You are on page 1of 4

1.

After the image label service is called, only one corresponding label value is returned: false
2. Logical reasoning is a part of AI: true
3. 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 two models can resolve the XOR problem: true
4. Automatic graph segmentation is one of the key technologies of MindSpore: true
5. Da Vinci architecture is the core of the Ascend AI computing engine and AI processor: true
6. Skit-learn is a free software machine learning library for python. It has powerful tools for data
analysis and mining: true
7. Backpropagation makes gradient descent more efficient by means of dynamic programming:
true
8. Pytorch supports GPU acceleration: true
9. In theory, a model composed of machine learning algorithms cannot represent but only
approximates to a real data distribution function: true
10. In a feedforward network, neurons in the same layer are connected to each other: false
11. Atlas series chips need the support of CUDA software package like GPU accelerator cards: false
12. In a medical image recognition task, we usually ae more concerned about the recall ratio than
the precision ratio when evaluating model performance: true
13. TensorFlow2.0 uses Keras and eager execution to easily model: true
14. IF the metrics parameter is not specified in model. Compile, no loss value will be output: false
15. The HiAI mobile computing platform architecture supports various mainstream front-end
frameworks, such as Tensorflow and Caffe: true
16. Softmax can’t be used for multiple classification problems: false
17. Voice recognition refers to the recognition of text data as audio data: false
18. In some image and voice fields, AI performs better than humans, so artificial intelligence has
entered the stage of strong artificial intelligence: false
19. The adam optimizer in tensorflow 2.0 is the built-in optimizer of the keras module: true
20. One of the disadvantages of deep learning is that the interpretability is very weak, but it is still
possible to visualize the model using methods such as local interpretable Model-Agnostic
Explanations (LIME): true
21. Which of the following is the representative of behaviorism: D reinforcement learning
22. Which of the following reasons is the main reason why deep learning developed rapidly after
2010: D Improvement of computing power
23. Which of the following networks is not a recurrent neural network: B Highway Network
24. Which of the following options is the architecture adopted by Huawei Atlas series chips: D
Ascend
25. What kind of operation does “concat” represent in Tensorflow 2.0: A Merge tensor
26. Which of the following statements about hyperparameters is incorrect: D in K-means, the initial
value of the clustering center is not a hyperparameter, and thus the training result is not
affected.
27. 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: A 3x3
28. Which of the following is the output of the ReLU function: A [0,+infini)
29. Which of the following is not a feature o the Atlas 500 AI edge stations: C High energy
consumption
30. When was Turning test proposed: A 1950
31. Which of the following learning methods is most suitable for training ML models of chessboard
games: D reinforcement learning
32. In the process of training the neural network, Which of the following methods is used to update
the parameters so as to minimize loss function: C Gradient descent
33. Import numpy as np np = np.arrange(24).reshape(2,-1,2,2) Which of the following is the
result of n.shape after running the preceding code: A (2,3,2,2)
34. Input_shape = (4,28,28,1) x=….. What is the output of the preceding code: B (4,26,26,2)
35. What operation does “isinstance” represent in Tensorflow 2.0: C Determine whether it is
tensorflow
36. Which of the following problems can’t be handled by the perceptron: C XOR
37. Which of the following network layers does not need to be included ina concolutional neural
network: C pooling layer
38. Which of the following is not Huawei’s full_stack AI solutions: D Spark
39. Which of the following options is the artificial intelligence development framework launched by
Huawei: C MindSpore
40. Which of the following is a function of HiAI foundation: B Quiclkly convert and migrate existing
models
41. If a model has a small bias and a large variance on the test set, what is the problem with the
model: B Overfitting
42. In Tensorflow, which order of tensors are scalars: A Rank-0
43. Factory A produces 60% …. What is the possibility that the vehicle is produced by factory A: B
1/3
44. Adadelt = tf.keras.optimizers.Adadelta() …. : A adam
45. Which of the following are AI applicaions ABCD Smart speaker, Smart customer device,
Autonomous driving, AlphaGo
46. What are the elements of AI : ABCD algorithms, computing power, data, scenarios
47. Compared with the batch gradient descent method, which of the following options are the
advantages of the mini batch descent method: AB Fast calculation speed in one time parameters
update, Skip local minimum
48. Suppose the following: now we need to classify the data,….: BC Logistic regression, Supposrt
vector machine
49. When using Huawei cloud services, which of the following information will be used for
authentification: AB Current time, AK
50. Which of the following are the problems you may encounter when training a generative
adversarial network: ABCD
51. Which of the following are major modules of the Ascend AI processor: ABCD chip
52. Which of the following are the major functions of mind Expression in MindSpore: ABCD
53. Which of the following are common Python computer vision libraries: CD Matplotlib, cv2
54. Which of the following options are hyperparameters: ABC K in KNN,K in k-means, C in SVM
55. Which of the following options belong to the activation function of the neural network: ABCD
56. Which of the following statements about LSTM are corrcect: BCD the input of time…, if peephole
is used…,in LSTM if the dimension….
57. Which of the following options belong to the application scenarios of natural language
processing: BD Machine translation, Public opinion analysis
58. Which of the following statements about tenserflow are correct: ABC Keras is a high-level API of
tensorflow, Tensorflow 1.0 uses statistic graphs, Tensorflow 2.0 uses Eager Executio by default
59. Which of the following tensor operations are supported by Tensorflow: ABCD
60. Which of the following env are available for creating notebooks on modelArts: ABD Python3,
Python2, TF-2.2.0&pytorch-1.4.0-python3.6
61. The speed of grid search must be faster than random search: false
62. HiLens can help small and micro companies realize smart access control: true
63. Polynomial regression cannot resolve nonlinear problem: false
64. To install TensorFlow of the GPU version, you need to install CUDA and cuDNN: true
65. In a feedforward network, neurons in the same layer ae connected to each other: false
66. Similar to a 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: false
67. Cifar10 is a built-in data set of keras: true
68. The generalization error will continue to decrease as the complexity of the model increases:
false
69. MindSpore supports GPU acceleration: true
70. In the KNN model, a small K value may cause model underfitting: false
71. Malformed data is not dirty data: false
72. Which of the following statements about dropout is incorrect: D dropout can’t be added during
inference
73. Which of the following methods cannot resolve overfitting in a decision tree: D Use the
integration algorithm to set up an AdaBoost model
74. What determines the number of neurons in the input layer of the neural network: A Data
dimension
75. Which of the following schools does the expert system belong to: A symbolism
76. Which of the following options is not the feature of ModelArts: C Support automatic tuning
77. In machine learning, how many support vectors can the SVM algorithm have: D Several
78. Which of the following options is a commonly used loss function for classification problems: B
Cross Entropy loss function
79. Which of the following models is not suitable for the scenario where the covariance between
data features is large: C naïve bayes
80. Which of the following options is Huawei’s artificial intelligence computing platform: D Atlas
81. A company wants to use the HiAI platform to develop its own APP. Which of the following IDEs
can be used in conjunction with the HiAI platform: B Android Studio
82. Which of the following options is the classification of the AI chip business from the application :
C training and inference
83. Which of the following options are the main advantages of the Keras interface in TensorFlow
2.0: ABC convenient for users, modular and combinable, easy to expand
84. Which of the following may cause oferfitting: BCD the model capacity is large, the training set
includes a large amount of dirty data, the number of training sets is small
85. Which of the following options belong to Huawei HiAI platform content: ABD HiAI service, HiAI
Foundation, HiAI Engine
86. Which of the following options are Huawei products: ABC Atlas 200DK, Ascend 310, Ascend 910
87. Compared with the sigmoid activation function,which of the following are the advantages of
ReLU: AD mitigating the problem of vanishing gradient, save calculation
88. Which of the following are deep learning frameworks: ABCD Tensorflow, pytorch, MXNet,
MindSpore

You might also like