ML/AI Flashcards
1
Q
What is the primary inspiration for the GWO algorithm?
a) Bird migration patterns
b) Social behavior of wolves
c) Genetic structures
d) Firefly communication
A
Answer: b
2
Q
In GWO, what does the prey represent in the optimization process?
a) The worst solution
b) The optimal solution
c) The alpha wolf
d) The average solution
A
Answer: b
3
Q
What is the significance of the social hierarchy in a wolf pack for the GWO algorithm?
a) It determines the pack’s hunting strategy
b) It establishes dominance for all wolves
c) It guides the optimization process
d) It influences the color of wolves
A
Answer: c
4
Q
In Grey Wolf Optimization the Alpha wolf
a) Ranks first and represents the best solution
b) Ranks second and assists the alpha by exploring the solution space
c) Ranks third and independently explores the solution space
d) None of the above
A
Answer: a
5
Q
In Grey Wolf Optimization the Beta wolf
a) Ranks first and represents the best solution
b) Ranks second and assists the alpha by exploring the solution space
c) Ranks third and independently explores the solution space
d) None of the above
A
Answer: b
6
Q
In Grey Wolf Optimization the Delta wolf
a) Ranks first and represents the best solution
b) Ranks second and assists the alpha by exploring the solution space
c) Ranks third and independently explores the solution space
d) None of the above
A
Answer: c
7
Q
In Grey Wolf Optimization the Omega wolf
a) Ranks first and represents the best solution
b) Ranks second and assists the alpha by exploring the solution space
c) Ranks third and independently explores the solution space
d) None of the above
A
Answer: d
8
Q
What is the time complexity of the GWO algorithm approximately expressed as?
a) O(n^2)
b) O(k+n)
c) O(k*n)
d) O(log n)
A
Answer: O(k*n)
9
Q
How is the position of wolves updated in GWO?
a) Only based on alpha wolf
b) Average of the three best wolves
c) Randomly without considering other wolves
d) According to the prey’s position
A
Answer: b
10
Q
What is the main application for GWO in machine learning?
a) Image recognition
b) Hyperparameter tuning
c) Natural language processing
d) Clustering
A
Answer: b
11
Q
How is GWO used in hyperparameter tuning?
a) By optimizing learning rate, number of trees, and maximum depth
b) By selecting the fastest algorithm
c) By adjusting image resolution
d) By tuning neural network weights
A
Answer: a
12
Q
What is a more feasible representation for feature selection in GWO?
a) Numeric values for each feature
b) A binary vector representing included features
c) Randomly selected features
d) A matrix of feature correlations
A
Answer: b
13
Q
What does GWO offer as a cost-effective solution for feature selection?
a) Exponential complexity
b) Linear complexity
c) Binary complexity
d) Hierarchical complexity
A
Answer: b
14
Q
In GWO, why is a small random shift added to the position update of wolves?
a) To confuse the prey
b) To avoid getting stuck in local optima
c) To increase dominance
d) To speed up convergence
A
Answer: b
15
Q
What does the gathering of omega wolves around the best three wolves indicate in GWO?
a) The end of the optimization process
b) Increased randomness
c) Failure of the algorithm
d) Lack of hierarchy
A
Answer: a
16
Q
What type of algorithm is SSA?
a) Genetic Algorithm
b) Population-based Algorithm
c) Neural Network
d) Decision Tree
A
Answer: b
17
Q
What does SSA mimic in its design?
a) Birds
b) Fish
c) Salp Swarms
d) Bees
A
Answer: c
18
Q
What is the natural behavior of salp chains in the depths of the oceans?
a) They move by flying in the air
b) They move by pushing water to find food
c) They move by crawling on the ocean floor
d) They move by floating on the water surface
A
Answer: b
19
Q
How long can the chains of salp swarms be?
a) Up to 2 feet
b) Up to 5 feet
c) Up to 10 feet
d) Up to 15 feet
A
Answer: d
20
Q
What is the main reason for the behavior of salp chains?
a) To avoid predators
b) To achieve better movement for finding food
c) To migrate to different oceans
d) To reproduce more efficiently
A
Answer: b
21
Q
How is the leader-follower structure mathematically described in the salp swarm?
a) The head salp is the follower, and others are leaders
b) The head salp is the leader, and others are followers
c) All salps in the chain are leaders
d) All salps in the chain are followers
A
Answer: b
22
Q
What is the main characteristic of the Salp Swarm Algorithm (SSA)?
a) It is a deterministic algorithm
b) It is a random (stochastic) algorithm
c) It is a rule-based algorithm
d) It is a supervised learning algorithm
A
Answer: b
23
Q
In the Salp Swarm Algorithm, what does the exploration stage focus on?
a) Finding better solutions
b) Exploiting local data to improve the current solution
c) Balancing between exploration and exploitation
d) Updating the position of the leader salp
A
Answer: a
24
Q
What is the purpose of the parameter initialization step in the Salp Swarm Algorithm?
a) It initializes the food source position
b) It initializes the population size and other parameters
c) It updates the position of the leader
d) It terminates the algorithm
A
Answer: b
25
How are individuals in the initial population generated in the Salp Swarm Algorithm?
a) By cloning the leader's position
b) By creating random solutions within a specified range
c) By following a deterministic rule
d) By selecting the best solution from a previous iteration
Answer: b
26
What happens if any solution violates the range of the search space during the update process in the Salp Swarm Algorithm?
a) It is discarded
b) It is returned back within the range of the problem
c) It becomes the new leader
d) It terminates the algorithm
Answer: b
27
What is the termination criteria in the Salp Swarm Algorithm?
a) The number of followers
b) The maximum number of iterations
c) The size of the search space
d) The quality of the leader's position
Answer: b
28
What is the primary inspiration for the Whale Optimization Algorithm?
a) Blue whales
b) Humpback whales
c) Beluga whales
d) Sperm whales
Answer: b
29
What are the two main strategies employed by the Whale Optimization Algorithm?
a) Expansion and Contraction
b) Exploration and Exploitation
c) Encircling and Attacking
d) Searching and Sorting
Answer: b
30
During the encircling phase of the Whale Optimization Algorithm, what do humpback whales consider as the best-obtained and near-optimal solution?
a) Random candidate solution
b) Current best-candidate solution
c) Average solution
d) Worst solution
Answer: b
31
What is the mechanism used by humpback whales during the exploitation phase of the Whale Optimization Algorithm?
a) Spiral mechanism
b) Encircling mechanism
c) Bubble-net mechanism
d) Linear mechanism
Answer: c
32
What is the initial step in the Whale Optimization Algorithm after randomly initializing the whale population?
a) Encircling prey
b) Evaluating fitness values
c) Updating positions
d) Searching for optimal solutions
Answer: b
33
How is the position of the whale updated during the exploitation phase of the Whale Optimization Algorithm?
a) Linearly
b) Randomly
c) Spirally
d) Exponentially
Answer: c
34
Which phase of the Whale Optimization Algorithm involves attacking prey using the bubble-net mechanism?
a) Exploration phase
b) Exploitation phase
c) Encircling phase
d) Searching phase
Answer: b
35
What is the purpose of adjusting the population size in Whale Optimization?
a) To maximize exploration
b) To minimize solution quality
c) To prevent premature convergence
d) To ignore problem-specific knowledge
Answer: c
36
What is a challenge faced by Whale Optimization in complex landscapes?
a) Lack of computational complexity
b) Limited exploration capabilities
c) Efficient exploitation of local regions
d) Lack of problem-specific guidance
Answer: b
37
What is the purpose of performing sensitivity analysis on Whale Optimization parameters?
a) To ignore optimal settings
b) To identify optimal settings for specific problems
c) To increase population diversity
d) To eliminate problem-specific knowledge
Answer: b
38
The Impact of Whale Optimization is measured based on which metric/s?
a) Solution Quality
b) Convergence Speed
c) Resource Utilization
d) All of the mentioned
Answer: d
39
How do some cuckoos reduce the probability of their eggs being abandoned?
a) By imitating the colors and patterns of host eggs
b) By building stronger nests
c) By laying fewer eggs
d) By choosing nests randomly
Answer: a
40
What happens if host birds discover that cuckoo eggs are not their own?
a) They throw the eggs away or abandon their nests and build new ones
b) They adopt the cuckoo eggs as their own
c) They protect the cuckoo eggs from predators
d) They throw away their own eggs
Answer: a
41
What action does a cuckoo chick take once it is hatched?
a) evict the eggs out of the nest
b) imitate the call of host chicks
c) both a and b
d) co-exist with the host eggs
Answer: c
42
In the Cuckoo Search Algorithm, what does each egg in a nest represent?
a) A mathematical equation
b) A new solution
c) The fitness of a solution
d) The best solution
Answer: b
43
In the Cuckoo Search Algorithm each nest can only contain 1 egg.
a) True
b) False
Answer: b
44
According to the Cuckoo Search Algorithm, what happens to the best nests with high-quality eggs?
a) They are abandoned
b) They are discarded
c) They carry over to the next generation
d) They are replaced immediately
Answer: c
45
What is the significance of Levy flight in the Cuckoo Search Algorithm?
a) It represents the flight patterns of cuckoos
b) It is used to calculate fitness
c) It generates a new cuckoo solution
d) It determines the number of available hosts
Answer: c
46
How is the termination criterion in the Cuckoo Search Algorithm defined?
a) By the number of cuckoos in a nest
b) By the fitness of the eggs
c) By the number of iterations
d) By the size of the nests
Answer: c
47
In what applications can the Cuckoo Search Algorithm be used?
a) Train neural network
b) Engineering optimization problems
c) both a and b
d) Culinary arts
Answer: c
48
How is the Cuckoo Search Algorithm extended in more complicated cases?
a) By decreasing the number of cuckoos in a nest
b) By introducing multiple eggs in each nest
c) By changing the color patterns of cuckoo eggs
d) By reducing the number of iterations
Answer: b
49
How is the fitness of a cuckoo's egg compared to the fitness of the host egg in the Cuckoo Search Algorithm?
a) By their size
b) By the number of iterations
c) By the color patterns
d) By using the fitness function
Answer: d
50
What action is taken if the fitness of cuckoo’s eggs is better than the host egg in the Cuckoo Search Algorithm?
a) The cuckoo's egg is discarded
b) The host egg is abandoned
c) The host egg is replaced by the cuckoo’s egg
d) A new nest is built immediately
Answer: c
51
What is the purpose of abandoning nests in the Cuckoo Search Algorithm?
a) To reduce the number of solutions
b) To increase local optimization
c) To build new nests with better solutions
d) To protect cuckoo eggs from predators
Answer: c
52
How are nests ranked in the Cuckoo Search Algorithm?
a) Based on the number of cuckoos in each nest
b) Randomly
c) According to the fitness of solutions
d) Based on the size of each egg
Answer: c
53
What are the two main behaviors that inspire firefly optimization?
a) Hovering and Pulsing
b) Flashing and Gliding
c) Fluttering and Glowing
d) Flashing and Lévy flight
Answer: d
54
What purpose did the ability to light up originally serve for fireflies?
a) Attracting mates
b) Warding off predators
c) Finding food
d) Establishing territory
Answer: b
55
What is the significance of the flashing behavior of fireflies?
a) It signals danger
b) It communicates with other species
c) It attracts mates
d) It scares off predators
Answer: c
56
How do female fireflies of a species respond to the flashing pattern of males?
a) They ignore it
b) They flee from it
c) They respond based on the pattern
d) They attack the males
Answer: c
57
What is the main characteristic observed in the flashing pattern in fireflies?
a) It is the same for all species
b) It is unique for each species
c) It is random
d) It is predictable
Answer: b
58
Why are fireflies considered unisex in the context of the firefly algorithm?
a) They lack distinct sexes
b) They can be attracted regardless of sex
c) They have both male and female characteristics
d) They change sexes over time
Answer: b
59
What determines the brightness of a firefly in the firefly algorithm?
a) The time of day
b) The landscape of the objective function
c) The type of vegetation
d) The distance to other fireflies
Answer: b
60
How does the firefly algorithm handle the movement of fireflies?
a) Randomly
b) Based on a fixed pattern
c) Towards the more bright firefly
d) Away from other fireflies
Answer: c
61
How does a firefly move in the Firefly Algorithm if there is no brighter firefly?
a) It doesn’t move
b) It moves towards the dimmest firefly
c) It moves randomly
d) It moves towards the light source
Answer: c
62
What is the role of attractiveness in the firefly algorithm?
a) It increases randomness
b) It determines movement
c) It decreases brightness
d) It affects the flashing pattern
Answer: b
63
What law governs the variation in light intensity in the firefly algorithm?
a) Direct square law
b) Inverse square law
c) Square root law
d) Direct proportional law
Answer: b
64
What is the impact of the distance between fireflies on attractiveness?
a) It increases attractiveness
b) It decreases attractiveness
c) It has no effect
d) It increases brightness
Answer: b
65
Which of the following is an application of the firefly algorithm?
a) Antenna design
b) Digital image compression
c) Semantic web composition
d) All of the above
Answer: d
66
Which of the following does the attractiveness of a firefly depend on?
a) The color of the firefly
b) The speed of the firefly
c) The intensity of the light the light the firefly emits.
d) None of the above
Answer: c
67
What does supervised machine learning seek to approximate?
a) Input variables
b) Target function
c) Residual errors
d) Noise in training data
Answer: b
68
What does the equation Y = f(X) represent in supervised machine learning?
a) Input variables
b) Target function
c) Residual errors
d) Noise in training data
Answer: b
69
What is induction in the context of machine learning?
a) Learning specific concepts from general rules
b) Learning general concepts from specific examples
c) Learning the form of the target function
d) Learning noise in training data
Answer: b
70
What is the goal of a good machine learning model?
a) Overfitting
b) Underfitting
c) Generalization
d) Memorization
Answer: c
71
What term is used to describe a model that models the training data too well?
a) Underfitting
b) Generalization
c) Overfitting
d) Deduction
Answer: c
72
Why is overfitting more likely with nonparametric and nonlinear models?
a) They are flexible and have more constraints
b) They are less flexible and have fewer constraints
c) They are sensitive to noise in training data
d) They do not learn the target function
Answer: a
73
What is the remedy for underfitting in machine learning?
a) Pruning the tree
b) Moving on and trying alternate algorithms
c) Using cross-validation
d) Holding back a validation dataset
Answer: b
74
In machine learning, what does a good fit aim for?
a) Underfitting
b) Overfitting
c) The sweet spot between underfitting and overfitting
d) Generalization
Answer: c
75
What is the sweet spot in machine learning?
a) The point where the model has poor skill on the training dataset
b) The point just before the errors on the test dataset start to increase
c) The point where the model overfits the training data
d) The point where the error on the test dataset is at its lowest
Answer: b
76
What is the purpose of using k-fold cross-validation in machine learning?
a) To estimate model accuracy on unseen data
b) To train and test the model on different subsets of training data
c) To hold back a validation dataset
d) To limit underfitting
Answer: b
77
What is a validation dataset in machine learning?
a) A dataset used for training the model
b) A subset of the training data held back until the end of the project
c) A dataset used to estimate model accuracy
d) A dataset used for pruning decision trees
Answer: b
78
What is the most common problem in applied machine learning?
a) Underfitting
b) Overfitting
c) Generalization
d) Deduction
Answer: b
79
What is the purpose of using resampling techniques in machine learning?
a) To estimate model accuracy on unseen data
b) To train and test the model on different subsets of training data
c) To hold back a validation dataset
d) To limit overfitting
Answer: a
80
What is the gold standard in applied machine learning for estimating model accuracy on unseen data?
a) Overfitting
b) Validation dataset
c) Resampling techniques
d) Cross-validation
Answer: d
81
What is the drawback of choosing the stopping point for training using the skill on the test dataset?
a) It leads to overfitting
b) It leaks knowledge about the test dataset into the training procedure
c) It improves model accuracy
d) It eliminates the need for validation datasets
Answer: b
82
What is the purpose of using a validation dataset in machine learning?
a) To estimate model accuracy on unseen data
b) To train and test the model on different subsets of training data
c) To hold back a subset of training data until the end of the project
d) To improve the model's ability to generalize
Answer: a
83
Why is overfitting a problem in machine learning?
a) It improves model accuracy
b) It negatively impacts the model's ability to generalize
c) It eliminates the need for validation datasets
d) It leads to underfitting
Answer: b
84
What characterizes a machine learning model that is underfit?
a) It models the training data too well
b) It has poor performance on the training data
c) It generalizes well to new data
d) It learns irrelevant detail and noise in the training dataset
Answer: b
85
What is the goal of a good machine learning model in terms of generalization?
a) To memorize the training data
b) To overfit the training data
c) To generalize well to any data from the problem domain
d) To underfit the training data
Answer: c
86
Why is generalization important in machine learning?
a) Because the data we collect is complete and noise-free
b) Because the data we collect is only a sample, incomplete, and noisy
c) Because machine learning algorithms always know the form of the target function
d) Because induction is not required in machine learning
Answer: b
87
What is the purpose of pruning a tree in machine learning?
a) To eliminate underfitting
b) To improve model accuracy
c) To address overfitting in decision trees
d) To memorize the training data
Answer: c
88
What is the primary purpose of metaheuristics?
a) Solving only combinatorial problems
b) Focusing solely on continuous optimization
c) General-purpose algorithms for various optimization problems
d) None of the above
Answer: c
89
What is the common inspiration source for metaheuristics?
a) Artificial intelligence
b) Natural phenomena
c) Machine learning
d) Database systems
Answer: b
90
Which strategy aims to preserve the diversity of solutions to avoid premature convergence?
a) Intensification
b) Diversification
c) Adaptive control
d) Randomization
Answer: b
91
How does diversity maintenance in metaheuristics achieve its goal?
a) Through crossover operations
b) By mutation and perturbation
c) Using adaptive control mechanisms
d) Via selection, replacement, or diversity operators
Answer: d
92
What does intensification in metaheuristics focus on?
a) Generating new solutions
b) Hill climbing and crossover
c) Preserving diversity
d) Mutation and perturbation
Answer: c
93
In metaheuristic search, what does diversification primarily involve?
a) Improving solution quality
b) Local search and hill climbing
c) Generating new solutions
d) Adaptive control mechanisms
Answer: c
94
What is adaptive control in the context of metaheuristics?
a) Preserving diversity
b) Dynamic adjustment based on feedback
c) Fixed number of iterations
d) Convergence threshold
Answer: b
95
What does termination criteria in metaheuristics influence?
a) Operator probabilities
b) Initial population
c) Trade-off between exploration and exploitation
d) Termination potential
Answer: c
96
What is the purpose of adaptive control mechanisms in metaheuristics?
a) Maintaining diversity
b) Fixing the number of iterations
c) Adjusting parameters dynamically
d) Enhancing convergence
Answer: c
97
What impact does a fixed number of iterations have on exploration potential?
a) Enhances exploration
b) Limits exploration
c) Improves convergence
d) Increases diversity
Answer: b
98
In metaheuristics, what may limit exploitation potential?
a) Adaptive control
b) Convergence threshold
c) Modality of the problem
d) Improvement threshold
Answer: b
99
What is the significance of randomness in metaheuristics?
a) Limits exploration potential
b) Enhances convergence
c) Preserves diversity
d) Reduces exploitation potential
Answer: c
100
Which of the following is true for an exploration problem?
a) State and actions are unknown to the agent
b) State and actions are known to the agent
c) Only actions are known to the agent
d) None of the above
Answer: a
101
________ involves choosing the best-known option based on past experiences, while ________ involves trying out new options that may lead to better outcomes in the future.
a) Optimization - Utilization
b) Convergence - Divergence
c) Exploration - Exploitation
d) Exploitation - Exploration
Answer: d
102
What is the process of simultaneously optimizing multiple conflicting objectives in a given problem?
a) Single-Objective Optimization
b) Bi-Objective Optimization
c) Multi-Objective Optimization
d) None of the above
Answer: c
103
What is the name of the strategy that preserves the diversity of solutions in the search population or memory?
a) Diversity maintenance
b) Intensification and diversification
c) Adaptive control
d) Local search
Answer: a
104
What is the name of the strategy that adapts the parameters or components of the metaheuristic according to feedback from the search process?
a) Diversity maintenance
b) Intensification and diversification
c) Adaptive control
d) Local search
Answer: c
105
What is the name of the criterion that can be used to adapt the parameters or components of the metaheuristic based on the quality of the solutions?
a) Performance
b) Diversity
c) Learning
d) All of the above
Answer: d
106
What is a characteristic of a rugged fitness landscape?
a) Many local peaks surrounded by deep valleys
b) All genotypes have the same replication rate
c) A flat and smooth surface
d) Maze-like properties
Answer: a
107
How does the concept of a fitness landscape apply to dynamic optimization?
a) By visualizing landscapes formed by expected fitness at each point
b) By using scalar-valued functions
c) By considering the changing environment
d) By introducing a fitness function
Answer: a
108
What is a relevant consideration regarding fitness landscapes being a relative function?
a) Fitness landscapes are always absolute
b) Fitness landscapes are multidimensional
c) Fitness landscapes are static in time
d) Fitness landscapes are relative, not absolute
Answer: d
109
What is the fundamental possibility with regard to measuring fitness landscapes?
a) Visualizing all dimensions
b) Visualizing all landscapes simultaneously
c) Measuring parameters of landscape ruggedness
d) Ignoring all limitations
Answer: c
110
How does the human mind struggle when thinking about highly multi-dimensional fitness landscapes?
a) It visualizes all dimensions easily
b) It does not struggle with multi-dimensional thinking
c) It can mislead when discussing highly multi-dimensional landscapes
d) It understands all dimensions simultaneously
Answer: c
111
What is the role of stochastic sampling in cases where a fitness function is hard to define?
a) It is used to create flat fitness landscapes
b) It samples from a known distribution
c) It is irrelevant to fitness landscapes
d) It samples from an unknown distribution at each point
Answer: d
112
What is the global maximum of a function?
a) The highest value in the entire domain
b) The highest value in a specific interval
c) The average value of the function
d) The derivative of the function
Answer: a
113
Which of the following statements is true regarding global minima?
a) They always occur at critical points
b) They can occur at endpoints of the domain
c) They are only found in continuous functions
d) They occur when the derivative is zero
Answer: b
114
In optimization problems, what is a local minimum?
a) The smallest value in a specific interval
b) The smallest value in the entire domain
c) A value that is smaller than the global minimum
d) A value that is not relevant to optimization
Answer: a
115
Which of the following functions is guaranteed to have a global minimum?
a) Polynomial functions of odd degree
b) Exponential functions
c) Trigonometric functions
d) Polynomial functions of even degree
Answer: d
116
When does a function have no global maximum or minimum?
a) When the function is continuous
b) When the function is differentiable
c) When the function is unbounded
d) When the function has critical points
Answer: c
117
Which optimization method is based on dividing the search space into subspaces?
a) Bisection method
b) Quasi-Newton method
c) Divide-and-conquer algorithm
d) Successive parabolic interpolation
Answer: a
118
What is the role of a cost function in optimization?
a) To maximize the objective function
b) To minimize the objective function
c) To randomize the search space
d) To introduce noise into the optimization process
Answer: b
119
What is the primary purpose of a heuristic function in optimization algorithms?
a) To guarantee convergence to the global minimum
b) To guide the search towards promising regions
c) To introduce randomness into the optimization process
d) To slow down the optimization process
Answer: b
120
What class does P represent in the context of computational complexity?
a) Polynomial time
b) Exponential time
c) Non-deterministic polynomial time
d) Exponential hierarchy
Answer: a
121
What does NP stand for in computational complexity?
a) Non-polynomial time
b) Non-deterministic polynomial time
c) Non-prime
d) New problem
Answer: b
122
What is the primary concern of the P vs NP problem?
a) Polynomial time algorithms
b) Non-polynomial time algorithms
c) Parallel algorithms
d) Probabilistic algorithms
Answer: b
123
The P vs NP problem asks whether:
a) P is equal to NP
b) P is a subset of NP
c) NP is a subset of P
d) P and NP are disjoint sets
Answer: a
124
The existence of which type of algorithm would prove P equals NP?
a) Polynomial-time algorithm
b) Exponential-time algorithm
c) Non-deterministic polynomial-time algorithm
d) Probabilistic polynomial-time algorithm
Answer: c
125
Which famous problem is known to be NP-complete and is often used in reductions?
a) Traveling Salesman Problem
b) Sorting Problem
c) Binary Search Problem
d) Fibonacci Sequence Problem
Answer: a
126
What is a common method used to show NP-completeness in reductions?
a) Cook-Levin Theorem
b) Euclidean Algorithm
c) Dijkstra's Algorithm
d) Floyd-Warshall Algorithm
Answer: a
127
If P equals NP, what would be the consequence for optimization problems?
a) Easier optimization problems
b) Harder optimization problems
c) No impact on optimization problems
d) Unsolvable optimization problems
Answer: c
128
If P equals NP, what would be the implication for cryptography?
a) Stronger encryption algorithms
b) Breakthrough in quantum cryptography
c) Infeasibility of secure encryption
d) No impact on cryptography
Answer: c
129
What is Machine learning?
a) The autonomous acquisition of knowledge through the use of computer programs
b) The autonomous acquisition of knowledge through the use of manual programs
c) The selective acquisition of knowledge through the use of computer programs
d) The selective acquisition of knowledge through the use of manual programs
Answer: a
130
What is optimization?
a) The process of finding the best solution while adhering to constraints
b) The process of finding the best solution while ignoring all constraints
c) Both a and b
d) None of the mentioned
Answer: a
131
In the context of AI, what is the term "bias" referring to?
a) Ethical considerations in AI development
b) Systematic and unfair prejudice in machine learning models
c) A type of algorithm
d) The speed of computation
Answer: b
132
Which of the following is a subfield of AI focused on teaching machines to learn from data?
a) Robotics
b) Machine Learning
c) Natural Language Processing
d) Expert Systems
Answer: b
133
In AI, what does the acronym NLP stand for?
a) Natural Language Processing
b) Neural Learning Platform
c) Networked Language Programming
d) Nonlinear Pattern Recognition
Answer: a
134
Which AI application involves machines understanding and generating human language?
a) Computer Vision
b) Speech Recognition
c) Natural Language Processing
d) Robotics
Answer: c
135
What is the primary purpose of computer vision in AI?
a) Recognizing and interpreting visual information
b) Generating random images
c) Enhancing computer graphics
d) Simulating human vision
Answer: a
136
What type of AI system is designed to mimic human decision-making abilities?
a) Expert System
b) Neural Network
c) Genetic Algorithm
d) Fuzzy Logic System
Answer: a
137
What is the role of reinforcement learning in AI?
a) Teaching machines to perform specific tasks without feedback
b) Learning from labeled training data
c) Learning from trial and error with feedback
d) Analyzing patterns in large datasets
Answer: c
138
Which AI application involves machines understanding and responding to human speech?
a) Speech Recognition
b) Natural Language Processing
c) Computer Vision
d) Genetic Algorithms
Answer: a
139
What is the primary goal of swarm intelligence in AI?
a) Mimicking the behavior of insect swarms
b) Enhancing communication between machines
c) Achieving collective problem-solving through decentralized systems
d) Simulating human social interactions
Answer: c
140
What is the role of fuzzy logic in AI?
a) Handling uncertainty and imprecision in decision-making
b) Creating clear and precise algorithms
c) Enhancing logical reasoning in machines
d) Simulating human emotions
Answer: a
141
Which AI application involves machines imitating human-like movements and actions?
a) Robotics
b) Natural Language Processing
c) Genetic Algorithms
d) Expert Systems
Answer: a
142
What is the primary focus of expert systems in AI?
a) Learning from experience
b) Replicating human intuition
c) Solving optimization problems
d) Mimicking human expertise in a specific domain
Answer: d
143
Which AI application involves machines understanding and interpreting visual information from the world?
a) Speech Recognition
b) Natural Language Processing
c) Computer Vision
d) Expert Systems
Answer: c
144
What does the term "unsupervised learning" refer to in machine learning?
a) Learning from labeled training data
b) Learning from trial and error with feedback
c) Learning without explicit supervision or labeled outputs
d) Learning from expert systems
Answer: c
145
Among the following, which is not a type of learning?
a) Unsupervised learning
b) Reinforcement learning
c) Supervised learning
d) Semi-unsupervised learning
Answer: d
146
Identify the type of learning in which labeled training data is used.
a) Unsupervised learning
b) Reinforcement learning
c) Supervised learning
d) Semi-unsupervised learning
Answer: c
147
What technology, prevalent in recent AI, enables machines to make decisions without explicit programming?
a) Expert systems
b) Reinforcement learning
c) Genetic algorithms
d) Symbolic reasoning
Answer: b
148
What aspect of recent AI contributes to the ability to generalize knowledge across different tasks?
a) Fixed architectures
b) Narrow specialization
c) Transfer learning
d) Symbolic reasoning
Answer: c
149
In contrast to traditional AI, what is a characteristic of recent AI in handling incomplete or noisy data?
a) Rule-based reasoning
b) Genetic algorithms
c) Probabilistic modeling
d) Expert systems
Answer: c
150
What is a distinguishing feature of recent AI in terms of adaptability to changing environments?
a) Fixed rule sets
b) Limited learning capabilities
c) Rule-based programming
d) Continuous learning mechanisms
Answer: d
151
Which machine learning algorithm is commonly used for binary classification tasks and is based on a logistic function?
a) Decision trees
b) Support Vector Machines (SVM)
c) K-Nearest Neighbors (KNN)
d) Logistic regression
Answer: d
152
In machine learning, what is a hyperparameter?
a) A parameter learned during training
b) A feature in the dataset
c) A parameter set by the machine learning engineer before training
d) An output of the model
Answer: c
153
In reinforcement learning, what is the role of an agent?
a) To label the training data
b) To make predictions on new data
c) To interact with the environment and learn from feedback
d) To cluster data points
Answer: c
154
155
