1, Which of the following is not a lhyperparameer for model trainiing?

QA oatch_si1ze in a t raining neurall netwo rk

0 B_ in K-means
@ C_ in the 0Jne-dimeni1sicmall linear mgmssion model Y=wX+ o
0 D_ Number of trees in a random forest
e Congratula ions, you got it! My Score : 20 Point(s) Cmrect answe r : C
2, The core of an A l framework and one of the decisive factors of the programming paradigm is the automatic differentiation used by the A l framework. Which of the

following is used by MindSpore?

A Graph kernel fusion

':' 8- Operator
over1oading
@ C. Source to Source (S2S)

*'¥*
8 O_ Just-in-time compilation
I Previous I
O Congratulations, you got I My Score : 20 Point(s) Correct answer: C

3, Which of the following statements is true about deep learning?

@ A. A lgorihms are used to automa1Jically extract features.
0 B. GPU parallel computing provides faster computing if there are more network parnmeters.
0 C. Parameters at each layer have specific meanings.
0 D. The feature interpretability is strong.
e
Previous
Congratulations, you got it! My Score: 20 Point(s) Conect answer : A
4, Dirty data refers to data with quality problems. Which of the following statements is false about the data quality?
0 A. Noise: Data contains incorrect records or exceptions.
0 B. Incompleteness : Data lacks attributes or contains some missing values.
@ C. Unprocessed: Data for which feature engineering has not been performed.
0 D. Inconsistency : Data contains conflicting records.
I Previous I -
9Congratulations, you got it! My Score: 20 Point(s) Correct answer : C
5, Which of the following commands can be executed to use the code of TensorFlow 1.X in TensorFlow 2?
@ A import tensorflow compatv1 as tf
0 B. import tensorflowv1 as 1f
0 c_ import !ensorflow as tf
0 D_ import v1 as tf
Previous I -
9Congratulations, you got it! My Score: 20 Point(s) Correct answer: A
6, Which of the following statements is true about the low-latency design of CPUs?
'.) A. The low clock frequency reduces the latency

0 B_ Complex logic control units increase the latency

0 c_ CPU is equipped with few ALUs and caches_ The caches merge access requests to the same data and then access the DRAM, reducing the latency_
@ D. The powerful ALU can complete the computation in a short clock cycle
I -
Previous

O Congratulations, you got rt! My Score: 20 Point(s) Correct answer : D

7 Which of the following oommands is used to install TensorF low?

0 A. conda crnate install tensorflow
0 B. python tensorflow setup_ py
0 C_ conda setup_ py tensortilow
@ D. pip iinstalll te11so1rflow
Previous
O Congra ulations, you got iU My Score: 20 Point(s) Cmrect answer : D
8 Which of the following statements is false about machine learning?
0 A_ Model generalization capability: tne extent to which a learned model can oe applied h
@ B. Error: differnnce between the prndiction of a learned model on a sample and tne actu
0 c_ UndmTiitting: Tne training mror is large_
0 D_ Overfirtting: Tne training error is l.mge_
Previous
OWrong answer My Score : 0 Point(s) Conect answer : D
9, Which of the following is a lightweight and high-performance service module that helps MindSpore developers efficiently deploy online inference services

environments?

0 A MindlR
@ B. MindSpore Serving

0 C M1ndArmour
0 D. Mindlnsight
I Previous I

O Congratulations , you got it! My Score: 20 Pmnt(s) Correct answer : B

1O, Which of the following functions is used in HUAWEI CLOUD General Text OCR experiments?
@ A. ocr_client.request_ocr_service_base64
() 6- ocr_clientocr_service_base64
0 C. ocr_client.request_ base64
0 D. ocr.request_ocr_service_basel28
Previous

8Congratulations, you got it! My Score: 20 Point(s) Correct answer : A

11, concat_sample_ 1 = tf.random.no rma1([4 ,100,100,3])</br>concat_sample_2 = tf.random.norma1([ 40 ,100,100,3])</br>concated_sample_ 1 =

tf .concat([concat _sample_ 1,concat_sample_ 2],axis=O)<lbr>print (concated_sample_ 1.shape)<lbr>Which of the following is the output for this code?

0 A. (44,200,200,6)
@ B (44 ,100,100,3)
0c (40,100,100,6)

0 D (44,100,100,6)
I Previous I
O Congratulations, you got rt! My Score: 20 Point(s) Correct answer: B

12, Which of the following HUAWEI CLOUD services is NOT a basic Al platform serv
0 A. ModelArts
0 B. Huawei Hi ens
@ C. Relational Database Service (RDS)
0 D. Graph Engine Service {GES)
Previous
O Congratulations, you got it! My Score: 20 Point(s) Cmrect answer: C
14, Which of the following is not a. network opera.tor in MindSpore?
0 A. Conv2D
0 B. SGD
@ C. ControlDepend
0 D_ Softmax
eCongratulations, you got iU
Previous
My Score: 20 Point(s) Correct answer : C
13, Which of the following is NOT an .application of speech processing?
0 A Speech recognition
0 B. Text to speech
Q C_ Voiceprint recognition
@ D. Named entity rncognirtion

eCongratulafons, you got ru

Previous
My Score: 20 Point(s) Correct answer : D
15, Intelligent quality inspection is based on the cloud-edge-device synergy of deep learning algorithms. Which of the follow ing operations is performed on edge devi<

@ A Model !raining

B. Data labeling

-=: C. Model delivery

D_ Onsite inference

I evious I Ml§F
0Wrong answer My Score : 0 Point(s} Correct answer : D

16, Which of the following statements is false about the ReLU function?
A The ReLU function is not differentiable at x = 0 and a derivative is forcibly defined at this point
1 B. The ReLU function effectively alleviates the vanishing gradient problem.
'.J C_ The surface defined at the zero point of the ReLU function is not smooth enough in some regression probl
=1 D_ Compared with Sigmoid and tanh, the convergence of the ReLU function is slow.
Previous
OWrong answer My Score : 0 Point(s) Correct answer : [)
17 , Which of the following is the purpose of running ocr_ client = HWOcrClientAKSK (ak, sk, region) when using HUAWEI CLOU
@ A Initialize the client

0 B_ Transfer images for recognition

0 C Delete image data
0 D Encode image data
Previous I -

O Congratulations, you got m My Score: 20 Point(s) Correct answer : A

18 Which of the following is a method of compressing an Al model?

@ A. Low rank approximation
0 B. Network integration
Q C_ Data a1.1gmeritation
Q D_ Data volume mduction
Previous
•Congratulations, you got iU My Score: 20 Point(s) Correct answer : A
19 Which of the followi ng cells are predefined in MindSpore?
IIA. MAELoss
0 B_ TFRecord
D c_ cos
IID. WithGrad Cell
Previous
•Corlgratulations, you got iU My Score: 20 Point(s) Correct answer : AD
20 Which of the following are major advantages of Keras?
IIA. Easy to use
D 8- Low-level coding style
IIC. Modular and combinab le
IID. Easy to scale
Previous
•Congra ulations, you got iU My Score: 20 Point(s) Correct answer : ACD
21, Which of the following statements are true about grid search based on hyperparameter tuning?
Iii A Grid search exhaustive ly searches for all possible hyperparameter combinations to form a hyperparameter value grid
Iii B. Grid search is expens ive and time-consuming_
Iii C Grid search works well when there are relatively few hyperparameters
0 D_ Grid search suits neural networks well.
I Previous I -
9 Congratulations, you got it! My Score: 20 Point(s) Correct answer : ABC
22, Which of the following methods provided by TensorFlow 2.0 cannot be used to check whether an object is a tensor?
0 A is_ tensor
l!I B. isinstance
l!I C device
l!I D. tflypes
I Previous I -
eCongratulations, you got it! My Score: 20 Point(s) Correct answer: BCD
23 Which of the following are supported by data management on ModelArts?
IIA. wide rrang,e of data fmmats
IIB. Team labeling
IIC. lrntelligent data filltering
D D_ Automatic data mig ratiori
Previous
•Corigra ulations, you got itJ My Score: 20 Point(s) Coned answer : ABC
24, In TensorFlow 2..0, which of the following cannot be used to create an all-zero tensor?
IIA. tf zeros()
IIB. tffill([1J,O)
IIC. tf.zeros_frame()
D D. tf.zems_tuple{)
Previous
O Wrong answer My Score : O Point(s) Conect answer : CD
25, Which of the following statements are true about hidden layers?
IIA. The more hidden layers, tile stronger the identi1iicatio11 capability_
D B. The mme hidden layers, the better.
IIC. Exoessive hidde11 layers may ca 1.1se ovm1iitting.
D D. llnsuftiicient hidden layers may cause overfitting_
Previous
O CoHgratulations, you got iU My Score: 20 Point(s) Correct answer : AC
26, Which of the following statements are false about GANs?
llA. The discr iminator input is mainly noise and real sample data
D B. GANs me a type of framework . They train the generator and discriminator through an adversar ial process
D C. GANs are trained using the BP algorithm
ll D_ The input of the discriminator is mainly sample data provided by the generator.
Previous

OCongratulations, you got it! My Score: 20 Point(s) Correct answer: AD

27, Which of the following statements are true about decision trees?
0 A. decision tree is a binary tree structure
Bl B Each non-leaf node in a decision tree represents a test on a feature attribute; each branch represents the output of the feature attribute
class label.
0 C. Except the root node, all nodes in a decision tree are called leaf nodes.
Bl D_ Building a decision tree means selecting and measuring feature attributes and determining their topology_
Previous I -
O Congratulations, you got it! My Score: 20 Point(s) Correct answer: BO

28, Which of the following are part of the Huawei full-stack Al solution?
Iii A. TBE
D
B. open Euler
Iii C. AscendCL
Iii D. Ascend
Previous
•Congratulations, you got itJ My Score: 20 Point(s) Correct: answer : ACD
29, Which of the following products can be equipped with Ascend 31O Processors?
IIA. Atlas 200 DK Al developer kit
IIB. Atlas 500 Al edge station
IIC Atlas 200 Al acceleratm module
IID. Atlas 300 Al acceleratm card
Previous
eCongratulations, you got iifJ My Score: 20 Point(s) Coned answer : ABCD
30, Which of the following statements are true about data preprocessing?
l!I A. Data cleansing is a process of fi lling in missing values , as well as detecting and eliminating noise data and exceptions_
l!I B. Data dimension reduction is to prevent dimension explosion by deleting data features
D c_ Data standardization is to reduce noise data and improve model accuracy
D D_ Mad1ine learning outputs results through models_ Therefore , model training is more important than data prepmcessing
Previous I -
OWrong answer My Score : 0 Point(s) Correct answer : AC
31, Which of the following statements are true about data conversion?

r.11 A Image data processing, such as color space conversion, grayscale image conversion , and geometric conversion, 1s a part of data conversion.
r.11 B I n machine learning, features are often normalized to ensure that the value ranges of different variables input to the same model are the same
0 C_ When the data dimension is high, the existing variables need to be combined or converted to generate new features , such as averages _
0 D_ All data needs to be converted.
Previous I -

8Congratulations, you got it! My Score: 20 Point(s) Correct answer : AB

32, Which of the following are common ensemble learning algorithms in machine learning?
l!I A. Ra11dom forest
l!I B. GBDT
D C. Polynomial regression
1,11 D. Adaboost
I Previous
O Congratu lations, you got it! My Score: 20 Point(s) Correct answer: AIBD
33, Which of the following are common tensor operations in MindSpore?
LI A. otypeO
LI B. astype()
LI C. swirtcil()
LI D. abs()
Previous
0Wmng answer My Score : 0 Point(s) Correct answer : ABD
34, Which of the following statements are true about linear regression?

l!I A The error of linear regression is affected by many independent factors . According to the central l1m1t theorem (CLT), the error follows normal distnbution.
l!I B The loss function of linear regression can be obtained using the normal distribution function and maximum likelihood estimation (MLE)
0 C.The result of multivariate linear regression analysis may be a hyperplane in a high-dimensional space.
l!I D. Due to algorithm complexity, linear regression cannot use the gradient descent method to calculate the weight parameter if the loss function reaches the minimum value.
Previous

OWrong answer My Score : 0 Pomt(s) Correct answer: ABC

35, AIR eliminates model differences between different backends through unified operator IR definitions. You can coordinate tasks on all platforms (device, edge, and
cloud) based on the same model file .

@ True

() False

Previous I Ml§F
eCongratulations, you got ! My Score: 10 Point(s) Correct answer : True

36 The XOR. problem can be solved by using the deep neural network.
@ Trne
0 False
Previous
•Corngratulations, you got iU My Score: 10 Point(s) Correct answer : True
37, One of the filter methods is using the Chi-Squared test to analyze the mapping between features and labels to implement feature selection.
@ True

0 False
Previous

0Congratulations, you got m My Score: 10 Point(s) Correct answer: True

38, The dropout echnology can be used only during Al model training.
0 Triue
@ False
Previous
•Congra ulations, you got iU My Score: 1O Point(s) Cmrect answer : False
39 , Filter methods for feature selection are restricted they do not tend to select redundant variables because they consider the relationships between features.

0 True
@ False

I Pre¥ious I

O Congratulations, you got it! My Score: 10 Point(s) Correct answer : False

40, All neurons on the same layer in a recurrent neural network share weights (U, V, and W)..
0 True
@ False
Previous
OWrong answer My Score : 0 Point(s) Correct answer : True
41, MindSpore can be quickly deployed on the cloud and edge , improving resource utilization and privacy protection, and helping developers focus on Al apps. Due to

size of the model, MindSpore cannot be deployed on mobile phones.

':' True

@ False

I Previous I Mi§F
OCongralulal10ns, you got rt• My Score: 10 Point(s) Correct answer : False
42, In a convolutional neural network , the size of each kernel is not necessarily the same as that of a pooling layer w indow, but their steps should be the same.

8 True
@ False
      Previous       *I§*
O Congratulations, you got ! My Score: 10 Poinl(s) Correct answer: False
43, MindSpore accelerates model convergence through automatic parallelism and second-order optimization.
@ True
0 False
Previous I -
O Congratulations, you got it! My Score: 1O Point(s) Cmrect answer: True
44 , A decision tree selects a label from features of the training data to work as the node splitt ing standard. Different label selection standards generate different decisi\ tree algorithms.
True

False

Previous I Mi§F
OCongratulations, you got ill My Score: 10 Point(s) Correct answer: False

45... Federated learning can protect user privacy to some extent.

@ True
0 False
e Congratulations, you got iU
Previous
My Score: 1O Point(s) Cmrect answer : True
46, MindSpore supports automatic graph splitting, which splits graphs based on the input and output data dimensions of the operators and integrates data parallelism

tensor parallelism, and model parallelism.

True

False

I Prous I Mi§F
GCongratulations, you got it! My Score: 10 Point(s) Correct answer : False

47, The added cost of L1 regularization is proportional to the absolute value of the weight coefficient.
@ True
0 False
Previous
eCongratulations, you got it! My Score: 10 Point(s) Correct answer : True
48, Poisson distribution is generally regarded as the common distribution of errors in linear regression.
0 True
@ False
Previous
8 Congratu lations, you got it! My Score: 10 Point(s) Correct answer: False
49, Although each industry and application Faces unique challenges , they are all gradually adopt ing Al-based workflows and solutions .
@ True

0 False
I Previous I -
8Congratulations, you got it! My Score: 10 Point(s) Correct answer : True
50, Boosting independently builds multiple basic learners and then averages their p1ed ictio111s.
@ Tme
0 False
Previous
OWrong answer My Score : O Point(s) Correct answer : False
51, In actual data processing, there are multiple trained models available so that you can choose the opt imal one.
@ True
0 False
Previous I -
O Congratulations, you got it! My Score: 1O Point(s) Correct answer: True
52, As the cornerstone of the Huawei full-stack all-scenario Al solution, Atlas provides modules, cards, and servers powered by Kunpeng Al processors to meet derr

for computing power in all scenarios.

) True

@ False

I Previous I Mi§F
O Congratulations , you got rt! My Score: 10 Po1nt(s) Correct answer : False

53., TensorFlow .2 .x is oompatib le with all l7ensorFlow 1..x modules ..

@ Trne
0 False
Previous
OWrong answer My Score : 0 Point(s) Coned a11swer : False
54, Object Storage Service (OBS) can be widely used in scenarios with abundant relational data such as social apps, enterprise relationsh ip analysis, knowledge gra
risk control, and recommendations.

@ True

0 False
I Previous I Mi§F
0Wrong answer My Score 0 Point(s) Correct answer : False
55, The average value of tensors generated using tf.random.normal([4], 2 , 1, tf.float32) is (). (Fill in the blank w it h a number.)
2
Correct answer : 2

Previous I -
O Congratulations, you got it! My Score: 20 Point(s)

56 Tensor [[[2J3]]] is a/an ()-dimensional tensor. (Fill in the blank with a number.)
1 thrne
Corrnct answer : 3
Previous
8Wrona answer Mv Score : 0 Point(s)
57 , mindspore .ops.GradOperation is the first-order derivative method in MindSpore to compute gradients. It contains multiple parameters. When get_all is set to (), th1

derivatives of all parameters are computed.

1 True

Correct answer· True

I ,Pre,ious
58 Mi§F
I built an open smart ecosystem based on the (), device, chip&quot; three-layered open architecture. It enables developers to quickly utilize Huawei'sl
HUAWE I HiAI

powerful A l processing capabilities to provide a better smart application experience for users.

1 cloud

Correct answer : Cloud,cloud;CLOUD

I Previous I *@*
59, Assume that there are 10,000 data pieces in a cancer data sample, of which 100 pieces are from cancer patients, and the other 9900 pieces are normal. If a

classification model predicts that 9000 out of the 9900 pieces are normal, and 90 out of the 100 pieces are from cancer patients, the accuracy rate of the model is

false

Garred answer: 90 9

       Pre11ious       

*i§
O Wrong answer My Score : O Point(s)
60, In TensorFlow 2.0, if tf .keras.layers. RNN is used to process timing information and you want to obtain the output status at each moment, set () to True.

return_sequence

Corred answer : return_sequence

Pre1.1ious
O Congratulations, you got it! My Score: 20 Point(s)
e data quality?

in TensorFlow 2?

DRAM, reducing the latency_

y online inference services

ated_sample_ 1 =

e output for this code?

atform service?
gression problems _

HUAWEI CLOUD OCR?

n object is a tensor?

ar ial process.

utput of the feature attribute

art of data conversion.

e model are the same

ches the minimum value.

al network.

er : True

lationships between features.

V, and W)..
-order optimization.

te different decisi\ tree algorithms. I

wer : True

ar regression.

orkflows and solutions .

p1ed ictio111s.

e opt imal one.

a number.)