Unit 1: Foundations and AI Problem Solving - Practice Quiz

CSE276 — Artificial Intelligence Foundations 60 Questions
0 Correct 0 Wrong 60 Left
0/60

1 What is the main goal of Artificial Intelligence?

Introduction to Artificial Intelligence Easy
A. To replace all forms of communication
B. To store files without processing them
C. To make machines perform tasks requiring human-like intelligence
D. To increase the physical size of computers

2 Which of the following is an example of an Artificial Intelligence task?

Introduction to Artificial Intelligence Easy
A. Connecting a power cable
B. Turning on a light switch
C. Printing a blank page
D. Recognizing speech

3 Who is commonly credited with coining the term "Artificial Intelligence"?

History and evolution of Artificial Intelligence Easy
A. John McCarthy
B. Alan Turing
C. Tim Berners-Lee
D. Charles Babbage

4 Which event is often considered the formal beginning of AI as an academic field?

History and evolution of Artificial Intelligence Easy
A. The Dartmouth workshop in 1956
B. The first commercial airplane flight
C. The launch of the World Wide Web
D. The invention of the telephone

5 Which type of AI is designed to perform a specific task?

Characteristics and types of Artificial Intelligence Easy
A. General AI
B. Narrow AI
C. Universal AI
D. Super AI

6 Which characteristic allows an AI system to improve from experience?

Characteristics and types of Artificial Intelligence Easy
A. Storage
B. Learning
C. Wiring
D. Printing

7 What is the relationship between Machine Learning and Artificial Intelligence?

Artificial Intelligence vs Machine Learning vs Deep Learning vs Data Science Easy
A. Machine Learning is a subset of AI
B. They are completely unrelated fields
C. AI is a subset of Machine Learning
D. Machine Learning only studies hardware

8 Deep Learning commonly uses which structures to learn from data?

Artificial Intelligence vs Machine Learning vs Deep Learning vs Data Science Easy
A. Mechanical gears
B. Electrical switches only
C. Artificial neural networks
D. Paper-based tables

9 Which field focuses on collecting, analyzing, and interpreting data?

Artificial Intelligence vs Machine Learning vs Deep Learning vs Data Science Easy
A. Civil Engineering
B. Network Cabling
C. Graphic Design
D. Data Science

10 What is an intelligent system?

Intelligent systems and Artificial Intelligence lifecycle Easy
A. A system that only stores documents
B. A system that senses, reasons, and acts
C. A system that works without any input
D. A system that only displays images

11 Which activity is usually part of the AI lifecycle?

Intelligent systems and Artificial Intelligence lifecycle Easy
A. Removing all system objectives
B. Collecting and preparing data
C. Deleting useful input data
D. Avoiding system evaluation

12 How can AI assist healthcare professionals?

Applications of Artificial Intelligence in healthcare, agriculture, finance, education, manufacturing, robotics and smart cities Easy
A. By removing patient records
B. By helping detect diseases in medical images
C. By preventing medical measurements
D. By replacing hospital buildings

13 Which is an example of AI use in agriculture?

Applications of Artificial Intelligence in healthcare, agriculture, finance, education, manufacturing, robotics and smart cities Easy
A. Painting farm fences
B. Detecting crop diseases
C. Increasing soil color
D. Replacing all farming tools

14 Which financial activity can use Artificial Intelligence?

Applications of Artificial Intelligence in healthcare, agriculture, finance, education, manufacturing, robotics and smart cities Easy
A. Changing bank building designs
B. Detecting fraudulent transactions
C. Removing account verification
D. Printing currency by hand

15 How can AI support education?

Applications of Artificial Intelligence in healthcare, agriculture, finance, education, manufacturing, robotics and smart cities Easy
A. By removing all assessments
B. By preventing student feedback
C. By replacing classroom furniture
D. By providing personalized learning

16 Which is a common use of AI in manufacturing?

Applications of Artificial Intelligence in healthcare, agriculture, finance, education, manufacturing, robotics and smart cities Easy
A. Changing factory colors
B. Manual product labeling
C. Predictive maintenance
D. Removing machine sensors

17 In an AI problem, what does the initial state represent?

Artificial Intelligence problem formulation and state-space representation Easy
A. The final answer
B. The search cost
C. The starting condition
D. The list of possible actions

18 What does a goal state represent in a problem-solving task?

Artificial Intelligence problem formulation and state-space representation Easy
A. A desired solution condition
B. An unused input value
C. A random starting point
D. An unavailable action

19 A problem is called deterministic when:

Problem characteristics Easy
A. The goal is never defined
B. No action can change the state
C. Every action has a random result
D. Each action has a predictable result

20 What is the search space of an AI problem?

Search space and solution space Easy
A. The set of states considered during search
B. The physical area of a laboratory
C. The memory size of the computer
D. The number of users of a system

21 A hospital system recommends possible diagnoses by comparing a patient's symptoms with patterns learned from previous cases. Which AI capability is primarily demonstrated?

Introduction to Artificial Intelligence Medium
A. Manual rule execution
B. Hardware resource allocation
C. Random data generation
D. Perception and classification

22 Why did expert systems become less popular during periods known as AI winters?

History and evolution of Artificial Intelligence Medium
A. They required extensive hand-coded knowledge
B. They depended entirely on image sensors
C. They were designed only for entertainment
D. They could not process any symbolic data

23 A recommendation system improves its suggestions as it receives more user feedback. Which characteristic of AI does this best illustrate?

Characteristics and types of Artificial Intelligence Medium
A. Adaptability through learning
B. Fixed behavior without updates
C. Dependence on physical movement
D. Restriction to arithmetic operations

24 A chess program is highly effective at chess but cannot diagnose diseases or understand ordinary conversation. How should this system be classified?

Characteristics and types of Artificial Intelligence Medium
A. Human-level artificial consciousness
B. General or strong AI
C. Self-aware superintelligence
D. Narrow or weak AI

25 Which statement best describes the relationship among AI, machine learning, and deep learning?

Artificial Intelligence vs Machine Learning vs Deep Learning vs Data Science Medium
A. Deep learning uses only manually written rules
B. Artificial intelligence is a subset of deep learning
C. Machine learning is unrelated to artificial intelligence
D. Deep learning is a subset of machine learning

26 A data science team cleans sales data, creates dashboards, identifies trends, and uses a predictive model to estimate future demand. Which conclusion is most accurate?

Artificial Intelligence vs Machine Learning vs Deep Learning vs Data Science Medium
A. The work combines data science and machine learning
B. The work represents only deep learning
C. The work excludes all forms of AI
D. The work is limited to robotics control

27 During an AI project, a model performs well during development but poorly after deployment because real-world data has changed. Which lifecycle activity should address this issue?

Intelligent systems and Artificial Intelligence lifecycle Medium
A. Continuous monitoring and model updating
B. Replacement of all input variables
C. Initial problem definition only
D. Permanent removal of evaluation data

28 An intelligent traffic system receives sensor data, evaluates congestion, changes signal timings, and observes the resulting traffic flow. Which property of an intelligent system is most evident?

Intelligent systems and Artificial Intelligence lifecycle Medium
A. Perception, decision, and action
B. One-time computation without feedback
C. Manual control without environmental input
D. Data storage without interpretation

29 A farming system combines satellite images, soil measurements, and weather forecasts to determine where irrigation is needed. What is the main AI benefit?

Applications of Artificial Intelligence in healthcare, agriculture, finance, education, manufacturing, robotics and smart cities Medium
A. Replacement of all agricultural workers
B. Complete elimination of weather uncertainty
C. Precision resource management
D. Conversion of soil into digital information

30 A bank uses an AI model to flag unusual transactions for review. Which design concern is especially important for this application?

Applications of Artificial Intelligence in healthcare, agriculture, finance, education, manufacturing, robotics and smart cities Medium
A. Ignoring historical transaction patterns
B. Maximizing the number of blocked customers
C. Using only the transaction amount
D. Balancing fraud detection with false alarms

31 In an adaptive education platform, a learner repeatedly makes errors with fractions. What is an appropriate AI-based response?

Applications of Artificial Intelligence in healthcare, agriculture, finance, education, manufacturing, robotics and smart cities Medium
A. Recommend targeted practice on fractions
B. Increase difficulty in unrelated subjects
C. Assign identical content to every learner
D. Remove all mathematical exercises

32 A manufacturing robot detects a defective component using a camera and removes it from a production line. Which combination of AI functions is involved?

Applications of Artificial Intelligence in healthcare, agriculture, finance, education, manufacturing, robotics and smart cities Medium
A. Audio compression and file transfer
B. Database backup and encryption
C. Visual perception and physical action
D. Text summarization and translation

33 For a route-finding problem, which set correctly represents the essential components of a state-space formulation?

Artificial Intelligence problem formulation and state-space representation Medium
A. Rows, columns, keys, and indexes
B. Training labels, epochs, gradients, and loss
C. Pixels, filters, pooling, and activation
D. Initial state, actions, transition model, and goal test

34 In a sliding-tile puzzle, two board arrangements differ only because one tile has moved into the empty position. How should this relationship be represented?

Artificial Intelligence problem formulation and state-space representation Medium
A. As a change in the goal definition
B. As two unrelated problem instances
C. As an increase in the heuristic value only
D. As a transition produced by a legal action

35 A delivery robot operates in a warehouse where other robots move unpredictably and sensor readings are sometimes incomplete. Which problem characteristics apply?

Problem characteristics Medium
A. Dynamic and partially observable
B. Static and fully observable
C. Single-state and independent
D. Deterministic and completely known

36 Why is an action sequence often preferred over a single action when solving a navigation problem?

Problem characteristics Medium
A. A single action always changes every state
B. Navigation problems have no possible final state
C. Actions cannot be represented in a search tree
D. The goal usually requires multiple state transitions

37 A search algorithm explores 1,000 reachable states, but only 18 of them satisfy the goal condition. What does the set of 18 states represent?

Search space and solution space Medium
A. The goal states within the search space
B. The branching factor of the search
C. The initial state of the problem
D. The complete set of expanded states

38 A problem has a branching factor of and a maximum search depth of . Approximately how many nodes can appear in the complete search tree, including the root?

Search space and solution space Medium
A.
B.
C.
D.

39 An emergency response AI must find routes quickly, but the routes should also be short and safe. Which performance evaluation is most appropriate?

Performance measures Medium
A. Measure only whether any route exists
B. Measure only the program's memory size
C. Measure time, route cost, and safety
D. Measure only the number of generated states

40 A search problem has a reliable heuristic estimating the remaining distance to the goal. Which approach can use this information to prioritize promising states?

Overview of Artificial Intelligence-based problem-solving approaches Medium
A. Manual enumeration without transitions
B. Heuristic or informed search
C. Blind search without state ordering
D. Random sampling without evaluation

41 A delivery agent selects the action with the highest expected utility given its sensor history and probabilistic model. The selected action later causes a poor outcome because an unlikely road closure occurred. Which conclusion is most accurate?

Introduction to Artificial Intelligence Hard
A. The agent is rational only if its internal model predicts every environmental event correctly
B. The agent can still be rational because rationality depends on expected performance given available evidence
C. The agent is non-intelligent because probabilistic models cannot support autonomous decisions
D. The agent is irrational because a rational agent must always achieve the best realized outcome

42 Which sequence places the following AI milestones in the correct chronological order?

History and evolution of Artificial Intelligence Hard
A. Dartmouth workshop, second AI winter, commercial expert systems, AlexNet breakthrough
B. Commercial expert systems, second AI winter, Dartmouth workshop, AlexNet breakthrough
C. Dartmouth workshop, commercial expert systems, second AI winter, AlexNet breakthrough
D. Commercial expert systems, Dartmouth workshop, second AI winter, AlexNet breakthrough

43 An AI laboratory demonstrates strong prototypes, but deployment collapses because computing costs are excessive, systems are brittle outside curated examples, and promised capabilities were overstated. Which historical pattern does this most closely reproduce?

History and evolution of Artificial Intelligence Hard
A. The rise of expert systems caused by universal success in transferring domain knowledge
B. The deep-learning resurgence caused by declining data availability and smaller computational models
C. An AI winter caused by unmet expectations, limited resources, and reduced institutional funding
D. The transition from symbolic AI to cybernetics caused by complete theoretical convergence

44 A system can outperform every human at chess but cannot transfer its learned strategies to contract analysis, household navigation, or medical diagnosis without task-specific redesign. How should it be classified?

Characteristics and types of Artificial Intelligence Hard
A. Narrow AI because its competence is exceptional but confined to a specialized domain
B. Self-aware AI because strategic search requires an explicit model of subjective experience
C. Strong AI because superhuman accuracy implies human-level understanding across domains
D. Artificial general intelligence because it exceeds human performance in one intellectual task

45 A team cleans transaction data, performs statistical exploration, trains a transformer to classify fraud, and embeds the model in an autonomous case-routing system. Which characterization is most precise?

Artificial Intelligence vs Machine Learning vs Deep Learning vs Data Science Hard
A. The entire project is deep learning because every project stage depends on a neural classifier
B. Data science supports analysis, deep learning is the model method, and the deployed agent is an AI system
C. Data science covers the full project, while AI, machine learning, and deep learning are equivalent terms
D. AI supplies data preparation, data science supplies autonomy, and deep learning supplies evaluation

46 Which system is an example of AI problem solving that does not inherently require machine learning?

Artificial Intelligence vs Machine Learning vs Deep Learning vs Data Science Hard
A. A clustering pipeline that discovers customer segments from unlabeled transaction vectors
B. A convolutional network that estimates product defects from labeled manufacturing images
C. A reinforcement-learning agent that improves routing through repeated simulated interactions
D. A rule-based planner that searches a known action model to construct a valid warehouse schedule

47 A loan model retains high aggregate accuracy after deployment, but approval errors rise sharply for a newly introduced applicant category. Which lifecycle response is most appropriate?

Intelligent systems and Artificial Intelligence lifecycle Hard
A. Replace all historical data with recent predictions because deployed outputs are reliable labels
B. Inspect subgroup drift and error costs, then retrain or recalibrate under updated governance checks
C. Remove monitoring because stable aggregate accuracy demonstrates that the model remains valid
D. Increase the classification threshold globally because aggregate accuracy conceals model complexity

48 A cleaning robot observes only its current camera image, while the same image may correspond to different room locations requiring different actions. What change most directly improves the robot's decision state?

Intelligent systems and Artificial Intelligence lifecycle Hard
A. Use the current image as a complete state because identical observations imply identical situations
B. Replace the performance measure with image resolution so perception becomes the primary objective
C. Store relevant observation and action history to maintain a belief about the hidden location
D. Increase actuator speed so that uncertainty is resolved through faster physical movement

49 A disease-screening model is deployed where prevalence is far lower than in its training hospital. Its sensitivity and specificity remain approximately unchanged. Which quantity is most likely to decrease substantially?

Applications of Artificial Intelligence in healthcare, agriculture, finance, education, manufacturing, robotics and smart cities Hard
A. The model's sensitivity among patients who truly have the disease
B. The model's specificity among patients who do not have the disease
C. The positive predictive value among patients receiving positive results
D. The negative predictive value among patients receiving negative results

50 An agricultural irrigation agent receives noisy moisture readings, cannot directly observe underground water distribution, and sees crop response only after several days. Which formulation best captures the task?

Applications of Artificial Intelligence in healthcare, agriculture, finance, education, manufacturing, robotics and smart cities Hard
A. A supervised classification problem where each sensor reading has an immediate correct action label
B. A fully observable deterministic search problem with irrigation volume as the only state
C. A static constraint problem because irrigation actions do not alter future environmental conditions
D. A partially observable sequential decision problem with delayed effects and uncertain transitions

51 In fraud detection, missing a fraudulent transfer costs , investigating a legitimate transfer costs , and predicted probabilities are calibrated. Ignoring other costs, at what fraud probability should a transaction be investigated?

Applications of Artificial Intelligence in healthcare, agriculture, finance, education, manufacturing, robotics and smart cities Hard
A. Investigate when
B. Investigate when
C. Investigate when
D. Investigate when

52 An educational model accurately predicts which students will fail. The institution wants to identify students whose outcomes would improve because of tutoring. Why is ranking only by predicted failure risk insufficient?

Applications of Artificial Intelligence in healthcare, agriculture, finance, education, manufacturing, robotics and smart cities Hard
A. Tutoring allocation cannot use historical data because educational outcomes are always nonstationary
B. High-risk students may not be those with the largest causal response to the tutoring intervention
C. Failure prediction estimates uncertainty, whereas tutoring decisions require deterministic student labels
D. Failure risk and tutoring benefit are identical only when the prediction model has low variance

53 A predictive-maintenance system can replace a component now for , while failure before the next inspection costs . If the estimated failure probability is calibrated and replacement eliminates that risk, what is the cost-minimizing policy?

Applications of Artificial Intelligence in healthcare, agriculture, finance, education, manufacturing, robotics and smart cities Hard
A. Replace when the failure probability exceeds
B. Replace when the failure probability exceeds
C. Replace when the failure probability exceeds
D. Replace when the failure probability exceeds

54 Independent traffic-signal agents minimize delay only at their own intersections. Each learns to release vehicles toward already congested downstream roads, increasing network-wide gridlock. Which modification best addresses the failure?

Applications of Artificial Intelligence in healthcare, agriculture, finance, education, manufacturing, robotics and smart cities Hard
A. Give each agent a network-aware reward and include downstream congestion in its observable state
B. Reduce every agent's observation to its local queue so state representations remain independent
C. Remove communication and randomize signal phases so agents cannot converge to similar policies
D. Optimize each intersection more frequently using the same local delay reward and local sensors

55 A route planner defines a state only as the current city, but fuel prices vary and the vehicle's remaining fuel affects which roads are feasible. What is the fundamental flaw?

Artificial Intelligence problem formulation and state-space representation Hard
A. The state contains too many variables, causing equivalent routes to become separate solutions
B. The transition model should ignore road feasibility to keep the search graph fully connected
C. The goal test must be removed because fuel constraints already define the destination implicitly
D. The state omits fuel level, so identical represented states can have different valid successors

56 In a block-stacking problem, two search nodes contain the same physical arrangement but were reached through different action sequences. Under the standard state-space formulation, how should they be treated?

Artificial Intelligence problem formulation and state-space representation Hard
A. As different goals because path histories determine whether an arrangement satisfies constraints
B. As different states because a state is defined by the complete sequence of preceding actions
C. As the same solution, even if neither arrangement satisfies the specified goal condition
D. As the same state, while retaining path information separately when path cost is relevant

57 An agent knows all possible outcomes of each action, but an action may lead to several outcomes chosen unpredictably. After every action, the resulting state becomes fully observable. Which approach is most appropriate?

Problem characteristics Hard
A. Unsupervised clustering because successor states have no externally supplied class labels
B. Binary classification because each action can be evaluated independently of future outcomes
C. AND-OR search because a valid plan must handle every possible action outcome
D. Linear single-path search because complete observability makes every action deterministic

58 A constraint satisfaction problem has variables, each with domain size . Before constraints are applied, there are complete assignments, but only satisfy all constraints. Which statement is correct?

Search space and solution space Hard
A. The search space may include partial and invalid assignments, while the solution space contains the valid assignments
B. The search and solution spaces are identical whenever every variable uses the same domain size
C. The solution space has size , while the search space contains only partial assignments
D. The search space contains only the satisfying assignments because invalid assignments are never states

59 Two medical triage agents have equal accuracy. Agent X has more false negatives, while Agent Y has more false positives. Missing an emergency is far costlier than unnecessary review. Which evaluation is most defensible?

Performance measures Hard
A. Prefer Agent Y if its cost-weighted expected loss is lower under the deployment prevalence
B. Treat the agents as equivalent because performance measures cannot encode asymmetric consequences
C. Prefer Agent X because false-positive counts should dominate all safety-critical evaluations
D. Prefer Agent X because equal accuracy proves both agents have equal expected clinical utility

60 Graph-search A* uses an admissible but inconsistent heuristic and permanently closes expanded states without reopening them when cheaper paths are found. Which guarantee can fail?

Overview of Artificial Intelligence-based problem-solving approaches Hard
A. Soundness can fail because A* may return a path that does not reach any goal state
B. Completeness can fail only because admissibility requires negative edge costs
C. Termination can fail because inconsistent heuristics always assign infinite values to goals
D. Optimality can fail because a cheaper path to a previously closed state may appear later