, w, *b

16. (Single-answer question) Which of the following errors cannot be captured using the try statement?
Answer: | = [1,2,3

Machine learning

1. (True or false) Gradient descent iteration is the only method of machine learning algorithms.
Answer: False
2. (Single-answer question) Which of the following algorithms is not supervised learning?
Answer: K-means

Deep Learning

1. (True or False) Compared with the recurrent neural network, the convolutional neural network is
more suitable for image recognition. Answer: True

2. (True or False) GAN is a deep learning model, which is one of the most promising methods for
unsupervised learning of complex distribution in recent years. Answer: True

3. (Single-choice) There are many types of deep learning neural networks. Which of the following is not
a deep learning neural network? Answer: Logistic

4. (Multi-choice) There are many important "components" in the convolutional neural network
architecture. Which of the following are the convolutional neural network "components"? Answer:
Activation Function, Convolutional kernel, Pooling, Fully connected layer

1. (True or False) In TensorFlow 2.x, eager execution is enabled by default. Answer: True

2. (Single-choice) Which of the following statements about tf.keras.Model and tf.keras.Sequential is


incorrect when the tf.keras interface is used to build a network model?

Answer: tf.keras.Sequential is recommended for model building when a sharing layer exists on the
network.

Deep Learning Open-Source Framework MindSpore


1. (Single-choice) In MindSpore, which of the following is the operation type of nn? Answer: Network

AI Computing Platform Atlas

1. (Single-choice) What are the main applications of Ascend 310? Answer: Model Inference

AI Development Platform for Smart Quiz

1. (True or false) Learning AI requires data foundation.

Answer: Correct

2. (Multiple-answer question) Which of the following are AI application industries?

Answer: City Brain, Smart City, Smart Furniture, Unmanned driving

3. (Single-answer question) Which of the following is not a Huawei HiAI module?

Answer: HiAI Framework

4. (True or false) The HUAWEI HiAI platform supports remote debugging of real devices.

Answer: Correct

5. (Multiple-answer question) What are the applications of HUAWEI HiAI Engine?

Answer: Table Identification, Keyword Extraction, Video Summary

Enterprise Smart Application Platform

1. (Multi-choice) Which of the following scenarios can EI be applied to? Answer:

Answer: Intelligent government, Intelligent city, Intelligent manufacturing, Intelligent financ

You might also like