You are on page 1of 2

Helwan University | Faculty of Computers & Artificial Intelligence | Mainstream Prog.

Artificial Intelligence (CS-361 | AI-301) Midterm Exam Thur. Nov. 16th, 2023
Time Allowed: 30 Minutes – Num. of Pages: 2 Pages (1 Sheet) – Marks: 15

Name: ........................................................................................... ID: ....................................................

Q.1 Choose the correct answer: [ 17 QUESTIONS – 17 MARKS – INCLUDING 2 BONUS MARKS ]

1) In the context of self-driving cars, an AI approach that is concerned with making the vehicle's actions
and decisions closely like those of a human driver is an example of _______, while an AI-powered
recommendation system that suggests movies for you to watch based on your preferences and viewing
history is an example of _______. (a) Acting Humanly / Acting Rationally (b) Acting Rationally / Acting
Humanly (c) Thinking Humanly / Thinking Rationally (d) Both are examples of Acting Humanly (e) Both are
examples of Acting Rationally

2) In a medical diagnosis system that learns from a database of patient records to predict diseases, the
concept that is primarily used is _______, while in a store, the system that identifies patterns in sales data
to optimize inventory and product placement is primarily applying _______. (a) Data Mining / Machine
Learning (b) Machine Learning / Data Mining

3) In a weather forecasting system that processes vast amounts of unstructured climate data for
predictions, the concept that is primarily used for handling this large dataset is _______. (a) Artificial
Intelligence (b) Data Mining (c) Big Data (d) Deep Learning (e) None of the above.

4) In a robot designed to clean your house, which component of PEAS represents the robot's cleaning
brushes? (a) Performance Measure (b) Environment (c) Sensors (d) Actuators

5) Which agent type are the following: (i) An energy-saving system for a smart home electricity
consumption while maintaining comfort levels – considering that the unsuccessful reduction is annoying
the residents. (ii) A language translation app aiming to provide accurate translations. (a) Cost-based /
Goal-based (b) Goal-based / Cost-based (c) Both are Goal-based (d) Both are Cost-based

6) Match the following AI concepts with their real-world characteristics and/or applications:

a. The Turing Test 1. AI systems with human-like understanding and feeling and general problem-
b. The Chinese solving abilities.
2. AI systems for specific tasks like virtual assistants or recommendation systems,
Room Argument
without human-like consciousness.
c. Weak AI
3. Challenging the idea that “computer manipulation equals true understanding”.
d. Strong AI 4. Assessing if a computer can mimic human conversation.
(a) 1 2 3 4 (b) 2 1 3 4 (c) 3 2 1 4 (d) 1 2 4 3 (e) 4 3 2 1

7) Match the following machine learning tasks with their descriptions:

a. Classification 1. Grouping similar data points into categories.


b. Regression 2. Predicting a continuous value output from input data.
3. Reducing the number of features or dimensions in data while
c. Dimensionality Reduction
preserving important information.
d. Clustering 4. Assigning data points to predefined categories or labels.
(a) 4 2 3 1 (b) 2 4 3 1 (c) 3 2 1 4 (d) 4 1 3 2 (e) 1 3 4 2

8) In a facial recognition AI system for security, which metric should be minimized to prevent unauthorized
access? (a) True Positives. (b) True Negatives. (c) False Positives. (d) False Negatives. (e) All of the above.

Page 1 of 2
Helwan University | Faculty of Computers & Artificial Intelligence | Mainstream Prog.
Artificial Intelligence (CS-361 | AI-301) Midterm Exam Thur. Nov. 16th, 2023
Time Allowed: 30 Minutes – Num. of Pages: 2 Pages (1 Sheet) – Marks: 15

9) Match the following machine learning techniques with their real-world applications:

1. Discovering hidden patterns in customer purchasing behaviour for


a. Supervised Learning
market segmentation.
b. Unsupervised Learning 2. Teaching a spam filter to classify emails as spam or not.
c. Reinforcement 3. Combining multiple weather prediction models to improve forecasting
Learning accuracy.
4. Training a robot to navigate through a maze by receiving rewards for
d. Ensemble Learning
successful movements.
(a) 2 3 4 1 (b) 2 1 4 3 (c) 3 2 1 4 (d) 4 1 2 3 (e) 2 4 3 1
A
10) Consider the game tree (right →) in which root is a maximizing node
and children are visited left to right. What nodes will be pruned by the
alpha-beta pruning? (a) I J (b) E F (c) E G (d) F H (e) None of these.

Sudoku is a logic-based number-placement puzzle where the objective is E F G I J


H
to fill a 9x9 grid with digits from 1 to 9 in such a way that each row, each
column, and each of the nine 3x3 sub-grids contains all the numbers from 2 3 1 2 4 6
1 to 9 without repetition. (For questions 11 to 17)

11) If you decided to use the Backtracking search algorithm to find a solution
for the given Sudoku puzzles, which of the two given states (puzzles) is likely
to require more time from the algorithm to solve? (a) Top. (b) Bottom.
(c) Both are likely to require a similar time to solve.

12) Using the heuristic h(N) “selecting the cell with the fewest possible
candidate values (i.e., the smallest domain) to fill first”, may reduce the
branching factor in the search space. (a) True. (e) False.

13) Which of the following search algorithms make use of this advice: "When
solving a Sudoku puzzle, filling first the cells that are part of a unit with fewer
filled cells is more likely to lead to a valid solution." (a) Backtracking.
(b) Bidirectional. (c) Minimax. (d) Depth-First. (e) Best-First.

14) When designing an AI-based solver, which of the following is similar to


Sudoku in terms of the number of start states? (a) Chess. (b) Tic-Tac-Toe.
(c) 8-puzzle. (d) Connect Four. (e) All of the above.

15) If the bottom state is the configuration (puzzle) given by the user to the
AI-agent to solve, what is its g(N): (a) 36 (b) 29 (c) 0 (d) 81 (e) 52

16) Which search algorithm will most likely get stuck in a local optimum when
solving Sudoku? (a) Breadth-First. (b) Depth-First. (c) Best-First. (d) Hill
Climbing. (e) Depth-First with Iterative Deepening.

17) An Intelligent Sudoku Solver’s environment is: (a) Episodic. (b) Strategic. (c) Stochastic. (d) Multi-
Agent. (e) Discrete.

Page 2 of 2

You might also like