5. A program has a one-dimensional array of integers A. Each element of the array is a four-byte integer. The array is stored in contiguous memory space. If the memory address of A[100] is 2000, then which of the following is the address of A[50]?
(A) 1000
(B) 1800
(C) 996
(D) 1004
(E) 1200

1. A software engineer is optimizing a real-time embedded system. The processor being used follows the RISC architecture. Which of the following optimizations is best suited for this processor? (A) Minimizing pipeline stalls by keeping instruction execution times uniform (B) Using highly optimized microcode for complex instructions (C) Utilizing variable-length instructions to reduce code size (D) Optimizing for micro-op fusion to minimize instruction count (E) Relying on Instruction-Level Parallelism (ILP) to reduce execution cycles

2. Which of the following is the goal of a page replacement algorithm?(A) Decrease disk access time (B) Increase CPU utilization(C) Reduce page faults (D) Reduce the number of processes(E) Increase memory fragmentation

3. Suppose you are using the MNIST dataset to train a neural network model to recognize handwritten digits. Which of the following settings is practical? (A) To accelerate convergence, you can set the learning rate to 256 initially. (B) To make the model more generalized, you can randomly apply image mirroring with a probability of 1/2. (C) To avoid gradient explosion, you can limit the gradient size to 2. (D) To avoid difficulty in training the model, the order of the images should be fixed during training. (E) Since the goal of training is to classify digits from 0 to 9, which is a regression problem, you should set the number of output nodes to 1 and use mean square error as the loss function.

4. In the supervised learning, what is the purpose of a grid search?(A) To shift the decision boundary or regression line.(B) To split the data into training and testing sets.(C) To select the best features for a model.(D) To prevent overfitting.(E) To search for the best combination of hyperparameters for a given model.

6. Huffman coding is an efficient method of compressing data without losing information. Symbols that appear more often will be encoded as a shorter bit string, while symbols that are not used much will be encoded with longer strings. If Huffman coding is used to compress a sequence with 5 different codewords, which of the following probability distributions will lead to the WORST compression ratio? (A) 0.2, 0.2, 0.2, 0.2, 0.2 (B) 0.02, 0.02, 0.03, 0.03, 0.9 (C) 0.1, 0.1, 0.2, 0.3, 0.3 (D) 0.1, 0.2, 0.2, 0.2, 0.3 (E) 0.05, 0.05, 0.2, 0.2, 0.5

7. A Distributed Denial of Service (DDoS) attack attempts to consume the target’s resources so that it cannot provide service. What is a DDoS attack and how does it work? (A) A type of malware that infects multiple machines and spreads through a network. (B) A type of cyber attack that targets an individual user's personal information. (C) A type of attack that attempts to overload a server or network with traffic from multiple sources. (D) A type of attack that exploits vulnerabilities in software to gain unauthorized access to a system. (E) A type of attack that intercepts and manipulates data transmitted over a network.

8. Which of the following statements is CORRECT about decision trees in data mining?(A) Decision trees cannot handle categorical data.(B) A deeper decision tree always improves model performance.(C) Decision trees do not support ensemble learning.(D) A decision tree must always be binary.(E) Pruning is used to reduce overfitting in decision trees.

9. As computer systems evolve, greater performance can be achieved by taking advantage of improvements in technology, such as faster circuitry. Which of the following is a method used in computer organization to increase the performance of a CPU by allowing multiple instructions to be executed simultaneously? (A) Multithreading (B) Multiprocessing (C) Pipelining (D) Supercomputing (E) Parallel computing

10. Which operating system scheduling algorithm may lead to starvation?(A) First-Come, First-Served (FCFS) (B) Shortest Job First (SJF)(C) Round Robin (D) Multilevel Queue(E) First-In, First-Out (FIFO)

11. Which of the following statements about machine learning is CORRECT? (A) We can generate new features by linearly or non-linearly combining the original features, which is a method of feature engineering. (B) In a dataset, all pairwise combinations of data point vectors are called the support vectors of the dataset. (C) During the training process of a neural network, the validation loss typically fluctuates, but the training loss will always decrease. (D) In k-NN classification, k should be set equal to the size of the dataset so that the entire dataset is considered before making a decision. (E) When training a deep learning model, if the accuracy of the test set is higher than that of the training set, it should be considered evidence of overfitting.

114年 - 114 學士後醫學系招生考試試題：計算機概論與程式設計#137105

113年 - 113 學士後醫學系招生考試試題：計算機概論與程式設計#119574

112年 - 112 學士後醫學系招生考試試題：計算機概論與程式設計#137106

111年 - 111 學士後醫學系招生考試試題：計算機概論與程式設計#137107

111年 - 111 國立中山大學_學士後醫學系招生考試：計算機概論與程式設計#113521

111年 - 111 學士後醫學系招生考試：計算機概論與程式設計#107630

110年 - 110 學士後_醫學系︰計算機概論與程式設計#100246

109年 - 109 高雄醫學大學_學士後醫學系招生考試：計算機概論與程式設計#86938