wayground logo

Free Printable Worksheets

Font size

S
M
L
XL
Worksheets

Huawei HCIA-AI V3.5 Demo

Total questions: 125

Worksheet time: 4hrs 10mins

Name
Class
Date
1.
Name:
4 lines
2.
1. What are the levels of driving automation defined by SAE International, formerly named the Society of Automotive Engineers (SAE), based on the degree of dependency on the system?
a)
A. L0~L4
b)
B. L0~L5
c)
C. L1~L5
d)
D. L1~L4
3.
2. Which of the following technologies is commonly used for image feature extraction and related research?
a)
A. Convolution neural network
b)
B. Naive Bayes classification algorithm
c)
C. Long short-term memory (LSTM) network
d)
D. Word2Vec
4.
3. Data is carrier and representation of information. Which of the following statements about data in AI applications is true?
a)
A. Data quality is not relevant to the quality of an AI model
b)
B. Data quality is important and as it determines the result of the model
c)
C. Data can be directly imported to a model without preprocessing
d)
D. Data refers only to that within an Excel file
5.
4. Consider a scenario where a machine learning algorithm is used to filter spam. According to the definition of machine learning, which of the following describes the experience E?
a)
A. Spam filtering
b)
B. Accuracy of spam filtering
c)
C. All tagged spam and genuine emails in the past three years
d)
D. Email addresses
6.
5. A computer uses labeled images to learn and determine which images contain apples and which contain pears. Which of the following types of machine learning is most applicable to this scenario?
a)
A. Supervised learning
b)
B. Unsupervised learning
c)
C. Semi-supervised learning
d)
D. Reinforcement learning
7.
6. Which of the following statements is true about classification models and regression models in machine learning?
a)
A. For regression problems, the output variables are discrete values. For classification problems, the output variables are continuous values.
b)
B. The most commonly used indicators for evaluating regression and classification problems are accuracy and recall rate.
c)
C. There may be overfitting in both regression and classification problems
d)
D. Logistic regression is a typical regression model
8.
7. Which of the following points constitute a support vector of the SVM algorithm without considering regularization terms?
a)
A. Points on the separating hyperplane
b)
B. Farthest points from the separating hyperplane
c)
C. Points closest to the separating hyperplane
d)
D. Points of a certain type
9.
8. Kernel functions allow algorithms to fit the largest hyperplane in a transformed high-dimensional feature space. Which of the following is not a common kernel function?
a)
A. Linear kernel function
b)
B. Polynomial kernel function
c)
C. Gaussian kernel function
d)
D. Poisson kernel function
10.
9. In 1958, Frank Rosenblatt invented the perceptron algorithm. 10 years later, Marvin Minsky questioned the perceptron's ability to solve non-linear classification problems, and proposed the question that signed the death warrant of perceptrons. Which of the following is the question?
a)
A. AND problem
b)
B. OR problem
c)
C. XOR problem
d)
D. XAND problem
11.
10. During neural network training, which of the following values is continuously updated by using the gradient descent method to minimize the loss function?
a)
A. Hyperparameter
b)
B. Feature
c)
C. Number of samples
d)
D. Parameter
12.
11. The ReLU function is commonly used in deep learning neural networks. Which of the following is the value range of this function?
a)
A. [0,+∞)
b)
B. [0,1]
c)
C. [-1,1]
d)
D. [-1,0]
13.
12. The sigmoid activation function is monotonic and continuous, has bounded outputs, and makes the network easy to converge. It was popular in a period of time. However, when the network is deep, what problems may sigmoid cause?
a)
A. Gradient reduction
b)
B. Vanishing gradient
c)
C. XOR
d)
D. Overfitting
14.
13. Which of the following are false about convolutional neural networks?
a)
A. A convolutional neural network may contain convolutional, pooling, and fully connected layers.
b)
B. Convolution kernels cannot extract global features of images.
c)
C. Common pooling includes max pooling and average pooling.
d)
D. When an image is processed, convolution is implemented by using a scanning window.
15.
14. Overfitting problems can be avoided through dataset expansion. Which of the following statements is true about dataset expansion?
a)
A. The larger the dataset, the lower the probability of overfitting.
b)
B. The larger the dataset, the higher the probability of overfitting.
c)
C. The smaller the dataset, the lower the probability of overfitting.
d)
D. The probability of overfitting decreases when the dataset increases or decreases.
16.
15. Which of the following is the shape of tensor [[[0,1],[2,3]],[[4,5],[6,7]]]?
a)
A. [3,3,2]
b)
B. [3,2,4]
c)
C. [2,3,4]
d)
D. [2,2,2]
17.
16. Which of the following statements about the running process of the MindArmour subsystem is false?
a)
A. Configuration policies: Define test policies based on threat vectors and trustworthiness certification requirements and select appropriate test data generation methods.
b)
B. Fuzzing execution: Generate trusted test data randomly based on the model coverage and configuration policies.
c)
C. Evaluation report: Generate an evaluation report based on built-in or user-defined trustworthiness metrics.
d)
D. Trustworthiness enhancement: Use preset methods to enhance the trustworthiness of AI models.
18.
17. Which of the following is NOT a complexity feature of AI computing?
a)
A. Mixed precision computing
b)
B. Parallel data and computing
c)
C. Parallel communication and computing
d)
D. Parallel processing of structured and unstructured data
19.
18. On-device execution refers to the execution of the entire graph. It makes full use of the computing power of the Ascend AI Processor, greatly reducing the interaction overhead and improving the accelerator usage. Which of the following is false about on-device execution?
a)
A. Challenges to model execution with powerful chip computing power: Memory wall problems, high interaction overhead, and difficult data supply. Some operations are performed on the host, while others are performed on the device. The interaction overhead is much greater than the execution overhead. As a result, the accelerator usage is low.
b)
B. The chip-oriented deep graph optimization is used to reduce synchronization waiting time and maximize the parallelism degree of "data-computing-communication". The training performance is equivalent to that of the graph scheduling mode on the host.
c)
C. Challenges to distributed gradient aggregation with powerful chip computing power:When a single iteration of ResNet-50 takes 20 ms, the central control overhead and communication overhead for frequent synchronization are generated. The traditional method requires three times of synchronization to complete AllReduce. The data-driven method automatically performs AllReduce without control overhead.
d)
D. MindSpore uses gradient-driven adaptive graph optimization to implement decentralized and autonomous All Reduce. The gradient aggregation step is consistent, and computing and communication are fully streamlined.
20.
19. Which of the following statements about the Da Vinci architecture is incorrect?
a)
A. A compute unit contains four types of basic compute resources.
b)
B. Control units are responsible for the running of AI Cores.
c)
C. The storage system consists of on-chip storage units of AI Core and corresponding data paths.
d)
D. Transfers data to the L1 buffer through the bus interface unit.
21.
20. Which of the following is not a component of the AI Core storage unit in the Da Vinci architecture?
a)
A. Storage control unit
b)
B. Buffer
c)
C. I-Cache
d)
D. Register
22.
21. Which of the following statements about the L1 buffer is true?
a)
A. Data in the L1 buffer needs to be read outside the AI Core over the bus interface each time.
b)
B. The L1 buffer can permanently retrain data that needs to be reused.
c)
C. The L1 buffer decreases data accesses over the bus and avoids bus congestion.
d)
D. L1 buffer is used to store the initial values in the neural network.
23.
22. In Huawei Cloud EI, which of the following is a one-stop AI development platform that supports large-volume data preprocessing, semi-automated data labeling, distributed training, automated model building, and on-demand model deployment across the device, edge, and cloud for machine learning and deep learning; and helps AI developers quickly build and deploy models and efficiently manage the AI development lifecycle?
a)
A. ModelArts
b)
B. MindSpore
c)
C. MySQL
d)
D. Ascend
24.
23. Which of the following statements about the three major schools of AI are true?
a)
A. Symbolicism puts more focus on logical inference. Connectionism is relatively weak in terms of being logical and interpretability. Instead, it focuses on results.
b)
B. Connectionism stems from the study of bionics, especially human brain models.
c)
C. Actionism puts more focus on practice. It believes in constantly adapting actions by learning from the environment.
d)
D. Symbolicism puts more focus on practice. It believes in constantly adapting actions by learning from the environment.
25.
24. Which of the following are topics of speech processing research?
a)
A. Speech recognition
b)
B. Voice processing
c)
C. Speech wake-up
d)
D. Voiceprint recognition
26.
25. Which of the following statements about some of AI use case are true?
a)
A. Smart home uses technologies such as the Internet of Things (IoT) and speech recognition.
b)
B. Smart healthcare uses AI technologies such as computer vision and data mining.
c)
C. Smart city is a comprehensive domain, covering all three basic technology fields of AI: speech, NLP, and computer vision.
d)
D. Smart education features openness, sharing, interaction, collaboration, and ubiquity. It aims to modernize education and upgrade traditional teaching methods with information technologies.
27.
26. Which of the following statements about datasets are true?
a)
A. A dataset is a collection of data used in machine learning tasks. Each piece of data is called a sample.
b)
B. Events or attributes that reflect the performance or nature of a sample in a particular aspect are called features.
c)
C. A training set is a dataset used in the training process, where each sample is referred to as a training sample.
d)
D. Learning (or training) is the process of building a model from data.
28.
27. Which of the following statements about data preprocessing are true?
a)
A. Data cleansing is a process of filling in missing values, as well as detecting and eliminating noise data and exceptions.
b)
B. The purpose of data dimension reduction is to simplify data attributes and avoid the curse of dimensionality.
c)
C. The purpose of data standardization is to reduce noise data and improve model accuracy by standardizing data.
d)
D. Machine learning outputs results through models. Therefore, model training is more important than data preprocessing.
29.
28. Which of the following statements are true about model parameters and hyperparameters?
a)
A. Models contain both parameters and hyperparameters.
b)
B. Hyperparameters are automatically learned by models.
c)
C. Hyperparameters are manually set.
d)
D. Hyperparameters can be used to control training.
30.
29. Which of the following are common kernel functions used by support vector machines?
a)
A. Gaussian kernel function
b)
B. Sigmoid kernel function
c)
C. Linear kernel function
d)
D. Polynomial kernel function
31.
30. Which of the following statements are true about the advantages of the Rectified Linear Unit (ReLU) activation function?
a)
A. The output is bounded, meaning the training is not easy to diverge.
b)
B. The calculation is simple.
c)
C. The vanishing gradient problem can be effectively alleviated.
d)
D. Dead neurons exist.
32.
31. Which of the following are common activation functions of neural networks?
a)
A. Dropout
b)
B. Sigmoid
c)
C. Tanh
d)
D. Leaky ReLU
33.
32. Which of the following are activation functions of deep learning algorithms?
a)
A. Sigmoid
b)
B. ReLU
c)
C. Tanh
d)
D. Sin
34.
33. Which of the following comprise the Adam optimizer?
a)
A. Momentum
b)
B. Adagrad
c)
C. RMSProp
d)
D. Nesterov
35.
34. Which of the following are characteristics of eager execution in TensorFlow 2.x?
a)
A. High performance
b)
B. Deployable
c)
C. Visualized
d)
D. Flexible
36.
35. Which of the following are activation function operators in MindSpore?
a)
A. Softmax
b)
B. MaxPool
c)
C. Sigmoid
d)
D. ReLU
37.
36. Which of the following frameworks innately support the distributed deep learning framework?
a)
A. TensorFlow
b)
B. MindSpore
c)
C. CNDK
d)
D. MXNet
38.
37. Which of the following are common evaluation metrics for object detection?
a)
A. mAP
b)
B. IOU value
c)
C. ROC
d)
D. BLEU
39.
38. Which of the following network model formats can be used to save training parameters and network models?
a)
A. Checkpoint
b)
B. MindIR
c)
C. ONNX
d)
D. AIR
40.
39. In a neural network based on connectionism, each node can express a specific meaning.
a)
True
b)
False
41.
40. As the cornerstone of Huawei's full-stack, all-scenario AI solution, Atlas provides modules, boards, and servers powered by the Ascend AI processor to meet customer demand for computing power in all scenarios.
a)
True
b)
False
42.
41. ModelArts is a one-stop development platform for AI developers. With large-volume data preprocessing, semi-automated data labeling, distributed training, automated model building, and on-demand model deployment across the device, edge, and cloud, ModelArts helps AI developers quickly build and deploy models and efficiently manage the AI development lifecycle.
a)
True
b)
False
43.
42. In terms of ensuring data privacy and security, the federated learning technology uses different data sources to train models, overcoming the data bottleneck as a result.
a)
True
b)
False
44.
43. Softmax regression is a generalization of logistic regression and applies only to binary classification.
a)
True
b)
False
45.
44. Lasso regression is a type of linear regression in which an absolute loss compensation term is added to the loss function.
a)
True
b)
False
46.
45. Assuming a dataset contains the areas and prices of 21,613 housing units in a city, you can use a classification model to predict the prices of other housing units in the city.
a)
True
b)
False
47.
46. The tanh function can effectively solve the vanishing gradient problem.
a)
True
b)
False
48.
47. Dropout can only be used in neural networks to avoid overfitting.
a)
True
b)
False
49.
48. L1 regularization is referred to as weight decay.
a)
True
b)
False
50.
49. When mindspore.nn.MaxPool2d is used to build a 2D maximum pooling layer, the received data format is [Height, Width, Channel, Quantity] by default.
a)
True
b)
False
51.
50. The vision and value of MindSpore is to lower the AI development threshold, and unleash the computing power of Ascend AI Processors, and empower inclusive AI.
a)
True
b)
False
52.
51. The MindInsight subsystem of MindSpore discovers model lineage and compares training results through the collected information about training hyperparameters, datasets, and data augmentation.
a)
True
b)
False
53.
52. MindSpore can be quickly deployed on the cloud, edge, and mobile phone, improving resource utilization and privacy protection and enabling developers to focus on developing AI apps.
a)
True
b)
False
54.
53. Products of Atlas series cover training and inference scenarios, including servers, edge computing devices, and acceleration modules.
a)
True
b)
False
55.
54. Huawei Cloud EI provides intelligent twins, AI development platforms, and general AI capabilities.
a)
True
b)
False
56.
55. The HMS Core supports HarmonyOS and Android, but does not support Windows.
a)
True
b)
False
57.
56. Compared with CPUs and GPUs, NPUs use synapse weights to integrate storage and compute, improving operating efficiency.
a)
True
b)
False
58.

57. The four elements of AI are data, compute power, (a)   , and application scenarios.

59.

58. The primary purpose of (a)   is to remove unimportant weights from the weight matrix and fine-tune the network again.

60.

59. The feature interpretability in deep learning is (a)   than in traditional machine learning. (Enter "stronger" or "weaker".)

61.

60. Typical training data processing pipelines include load, shuffle, map, batch, and (a)  

62.

61. Among the MindSpore's Python APIs, mindspore. (a)   defines loss functions, optimizers, and computing units for constructing networks.

63.

62. In Huawei Cloud NLP service, the (a)   technology can divide text into separate words as a sequence. For example, in English text, spaces are natural delimiters between words.

64.

63. What are the levels of driving automation defined by SAE International, formerly named the Society of Automotive Engineers (SAE), based on the degree of dependency on the system?

a)
A. L0~L4
b)
B. L0~L5
c)
C. L1~L5
d)
D. L1~L4
65.

64."Batch inference is a batch job that performs inference on batch data. There is no need for model training before using batch inference." Which of the following is true about this statement?

a)
A.This statement is correct. With batch inference, training is no longer required.
b)
B.This statement is correct. Inference means the end of training.
c)
C.This statement is incorrect. Model training is required before inference is performed.
d)
D.This statement is incorrect. No training is required before batch inference.
66.

65. Data is carrier and representation of information. Which of the following statements about data in AI applications is true?

a)
A. Data quality is not relevant to the quality of an AI model
b)
B. Data quality is important and as it determines the result of the model
c)
C. Data can be directly imported to a model without preprocessing
d)
D. Data refers only to that within an Excel file
67.

66. A computer uses labeled images to learn and determine which images contain apples and which contain pears. Which of the following types of machine learning is most applicable to this scenario?

a)
A. Supervised learning
b)
B. Unsupervised learning
c)
C. Semi-supervised learning
d)
D. Reinforcement learning
68.

67. Which of the following statements is true about classification models and regression models in machine learning?

a)
A. For regression problems, the output variables are discrete values. For classification problems, the output variables are continuous values.
b)
B. The most commonly used indicators for evaluating regression and classification problems are accuracy and recall rate.
c)
C. There may be overfitting in both regression and classification problems
d)
D. Logistic regression is a typical regression model
69.

68. Which of the following points constitute a support vector of the SVM algorithm without considering regularization terms?

a)
A. Points on the separating hyperplane
b)
B. Farthest points from the separating hyperplane
c)
C. Points closest to the separating hyperplane
d)
D. Points of a certain type
70.

69. Which of the following statements is false about support vector machines (SVMs)?

a)
A. SVMs are classification models. Their basic model is the linear classifier that maximizes the width of the gap between the two categories in the feature space.
b)
B. SVMs also have a kernel trick, which makes them non-linear classifiers.
c)
C. In the case of linear inseparability, non-linear mapping algorithms are used to convert the linearly inseparable samples of low-dimensional input space into samples of high-dimensional feature space. In this way, samples become linearly separable.
d)
D. SVMs only apply to linear classification.
71.

70. Kernel functions allow algorithms to fit the largest hyperplane in a transformed high-dimensional feature space. Which of the following is not a common kernel function?

a)
A. Linear kernel function
b)
B. Polynomial kernel function
c)
C. Gaussian kernel function
d)
D. Poisson kernel function
72.

71. In 1958, Frank Rosenblatt invented the perceptron algorithm. 10 years later, Marvin Minsky questioned the perceptron's ability to solve non-linear classification problems, and proposed the question that signed the death warrant of perceptrons. Which of the following is the question?

a)
A. AND problem
b)
B. OR problem
c)
C. XOR problem
d)
D. XAND problem
73.

72. During neural network training, which of the following values is continuously updated by using the gradient descent method to minimize the loss function?

a)
A. Hyperparameter
b)
B. Feature
c)
C. Number of samples
d)
D. Parameter
74.

73. The ReLU function is commonly used in deep learning neural networks. Which of the following is the value range of this function?

a)
A. [0,+∞)
b)
B. [0,1]
c)
C. [-1,1]
d)
D. [-1,0]
75.

74. The sigmoid activation function is monotonic and continuous, has bounded outputs, and makes the network easy to converge. It was popular in a period of time. However, when the network is deep, what problems may sigmoid cause?

a)
A. Gradient reduction
b)
B. Vanishing gradient
c)
C. XOR
d)
D. Overfitting
76.

75. Which of the following are false about convolutional neural networks?

a)
A. A convolutional neural network may contain convolutional, pooling, and fully connected layers.
b)
B. Convolution kernels cannot extract global features of images.
c)
C. Common pooling includes max pooling and average pooling.
d)
D. When an image is processed, convolution is implemented by using a scanning window.
77.

76. Overfitting problems can be avoided through dataset expansion. Which of the following statements is true about dataset expansion?

a)
A. The larger the dataset, the lower the probability of overfitting.
b)
B. The larger the dataset, the higher the probability of overfitting.
c)
C. The smaller the dataset, the lower the probability of overfitting.
d)
D. The probability of overfitting decreases when the dataset increases or decreases.
78.

77. Which of the following functions can be used to alleviate the vanishing gradient problem?

a)
A. Sigmoid
b)
B. Tanh
c)
C. Softsign
d)
D. ReLU
79.

78. Which of the following is the shape of tensor [[[0,1],[2,3]],[[4,5],[6,7]]]?

a)
A. [3,3,2]
b)
B. [3,2,4]
c)
C. [2,3,4]
d)
D. [2,2,2]
80.

79. Which of the following statements about the running process of the MindArmour subsystem is false?

a)
A. Configuration policies: Define test policies based on threat vectors and trustworthiness certification requirements and select appropriate test data generation methods.
b)
B. Fuzzing execution: Generate trusted test data randomly based on the model coverage and configuration policies.
c)
C. Evaluation report: Generate an evaluation report based on built-in or user-defined trustworthiness metrics.
d)
D. Trustworthiness enhancement: Use preset methods to enhance the trustworthiness of AI models.
81.

80. Which of the following is NOT a complexity feature of AI computing?

a)
A. Mixed precision computing
b)
B. Parallel data and computing
c)
C. Parallel communication and computing
d)
D. Parallel processing of structured and unstructured data
82.

81. On-device execution refers to the execution of the entire graph. It makes full use of the computing power of the Ascend AI Processor, greatly reducing the interaction overhead and improving the accelerator usage. Which of the following is false about on-device execution?

a)
A. Challenges to model execution with powerful chip computing power: Memory wall problems, high interaction overhead, and difficult data supply. Some operations are performed on the host, while others are performed on the device. The interaction overhead is much greater than the execution overhead. As a result, the accelerator usage is low.
b)
B. The chip-oriented deep graph optimization is used to reduce synchronization waiting time and maximize the parallelism degree of "data-computing-communication". The training performance is equivalent to that of the graph scheduling mode on the host.
c)
C. Challenges to distributed gradient aggregation with powerful chip computing power:When a single iteration of ResNet-50 takes 20 ms, the central control overhead and communication overhead for frequent synchronization are generated. The traditional method requires three times of synchronization to complete AllReduce. The data-driven method automatically performs AllReduce without control overhead.
d)
D. MindSpore uses gradient-driven adaptive graph optimization to implement decentralized and autonomous All Reduce. The gradient aggregation step is consistent, and computing and communication are fully streamlined.
83.

82. Which of the following statements about the Da Vinci architecture is incorrect?

a)
A. A compute unit contains four types of basic compute resources.
b)
B. Control units are responsible for the running of AI Cores.
c)
C. The storage system consists of on-chip storage units of AI Core and corresponding data paths.
d)
D. Transfers data to the L1 buffer through the bus interface unit.
84.

83. Which of the following statements about the L1 buffer is true?

a)
A. Data in the L1 buffer needs to be read outside the AI Core over the bus interface each time.
b)
B. The L1 buffer can permanently retrain data that needs to be reused.
c)
C. The L1 buffer decreases data accesses over the bus and avoids bus congestion.
d)
D. L1 buffer is used to store the initial values in the neural network.
85.

84. In Huawei Cloud EI, which of the following is a one-stop AI development platform that supports large-volume data preprocessing, semi-automated data labeling, distributed training, automated model building, and on-demand model deployment across the device, edge, and cloud for machine learning and deep learning; and helps AI developers quickly build and deploy models and efficiently manage the AI development lifecycle?

a)
A. ModelArts
b)
B. MindSpore
c)
C. MySQL
d)
D. Ascend
86.

85. Which of the following statements about some of the subfields of AI are true?

a)
A. Computer vision is a science that studies how to make computers "see" things.
b)
B. Speech processing is a general term for different speech processing technologies, such as the study of vocalization, the collection of statistics related to speech signals, speech recognition, machine synthesis, and speech perception.
c)
C. Natural language processing (NLP) studies how to use computer technology to understand and use human languages.
d)
D. Autonomous driving does not require speech processing or computer vision.
87.

86. Which of the following are topics of speech processing research?

a)
A. Speech recognition
b)
B. Voice processing
c)
C. Speech wake-up
d)
D. Voiceprint recognition
88.

87. Which of the following statements about some of AI use case are true?

a)
A. Smart home uses technologies such as the Internet of Things (IoT) and speech recognition.
b)
B. Smart healthcare uses AI technologies such as computer vision and data mining.
c)
C. Smart city is a comprehensive domain, covering all three basic technology fields of AI: speech, NLP, and computer vision.
d)
D. Smart education features openness, sharing, interaction, collaboration, and ubiquity. It aims to modernize education and upgrade traditional teaching methods with information technologies.
89.

88. Which of the following statements about datasets are true?

a)
A. A dataset is a collection of data used in machine learning tasks. Each piece of data is called a sample.
b)
B. Events or attributes that reflect the performance or nature of a sample in a particular aspect are called features.
c)
C. A training set is a dataset used in the training process, where each sample is referred to as a training sample.
d)
D. Learning (or training) is the process of building a model from data.
90.

89. Which of the following statements about data preprocessing are true?

a)
A. Data cleansing is a process of filling in missing values, as well as detecting and eliminating noise data and exceptions.
b)
B. The purpose of data dimension reduction is to simplify data attributes and avoid the curse of dimensionality.
c)
C. The purpose of data standardization is to reduce noise data and improve model accuracy by standardizing data.
d)
D. Machine learning outputs results through models. Therefore, model training is more important than data preprocessing.
91.

90. Which of the following statements are true about model parameters and hyperparameters?

a)
A. Models contain both parameters and hyperparameters.
b)
B. Hyperparameters are automatically learned by models.
c)
C. Hyperparameters are manually set.
d)
D. Hyperparameters can be used to control training.
92.

91. Which of the following are common kernel functions used by support vector machines?

a)
A. Gaussian kernel function
b)
B. Sigmoid kernel function
c)
C. Linear kernel function
d)
D. Polynomial kernel function
93.

92. Which of the following statements are true about the advantages of the Rectified Linear Unit (ReLU) activation function?

a)
A. The output is bounded, meaning the training is not easy to diverge.
b)
B. The calculation is simple.
c)
C. The vanishing gradient problem can be effectively alleviated.
d)
D. Dead neurons exist.
94.

93. Which of the following are common activation functions of neural networks?

a)
A. Dropout
b)
B. Sigmoid
c)
C. Tanh
d)
D. Leaky ReLU
95.

94. Which of the following are activation functions of deep learning algorithms?

a)
A. Sigmoid
b)
B. ReLU
c)
C. Tanh
d)
D. Sin
96.

95. Which of the following statements are true about commonly used optimizers?

a)
A. The momentum optimizer updates parameters with the same learning rate, while the momentum coefficient is adjusted with each iteration.
b)
B. The idea behind the Adagrad optimizer is to set different learning rates for different parameters.
c)
C. One drawback of the Adagrad optimizer is that it ends the optimization process too early.
d)
D. The RMSProp optimizer introduces an attenuation coefficient to enable the gradients to attenuate by a certain proportion in each round.
97.

96. Which of the following comprise the Adam optimizer?

a)
A. Momentum
b)
B. Adagrad
c)
C. RMSProp
d)
D. Nesterov
98.

97. Which of the following are characteristics of eager execution in TensorFlow 2.x?

a)
A. High performance
b)
B. Deployable
c)
C. Visualized
d)
D. Flexible
99.

98. Which of the following are activation function operators in MindSpore?

a)
A. Softmax
b)
B. MaxPool
c)
C. Sigmoid
d)
D. ReLU
100.

99. Which of the following network model formats can be used to save training parameters and network models?

a)
A. Checkpoint
b)
B. MindIR
c)
C. ONNX
d)
D. AIR
101.

100. Which of the following hardware supports MindSpore training?

a)
A. CPU
b)
B. GPU
c)
C. NPU
d)
D. TPU
102.

101. In a neural network based on connectionism, each node can express a specific meaning.

a)
True
b)
False
103.

102. As the cornerstone of Huawei's full-stack, all-scenario AI solution, Atlas provides modules, boards, and servers powered by the Ascend AI processor to meet customer demand for computing power in all scenarios.

a)
True
b)
False
104.

103. ModelArts is a one-stop development platform for AI developers. With large-volume data preprocessing, semi-automated data labeling, distributed training, automated model building, and on-demand model deployment across the device, edge, and cloud, ModelArts helps AI developers quickly build and deploy models and efficiently manage the AI development lifecycle.

a)
True
b)
False
105.

104. In terms of ensuring data privacy and security, the federated learning technology uses different data sources to train models, overcoming the data bottleneck as a result.

a)
True
b)
False
106.

105. Softmax regression is a generalization of logistic regression and applies only to binary classification.

a)
True
b)
False
107.

106. Lasso regression is a type of linear regression in which an absolute loss compensation term is added to the loss function.

a)
True
b)
False
108.

107. Assuming a dataset contains the areas and prices of 21,613 housing units in a city, you can use a classification model to predict the prices of other housing units in the city.

a)
True
b)
False
109.

108. The tanh function can effectively solve the vanishing gradient problem.

a)
True
b)
False
110.

109. Dropout can only be used in neural networks to avoid overfitting.

a)
True
b)
False
111.

110. L1 regularization is referred to as weight decay.

a)
True
b)
False
112.

111. When mindspore.nn.MaxPool2d is used to build a 2D maximum pooling layer, the received data format is [Height, Width, Channel, Quantity] by default.

a)
True
b)
False
113.

112. The vision and value of MindSpore is to lower the AI development threshold, and unleash the computing power of Ascend AI Processors, and empower inclusive AI.

a)
True
b)
False
114.

113. The MindInsight subsystem of MindSpore discovers model lineage and compares training results through the collected information about training hyperparameters, datasets, and data augmentation.

a)
True
b)
False
115.

114. MindSpore can be quickly deployed on the cloud, edge, and mobile phone, improving resource utilization and privacy protection and enabling developers to focus on developing AI apps.

a)
True
b)
False
116.

115. Products of Atlas series cover training and inference scenarios, including servers, edge computing devices, and acceleration modules.

a)
True
b)
False
117.

116. Huawei Cloud EI provides intelligent twins, AI development platforms, and general AI capabilities.

a)
True
b)
False
118.

117. The HMS Core supports HarmonyOS and Android, but does not support Windows.

a)
True
b)
False
119.

118. Compared with CPUs and GPUs, NPUs use synapse weights to integrate storage and compute, improving operating efficiency.

a)
True
b)
False
120.

119. The primary purpose of (a)   is to remove unimportant weights from the weight matrix and fine-tune the network again.

121.

120. The k value in the k-nearest neighbors (k-NN) algorithm is manually set. It is a/an (a)   of a model. (Enter "parameter" or "hyperparameter".)

122.

121. In the CART algorithm, the classification tree measures the purity mainly through the () coefficient.

(a)  

123.

122. The feature interpretability in deep learning is (a)   than in traditional machine learning. (Enter "stronger" or "weaker".)

124.

123. CNN is an abbreviation for (a)  

125.

124. Among the MindSpore's Python APIs, mindspore. (a)   defines loss functions, optimizers, and computing units for constructing networks.

Similar Resources on Wayground

Mega Test

122 questions

Mega Test

quiz

Worksheet

•

9th - 12th Grade

All Clicks - Fall Final - Legal Proceedings

125 questions

All Clicks - Fall Final - Legal Proceedings

quiz

Worksheet

•

11th Grade

ECOEMS 2025 FINAL

128 questions

ECOEMS 2025 FINAL

quiz

Worksheet

•

9th - 12th Grade

mcpc

123 questions

mcpc

quiz

Worksheet

•

KG - Professional Dev...

Hunter Safety Wildlife ID Test Review

130 questions

Hunter Safety Wildlife ID Test Review

quiz

Worksheet

•

9th - 12th Grade

Psychology Final Review

122 questions

Psychology Final Review

quiz

Worksheet

•

KG - University

Τροχός τα Ανθρώπινα - Myth2Space edition

121 questions

Τροχός τα Ανθρώπινα - Myth2Space edition

quiz

Worksheet

•

8th Grade - University

Đề cương ôn tập KTCK1 – Tin học 11 (trắc nghiệm)

120 questions

Đề cương ôn tập KTCK1 – Tin học 11 (trắc nghiệm)

quiz

Worksheet

•

11th Grade