AI Mock Exam. All the Exams. HCIA-AI V1.0 Mock Exam. HUAWEI H13-311-ENU With Passcert. Over 500 Questions and Answers.

Document Content and Description Below

AI Mock Exam Right or Wrong 1. The learning rate is a hyperparameter. 2. Deep learning is one of unsupervised learning. 3. All convolution kernels of the same layer in a convolutional neural... network are weight shared. 4. There are two main types of neural network calculations: forward propagation and back propagation. 5. We have regression functions A and B. If A is simpler than B, A will almost perform better than B on the test set. Single Choice 1. If we want to train decision trees on big data sets with less time, we can A. Reduce the number of trees B. Increase the depth of the tree C. Reduce the depth of the tree D. Increase learning rate (learning rate) 2. Which of the following techniques can you use if you need to change parameters to minimize the cost function? A. Any of the above B. Random search C. Exhaustive search D. Bayesian optimization 3. Which of the following operations achieves a similar effect to Dropout in a neural network? A. Bagging B. None of the above C. Boosting D. Stacking 4. which of the following statements of recurrent neural network is wrong? A. Recurrent neural networks can be expanded according to the time step B. LSTM is also a kind of recurrent neural network C. Recurrent neural network can be abbreviated as RNN D. LSTM cannot solve the problem of gradient disappearance 5. What is the correct step of the gradient descent algorithm?1. Calculate the error between the output value and the true value2. Iterate until the best weight is found3. forward the input into the network and get the output value4. Initialize the weights and biases in network by random5. For each neuron that produces an error, change the corresponding (weight) value to reduce the error. A. 4, 3, 1, 5, 2 B. 5, 4, 3, 2, 1 C. 1, 2, 3, 4, 5 D. 3, 2, 1, 5, 4 6. How to determine the K value in the K-means Clustering algorithm? A. random B. A known C. Automatic computer recognition D. Artificial 7. We usually use mini bacth for gradient descent in DNN training. What does the mini batch mean? A. Use part of the hyperparameters for training at each iteration B. Use part of the data set for training at each iteration C. Use part of the features for training at each iteration D. Use part of the erros for training at each iteration 8. What is the result of the following code execution?[i**i for i in xrange(3)] A. [1, 2,3] B. [0, 1, 4] C. [1, 1, 4] D. （ 1,1,4） 9. The following statements are the accuracy、 recall and F1 value of the classification algorithm. what is wrong? A. The recall is the ratio of the number of related documents retrieved to the number of related documents in the document library, which measure the retrieval rate of the system. B. In order to solve the conflict problem of accuracy and recall, F1 score is introduced. C. The accuracy, recall and F1 score are between 0 and 1. The more close the value is to 0, the better perfoemence of the model is . D. The accuracy rate is the ratio of the number of related documents retrieved to the total number of documents retrieved, which measures how accurate the retrival system is. 10. Which of the following techniques is used to solve overfitting in neural networks? A. Batch normalization B. Regularization C. Dropout D. all of the above 11. What is the type of labels predicted by the regression algorithms? A. Discrete type B. Self-modification C. Continuous type D. Resilience 12. Which version of Python is generally recommended as a production environment? A. Python 2.6.* B. Python 3.* C. Python 2.5.* D. Python 2.7.* 13. Which method can be used for tensor merging in TensorFlow 2.0? A. concat B. unstack C. split D. join 14. For image recognition problems (such as identifying cats in photos), which kind of neural networks is the most suitable one ? A. Multilayer perceptron B. RNN C. perceptron D. Convolutional neural network 15. Which of the following is not a legal identifier for python？ A. 40XL B. self C. int32 D. __name__ 16. What is the result of the following code execution?x=1def change(a):x+=1print xchange(x) A. 2 B. 1 C. Error D. 3 17. How can we solve the high bias problem of the model? A. Add features to the feature space B. Increase data points C. B and C D. Reduce features in the feature space 18. What is the L1 norm of vector x=[1,2,3,4,-9,0] A. 19 B. sqrt(111) C. 1 D. 6 19. How does Python generally perform grammatical boundary indentation? A. Eight spaces B. Two spaces C. TAB D. Four spaces 20. Which of the following statements about supervised learning is correct? A. Decision tree is a kind of supervised learning B. Supervised learning can be trained without labels C. Supervised learning cannot be trained using cross-validation D. Supervised learning is a rule-based algorithm 21. which one of the following indexes is used to select frequent sets in Apriori？ A. Support B. Accuracy C. Confidence D. alpha 22. Which operator in Python can be used to calculate the integer quotient? A. / B. // C. % D. ? 23. which one of the following statements of SVM is wrong? A. The support vector is the vector point closest to the hyperplane B. SVM cannot be used for the classification of nonlinear data sets C. SVM can use multiple kernel functions D. SVM can be used to solve classification problems 24. which operation is uesed in Bagging to reduce the varience? A. optimization B. more data C. average D. Multiple iterations 25. s=’ abcdefg’ , then the value of s[3:5] is A. de B. def C. cd D. cde Multiple Choice 1. In the user-based collaborative filtering algorithm, how to calculate the similarity between two users? A. Based on pearson correlation coefficient B. Based on the Jaccard formula C. Euclidean distance similarity D. Cosine similarity D 2. Common regression algorithms include A. Logistic regression B. Polynomial regression C. Ridge regression D. Linear regression B C D 3. What are the loss functions of GBDT? A. Exponential loss function B. Huber loss function C. Mean square error D. Quantile loss function B C D 4. What are the optimizers for deep learning? A. Adagrad B. Ordinary GD optimizer C. Nesterov D. Momentum optimizer B C D 5. Which of the following are the disadvantages of the gradient descent method? A. Convergence slows down near the minimum. B. It may fall "zigzag". C. Gradient descent can only find local optimal solutions. D. There may be some problems when searching in a straight line. B D 6. Which of the following are not included in the clustering algorithm? A. PCA B. K-means C. Hierarchical D. LLE D 7. Which of the following are necessary conditions for a positive definite matrix? A. The order master sub-form is all greater than 0 B. Is a measure matrix of a base. C. The eigenvalues are all positive. D. There is no reversible matrix. B D 8. What are the three elements of a decision tree? A. Feature selection B. Decision tree generation C. optimization D. Pruning B D 9. Which of the following activation functions are prone to gradient disappearance? A. ReLU B. Tanh C. Sigmoid D. Sogtplus C 10. What are the main methods of content-based recommendation? A. Relevance B. Heuristic approach C. Open D. Model-based approach. D 11. What are the advantages of PCA? A. It is only necessary to measure the amount of information by variance, independent of factors outside the data set. B. The meaning of each feature dimension of the principal component has a certain ambiguity, which is not as strong as the original sample feature. C. Orthogonality between principal components to eliminate the interaction between raw data components D. The calculation method is simple, and the main operation is eigenvalue decomposition, which is easy to implement. C D 12. What operations can an iterator use to traverse? A. while B. do while C. for D. next() D 13. If a = (1, 2, 3), which of the following operations are legal? A. list(a) B. a[2] = 4 C. a[1:-1] D. a*3 C D 14. In which case, the deep learning is not suitable? A. Data has local relevance B. The data set is too small C. The data set has no local correlation. D. The data set category is relatively single. C 15. Which of the following numeric types does Python support? A. long B. float C. int D. complex B C D True or false questions Questions Response (right /wrong) 1. The learning rate is a hyperparameter True 2. Deep learning is one of unsupervised learning True 3. All convolution kernels of the same layer in a convolutional neural network are weight shared True 4. There are two main types of neural network calculations: forward propagation and back propagation True 5. We have regression functions A and B. If A is simpler than B, A will almost perform better than B on the test set. ?? Currently, AI is suitable for scenarios with a known environment, clear targets, and predictable actions. Deep learning deals with image recognition, speech recognition, and translation, with AI having the same or in some cases, better recognition abilities than human beings. These abilities are applied in many scenarios like medical treatment and public safety but are still weak in reasoning and cognition. 6. True AI is a new science of researching and developing theories, methods, and systems to imitate and enhance human intelligence. It is one of the core research fields of machine learning 7. False Python lists can be delimited by square brackets ([ ]), with the first element at index 0. 8. False The for loop in Python can iterate over elements of any list. 9. True 10. Common mathematical functions in Python are stored in the math and cmath modules True Python tuples are enclosed in parentheses (||) and the internal elements are separated by semicolons (;) 11. False Time intervals in Python are floating point numbers in units of seconds 12. True Python identifiers include numbers, digits and underscores 13. True 14. The Python author designed the restrictive syntax to filter out bad programming habits (like failing to add indentation on the right of the next line of the if statement) False 15. Continue statements in Python are used for jumping out the current loop and entering the next loop True 16. Print in Python 3 must be enclosed in parentheses () True 17. The trace operation returns the sum of the diagonal elements of the matrix. So the trace of matrix A is the same as that of its transposed matrix. True 18. 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 linearly variables called principal components False Eigenvalue decomposition can obtain eigenvalues and eigenvectors 19. True The determinant of array A is a scalar 20. True The loss function is used to evaluate the variance of the predicted value and the true value. Usually the betterof the loss function, the better the performance of the model 21. False 22. There are already n data points, if half of it is used for training and the other half is used for testing. The difference between training error and test error will decrease if n increases False 23. TensorFlow is Google’s first-generation proprietary machine learning False (sec gen) 24. TensorFlow supports multiple TPU clusters computing True 25. Data in TensorFlow is represented by tensors and computational graphs True 26. Operations and computation graph of TensorFlow do not necessarily run in the Session False 27. Increasing the size of the convolution kernel will inevitably improve the performance of the convolutional neural network True 28. Recurrent neutral networks are more suitable for image recognition problems False (CNN) 29. Deep learning algorithms can be categorized into supervised learning or unsupervised learning False 30. GAN is a deep learning model. Recently, it is one of the most promising methods for unsupervised learning in complex distribution. True 31. Convolutional neural networks are more suitable for speech recognition problems. False 32. If the convolutional neural network is trained on the ImageNet dataset (object recognition dataset) and we give this model a completely white image as input, then the output probability of this input is equal for all classes False 33. Deep learning can be used to easily export simple mathematical functions from a large amount of multi-dimensional data to describe complicated decision-making interfaces True 34. 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 True 35. Multiple Python statements can be written in one line with commas (,) between the statements. () False 36. Correlation coefficient, also called the linear correlation coefficient, measures a linear relationship between two variables. Its value is a real number greater than 0. () False 37. Convolutional neural networks (CNNs) are better than recurrent neural networks (RNNs) at processing image recognition problems. True Single choice questions : Questions choices Right choice 1. If we want to train decision trees on big data sets with less time, we can A. Reduce the number of trees B. Increase the depth of the tree C. Reduce the depth of the tree D. Increase learning rate (learning rate) D ?? /C Which of the following techniques can you use if you need to change parameters to minimize the cost function? 2. A. Any of the above B. Random search C. Exhaustive search D. Bayesian optimization A or C ? Which of the following operations achieves a similar effect to Dropout in a neural network? 3. A. Bagging B. None of the above C. Boosting D. Stacking B ??? which of the following statements of recurrent neural network is wrong? 4. A. Recurrent neural networks can be expanded according to the time step B. LSTM is also a kind of recurrent neural network C. Recurrent neural network can be abbreviated as RNN D. LSTM cannot solve the problem of gradient disappearance D What is the correct step of the gradient descent algorithm? 1. Calculate the error between the output value and the true value 2. Iterate until the best weight is found 3.forward the input into the network and get the output value 4. Initialize the weights and biases in network by random 5. For each neuron that produces an error, change the corresponding (weight) value to reduce the error. 5. A. 4, 3, 1, 5, 2 B. 5, 4, 3, 2, 1 C. 1, 2, 3, 4, 5 D. 3, 2, 1, 5, 4 A. 4, 3, 1, 5, 2 How to determine the K value in the K means Clustering algorithm? 6. A. random B. A known ????? C. Automatic computer recognition D. Artificial We usually use mini bacth for gradient descent in DNN training. What does the mini batch mean? 7. A. Use part of the hyperparameters for training at each iteration B. Use part of the data set for training at each iteration C. Use part of the features for training at each iteration D. Use part of the erros for training at each iteration B. Use part of the data set for training at each iteration What is the result of the following code execution?[i**i for i in xrange(3)] (ce x n’existe peas?) 8. A. [1, 2,3] B. [0, 1, 4] C. [1, 1, 4] D. （ 1,1,4） C. [1, 1, 4] The following statements are the accuracy、 recall and F1 value of the classification algorithm. what is wrong? 9. A. The recall is the ratio of the number of related documents retrieved to the number of related documents in the document library, which measure the retrieval rate of the system. B. In order to solve the conflict problem of accuracy and recall, F1 score is introduced. C. The accuracy, recall and F1 score are between 0 and 1. The more close the value is to 0, the better performance of the model is . D. The accuracy rate is the ratio of the number of related documents retrieved to the total number of documents retrieved, which measures how accurate the retrival system is. C. The accuracy, recall and F1 score are between 0 and 1. The more close the value is to 0, the better performance of the model is . Which of the following techniques is used to solve overfitting in neural networks? 10. A. Batch normalization B. Regularization C. Dropout D. all of the above A.Batch normalisation B. Regularization C. Dropout What is the type of labels predicted by the regression algorithms? 11. A. Discrete type B. Self-modification C. Continuous type D. Resilience C. Continuous type Which version of Python is generally recommended as a production environment? 12. A. Python 2.6.* B. Python 3.* C. Python 2.5.* D. Python 2.7.* B. Python 3.* Which method can be used for tensor merging in TensorFlow 2.0? 13. A. concat B. unstack C. split D. join A. concat For image recognition problems (such as identifying cats in photos), which kind of neural networks is the most suitable one ? 14. A. Multilayer perceptron B. RNN C. perceptron D. Convolutional neural network D. Convolutional neural network Which of the following is not a legal identifier for python？ 15. A. 40XL B. self C. int32 D. __name__ A. 40XL (it cannot start with a number ) What is the result of the following code execution?x=1def change(a):x+=1print xchange(x) 16. A. 2 B. 1 C. Error D. 3 C What is the L1 norm of vector x=[1,2,3,4,-9,0] 17. A. 19 B. sqrt(111) C. 1 D. 6 A How does Python generally perform grammatical boundary indentation? 18. A. Eight spaces B. Two spaces C. TAB D. Four spaces C. TAB Which of the following statements about supervised learning is correct? 19. A. Decision tree is a kind of supervised learning B. Supervised learning can be trained without labels C. Supervised learning cannot be trained using cross-validation D. Supervised learning is a rule-based algorithm A. Decision tree is a kind of supervised learning which one of the following indexes is used to select frequent sets in Apriori？ 20. A. Support B. Accuracy C. Confidence D. alpha A??? Which operator in Python can be used to calculate the integer quotient? A. / B. // B 21. C. % D. ? which one of the following statements of SVM is wrong? 22. A. The support vector is the vector point closest to the hyperplane B. SVM cannot be used for the classification of nonlinear data sets C. SVM can use multiple kernel functions D. SVM can be used to solve classification problems B. SVM cannot be used for the classification of nonlinear data sets which operation is uesed in Bagging to reduce the varience? 23. A. optimization B. more data C. average D. Multiple iterations(solvve underfitting) C ?? s=’ abcdefg’ , then the value of s[3:5] is 24. A. de B. def C. cd D. cde C 25. In May 1997, a computer defeated the reigning world chess champion Garry Kasparov by a score of 3.5-2.5 in the famous "Man vs Machine" chess match. What is the name of this computer? Deep blue Alpha go (2016) 26. Who first defined AI in 1950 and proposed a machine intelligence test model? Alan Mathison Turing 27. Machines must be knowledgeable before they become smart. A research area of AI mainly deals with automatic computer acquisition of knowledge and skills for self improvement. What is this research branch? Machine Learning 28. What kind of academic discipline is AI ? A. Psychology and Physiology B. Mathematics and Physiology C. Linguistics D. An interdisciplinary subject D . An interdisciplinary subject Which of the following is not a Python operator? 29. Inference operator 30. Which of the following statements about Python regular expression is not true? re regular expression can only process string data, not value data 31. Which of the following is not a Python file operation step? delete 32. Linear algebra is the branch of mathematics concerning linear equations, functions, and their representations. In a linear relationship, all variables must be to the ( ) power first [Show More]

Last updated: 1 year ago

Preview 1 out of 244 pages