ECON 100 UNR ACTUAL TEST PAPER 2026
COMPLETE QUESTIONS AND CORRECT
ANSWERS GRADED A+
▶ During neural network training, which of the following values is
continuously updated using the gradient descent method to minimize the
loss function?
A. Hyperparameters
B. Feature value
C. Number of samples
D. Parameters - Answer: D
▶ Without considering any regularization terms, the support vectors of a
support vector machine (SVM) are composed of ( ).
A. Points on the separating hyperplane
B. The points farthest from the separating hyperplane
C. The points closest to the separating hyperplane
D. Points of a certain type - Answer: C
▶ Within the broad landscape of large language models built on the
Transformer architecture, which roadmap does GPT belong to?
A. Encoder-Only
B. Decoder-Only
C. Encoder-Decoder
D. Decoder-Encoder - Answer: B
▶ Which of the following is NOT a use case for natural language
processing?
A. Public opinion analysis
B. Machine translation
C. Text classification
D. Image recognition - Answer: D
▶ Which of the following statements about different optimizers is false?
A. RMSprop solves the problem of the Adagrad optimizer ending too early.
, B. Compared with RMSprop, Adagrad is more sensitive to gradient
changes.
C. Both Adagrad and RMSprop can set adaptive learning rate for
parameters.
D. Adam requires bias correction for initial iterations. - Answer: B
▶ Which of the following is NOT an Al deep learning framework?
A. TensorFlow
B. Scikit-learn
C. Pytorch
D. Caffe - Answer: B
▶ Which of the following statements is true about classification models and
regression models in machine learning?
A. For regression problems, the output variables are discrete values. For
classification problems, the output variables are continuous.
B. The most commonly used indicators for evaluating regression and
classification problems are the accuracy and the recall rate.
C. Overfitting may occur in both regression and classification problems.
D. Logistic regression is a typical regression model. - Answer: C
▶ Which of the following are used in TensorFlow to describe the
computation process?
A. Parameter
B. Session
C. Data flow
D. Tensor - Answer: C
▶ Which of the following is NOT a characteristic of PyTorch?
A. Python first
B. Easy to debug
C. Extensive ecosystem
D. Default use of static maps - Answer: D
▶ Which of the following statements is true about the training of large
models?
A. Foundation models cannot be directly deployed.
B. The pre-training of large models requires a large amount of high-quality
data.
COMPLETE QUESTIONS AND CORRECT
ANSWERS GRADED A+
▶ During neural network training, which of the following values is
continuously updated using the gradient descent method to minimize the
loss function?
A. Hyperparameters
B. Feature value
C. Number of samples
D. Parameters - Answer: D
▶ Without considering any regularization terms, the support vectors of a
support vector machine (SVM) are composed of ( ).
A. Points on the separating hyperplane
B. The points farthest from the separating hyperplane
C. The points closest to the separating hyperplane
D. Points of a certain type - Answer: C
▶ Within the broad landscape of large language models built on the
Transformer architecture, which roadmap does GPT belong to?
A. Encoder-Only
B. Decoder-Only
C. Encoder-Decoder
D. Decoder-Encoder - Answer: B
▶ Which of the following is NOT a use case for natural language
processing?
A. Public opinion analysis
B. Machine translation
C. Text classification
D. Image recognition - Answer: D
▶ Which of the following statements about different optimizers is false?
A. RMSprop solves the problem of the Adagrad optimizer ending too early.
, B. Compared with RMSprop, Adagrad is more sensitive to gradient
changes.
C. Both Adagrad and RMSprop can set adaptive learning rate for
parameters.
D. Adam requires bias correction for initial iterations. - Answer: B
▶ Which of the following is NOT an Al deep learning framework?
A. TensorFlow
B. Scikit-learn
C. Pytorch
D. Caffe - Answer: B
▶ Which of the following statements is true about classification models and
regression models in machine learning?
A. For regression problems, the output variables are discrete values. For
classification problems, the output variables are continuous.
B. The most commonly used indicators for evaluating regression and
classification problems are the accuracy and the recall rate.
C. Overfitting may occur in both regression and classification problems.
D. Logistic regression is a typical regression model. - Answer: C
▶ Which of the following are used in TensorFlow to describe the
computation process?
A. Parameter
B. Session
C. Data flow
D. Tensor - Answer: C
▶ Which of the following is NOT a characteristic of PyTorch?
A. Python first
B. Easy to debug
C. Extensive ecosystem
D. Default use of static maps - Answer: D
▶ Which of the following statements is true about the training of large
models?
A. Foundation models cannot be directly deployed.
B. The pre-training of large models requires a large amount of high-quality
data.
