这是一篇来自美国的关于人工智能2022年学期秋季期末试卷的AI代写
Instructions:
- Before solving the exam, you should read the Ed post titled “Final Exam Next Week”.
- You should fill out this PDF and submit it on BOTH Gradescope and Canvas. We reserve the right to deduct points if you don’t submit on both platforms.
- You have an unlimited number of submissions until the deadline. You can either type directly into the PDF or you can print, hand-write, and scan your solutions. If you decide to type into the PDF, we strongly recommend using Adobe Acrobat Reader DC. If you are on MacOS, please do not use Preview, as we have seen major issues with it in the past. Other programs may not save your answers.
Thus, always make sure to keep a backup of your answers.
- You should submit only a single PDF. Also, make sure your typed answers appear clearly in the Gradescope Preview. After each question, a blank page is provided if you wish to highlight some of your calculations. However, you should not expect any points to be awarded for your calculations (only your final answers will be considered). The TAs reserve the right to assign partial credits based on your calculations. Make sure to submit ALL the pages of the exam, not only the completed ones.
- You must fill out the checklist at the end of the exam. The exam may not be graded if it is left blank.
- The exam is open-book, open-note, and open video lectures, with no time limit aside from the duration of the exam. No internet use is allowed, except for e-text versions of the textbook, this semester’s CS6601 course materials, Ed, and any links provided in the PDF itself. No resources outside this semester’s 6601 class should be used. Do not discuss the exam on Ed, Slack, or any other platform. In particular, do not post publicly about the exam. If there is a question for the teaching staff, please make it private on Ed and tag it as Final Exam with the question number in the subject line (for example, a question on Search would be “Final Exam #2”). Please make different posts for different questions.
- Global Rounding Rule: Please round all your final answers to 6 decimal places. Don’t round intermediate results. You can use round(answer, 6) function in Python for help. You may not receive full credit if your answers are not given to the specified precision.
- Points breakdown is provided below.
Problem 1: Game Playing [7 points]
Harry is a struggling sophomore and loves playing games. Often, this habit affects his progress in college as he is unable to turn in assignments on time as he is caught up in completing game missions instead.
This semester Harry took a CS course called Game Playing and Artificial Intelligence where he is learning about how two player and zero-sum games use adversarial search to apply AI and how modern algorithms and machines use these algorithms in everyday applications. Surprisingly, Harry loves this course and has decided to focus on his studies this semester but is unable to solve and understand some questions about the concept of Iterative Deepening. Therefore, he needs a peer to explain to him and solve such questions.
Help Harry gain interest in studies and his potential to be a shining student again by solving the following questions.
(a) (1.5 points) Which of the following is (are) the best ordering of the nodes A, B, C, and D (best ordering is the ordering that produces the most pruning) for alpha-beta pruning evaluation up to depth 2 with Iterative Deepening in an attempt to improve the effectiveness of alpha-beta pruning when running up to depth 3 (see Figure 1). Assume that the procedure uses the information it has acquired up to a given depth to try to improve the order of evaluations later. The static values for nodes A, B, C, and D are given with the definition of S and the depths for the tree are given with definition of D.
(b) (1.5 points) Which of the following is (are) the best ordering of the nodes A, B, C, and D (best ordering is the ordering that produces the most pruning) for alpha-beta pruning evaluation up to depth 2 with Iterative Deepening in an attempt to improve the effectiveness of alpha-beta pruning when running up to depth 3 (see Figure 2). Again, assume the procedure uses the information it has acquired up to a given depth to try to improve the order of evaluations later. The static values for nodes A, B, C, and D are given with the definition of S and the depths for the tree are given with the definition of D.
(c) (1.5 points) Solve the given game tree in Figure 3 using minimax:
Enter the final propagated value at the top:
Enter the number of leaf nodes pruned if alpha-beta pruning is applied from left to right to solve the game tree in Figure 3 (we are only counting leaf (terminal) nodes and not branches)
(d) (2.5 points) For Figure 1 and Figure 2, assume that we run the alpha-beta pruning to depth 3 and ignore the static values assigned at depth 2 previously used for iterative deepening.
How many more leaf nodes (compared to without reordering and using the left-to-right evaluation rule) will be pruned if the correct reordering (from the answer to question (a)) at depth 2 is applied (see Figure 1 and ignore the static values assigned at depth 2 previously used for iterative deepening.)?
How many more leaf nodes (compared to without reordering and using the left-to-right evaluation rule) will be pruned if the correct reordering (from the answer to question (b)) at depth 2 is applied (see Figure 2 and ignore the static values assigned at depth 2 previously used for iterative deepening.)?