18. Perform a DFS (Depth-First Search) traversal of this graph by starting at node 1 and going in the order of numbers in Arabic when picking which neighbor to visit first. Which of the following is the CORRECT traversal order?
  
(A) 1, 2, 6, 4, 5, 3, 7
(B) 1, 2, 4, 6, 5, 3, 7
(C) 1, 2, 4, 5, 6, 3, 7
(D) 1, 2, 4, 3, 7, 5, 6
(E) 1, 2, 4, 5, 3, 7, 6

14. Let AB*CD-/E+ be an postfix expression. If A=6, B=3, C=4, D=5, and E=8. What is the result of this postfix expression after evaluation? (A) -10 (B) -8 (C) 8 (D) 9 (E) 10

15. Let the height of a tree be the number of nodes along the longest path from the root node to the leaf nodes. Consider the integers 30, 41, 25, 29, 94, 37, 70, 23, 65, 75 in the specified order to develop a binary search tree. Which of the following is TRUE? (A) The node for 37 is an internal node. (B) The root node is 41. (C) The node for 70 has only one child. (D) The height of the tree is 5. (E) The node for 75 is a child of the node for 94.

16. What will be the output of the following C code ?   (A) 2 (B) 3 (C) 4 (D) 5 (E) None of the above 

17. Consider the following instruction mix for a processor: ALU operations: 40%, uses 4 cycles. Branch operations: 30%, uses 4 cycles. Memory references: 30%, uses 5 cycles. The un-pipelined processor has a clock cycle time of 1 ns. The pipelined processor has a clock cycle time of 1.2 ns. Suppose that we ignore any latency and hazards, and assume that the pipelined processor has an ideal CPI (cycles per instruction) of 1. How much speedup can be achieved when comparing the un-pipelined processor and the pipelined processor? (A) 3.28 (B) 3.35 (C) 3.58 (D) 3.86 (E) 3.98

19. Consider the following C code:When the code is executed, what is the final value of x? (A) 20 (B) 24 (C) 28 (D) 32 (E) 36

20. What will be the output of the following C code? (A) 36 (B) 39 (C) 93 (D) 63 (E) 96

21. In machine learning, one of the most common problems that will occur when we are training a model is overfitting. There are some techniques that could ease or even solve this problem. Which of the following technique does NOT alleviate overfitting problem? (A) Adding dropout between layers. (B) Increasing training epochs. (C) Data augmentation. (D) Early stopping while training model. (E) Using L1/L2 regularization.

22. Consider the following C program code and statements (1)~(5). (1) node[1].num = 0 (2) node[1].num = 1 (3) node[2].ptr-＞num = 9 (4) node[9].ptr-＞ptr-＞num = 121 (5) node[9].ptr-＞num = 100 Which of the following statements is CORRECT? (A) (1)(3)(4)(5) (B) (1)(3) (C) (2)(3) (D) (2)(3)(4)(5) (E) (4)(5)

23. The following is a C program code: Which one of the following statements is CORRECT? (A) When a=8, b=5, c=6, then A is 81, B is 5. (B) When a=8, b=5, c=6, then A is 81, B is 6. (C) When a=8, b=5, c=6, then A is 17, B is 5. (D) When a=8, b=5, c=6, then A is 17, B is 6. (E) When a=8, b=5, c=6, then A is 16, B is 6.

24. Suppose there are 16 users in the system, and they use symmetric encryption to achieve confidential communication, how many keys need to be managed? (A) 125 (B) 256 (C) 120 (D) 128 (E) 160