Q#1: Probabilistic models in AI represent:
(A) Uncertainty in data and relationships
(B) Deterministic outcomes only
(C) BFS only
(D) DFS only
Answer: (A) Uncertainty in data and relationships
Q#2: A Bayesian network is:
(A) A directed acyclic graph representing dependencies among variables
(B) A decision tree only
(C) BFS only
(D) DFS only
Answer: (A) A directed acyclic graph representing dependencies among variables
Q#3: Nodes in a Bayesian network represent:
(A) Random variables
(B) Deterministic actions
(C) BFS only
(D) DFS only
Answer: (A) Random variables
Q#4: Edges in a Bayesian network represent:
(A) Conditional dependencies between variables
(B) Independent variables only
(C) BFS only
(D) DFS only
Answer: (A) Conditional dependencies between variables
Q#5: Conditional probability tables (CPTs) specify:
(A) Probabilities of each node given its parents
(B) BFS only
(C) DFS only
(D) Random guesses
Answer: (A) Probabilities of each node given its parents
Q#6: Learning probabilistic models can involve:
(A) Estimating parameters from data
(B) BFS only
(C) DFS only
(D) Random guesses
Answer: (A) Estimating parameters from data
Q#7: Parameter learning assumes:
(A) Known structure of the probabilistic model
(B) Random structure
(C) BFS only
(D) DFS only
Answer: (A) Known structure of the probabilistic model
Q#8: Structure learning aims to:
(A) Learn the dependency structure from data
(B) BFS only
(C) DFS only
(D) Random guesses
Answer: (A) Learn the dependency structure from data
Q#9: Maximum likelihood estimation (MLE) is used to:
(A) Estimate parameters that maximize probability of observed data
(B) BFS only
(C) DFS only
(D) Random guesses
Answer: (A) Estimate parameters that maximize probability of observed data
Q#10: Bayesian parameter estimation incorporates:
(A) Prior knowledge and observed data
(B) BFS only
(C) DFS only
(D) Random guesses
Answer: (A) Prior knowledge and observed data
Q#11: Hidden variables are:
(A) Variables not directly observed in data
(B) BFS only
(C) DFS only
(D) Observed variables only
Answer: (A) Variables not directly observed in data
Q#12: EM (Expectation-Maximization) algorithm is used for:
(A) Learning parameters with hidden variables
(B) BFS only
(C) DFS only
(D) Random guesses
Answer: (A) Learning parameters with hidden variables
Q#13: In EM, the E-step computes:
(A) Expected values of hidden variables
(B) Maximizes parameters directly
(C) BFS only
(D) DFS only
Answer: (A) Expected values of hidden variables
Q#14: In EM, the M-step computes:
(A) Parameters that maximize expected log-likelihood
(B) Expected values of hidden variables
(C) BFS only
(D) DFS only
Answer: (A) Parameters that maximize expected log-likelihood
Q#15: Probabilistic graphical models include:
(A) Bayesian networks and Markov networks
(B) Decision trees only
(C) BFS only
(D) DFS only
Answer: (A) Bayesian networks and Markov networks
Q#16: Markov networks represent:
(A) Undirected dependencies among variables
(B) Directed acyclic graphs
(C) BFS only
(D) DFS only
Answer: (A) Undirected dependencies among variables
Q#17: Factor graphs are used to:
(A) Represent factorization of joint probability distributions
(B) BFS only
(C) DFS only
(D) Random guesses
Answer: (A) Represent factorization of joint probability distributions
Q#18: Conditional independence reduces:
(A) Complexity of probabilistic models
(B) BFS only
(C) DFS only
(D) Random guesses
Answer: (A) Complexity of probabilistic models
Q#19: Learning probabilistic models requires:
(A) Estimating structure and parameters from data
(B) BFS only
(C) DFS only
(D) Random guesses
Answer: (A) Estimating structure and parameters from data
Q#20: Naive Bayes classifier assumes:
(A) Conditional independence of features given class
(B) No independence assumption
(C) BFS only
(D) DFS only
Answer: (A) Conditional independence of features given class
Q#21: Naive Bayes is used for:
(A) Classification tasks
(B) Regression only
(C) BFS only
(D) DFS only
Answer: (A) Classification tasks
Q#22: Maximum a posteriori (MAP) estimation chooses:
(A) Parameter values that maximize posterior probability
(B) BFS only
(C) DFS only
(D) Random guesses
Answer: (A) Parameter values that maximize posterior probability
Q#23: Learning Bayesian networks from complete data involves:
(A) Counting occurrences of variable combinations
(B) BFS only
(C) DFS only
(D) Random guesses
Answer: (A) Counting occurrences of variable combinations
Q#24: Incomplete data requires:
(A) Algorithms like EM to handle missing values
(B) BFS only
(C) DFS only
(D) Random guesses
Answer: (A) Algorithms like EM to handle missing values
Q#25: Probabilistic inference computes:
(A) Probability of queries given observed evidence
(B) BFS only
(C) DFS only
(D) Random guesses
Answer: (A) Probability of queries given observed evidence
Q#26: Exact inference in Bayesian networks is:
(A) Often computationally expensive
(B) Always trivial
(C) BFS only
(D) DFS only
Answer: (A) Often computationally expensive
Q#27: Approximate inference methods include:
(A) Sampling and variational methods
(B) BFS only
(C) DFS only
(D) Random guesses
Answer: (A) Sampling and variational methods
Q#28: Gibbs sampling is:
(A) A Markov chain Monte Carlo method
(B) BFS only
(C) DFS only
(D) Random guesses
Answer: (A) A Markov chain Monte Carlo method
Q#29: Learning probabilistic relational models extends:
(A) Probabilistic models to relational data
(B) BFS only
(C) DFS only
(D) Random guesses
Answer: (A) Probabilistic models to relational data
Q#30: Hidden Markov Models (HMMs) are used for:
(A) Sequential data with hidden states
(B) Single-step classification only
(C) BFS only
(D) DFS only
Answer: (A) Sequential data with hidden states
Q#31: HMMs consist of:
(A) Hidden states, observed symbols, and transition/emission probabilities
(B) BFS only
(C) DFS only
(D) Random guesses
Answer: (A) Hidden states, observed symbols, and transition/emission probabilities
Q#32: Forward algorithm in HMMs computes:
(A) Probability of observed sequence
(B) Random guesses
(C) BFS only
(D) DFS only
Answer: (A) Probability of observed sequence
Q#33: Viterbi algorithm in HMMs finds:
(A) Most likely sequence of hidden states
(B) BFS only
(C) DFS only
(D) Random guesses
Answer: (A) Most likely sequence of hidden states
Q#34: Learning HMM parameters uses:
(A) EM algorithm or Baum-Welch algorithm
(B) BFS only
(C) DFS only
(D) Random guesses
Answer: (A) EM algorithm or Baum-Welch algorithm
Q#35: Probabilistic models can capture:
(A) Uncertainty in predictions
(B) Deterministic outcomes only
(C) BFS only
(D) DFS only
Answer: (A) Uncertainty in predictions
Q#36: Bayesian model averaging improves:
(A) Predictions by combining multiple models weighted by posterior probability
(B) BFS only
(C) DFS only
(D) Random guesses
Answer: (A) Predictions by combining multiple models weighted by posterior probability
Q#37: Learning graphical models involves:
(A) Structure learning and parameter estimation
(B) BFS only
(C) DFS only
(D) Random guesses
Answer: (A) Structure learning and parameter estimation
Q#38: Probabilistic reasoning supports:
(A) Decision-making under uncertainty
(B) Deterministic reasoning only
(C) BFS only
(D) DFS only
Answer: (A) Decision-making under uncertainty
Q#39: Probabilistic models are evaluated using:
(A) Likelihood, log-likelihood, and predictive accuracy
(B) BFS only
(C) DFS only
(D) Random guesses
Answer: (A) Likelihood, log-likelihood, and predictive accuracy
Q#40: Bayesian networks encode:
(A) Joint probability distributions compactly
(B) BFS only
(C) DFS only
(D) Random guesses
Answer: (A) Joint probability distributions compactly
Q#41: Directed graphical models are:
(A) Bayesian networks
(B) Markov networks
(C) BFS only
(D) DFS only
Answer: (A) Bayesian networks
Q#42: Undirected graphical models are:
(A) Markov networks
(B) Bayesian networks
(C) BFS only
(D) DFS only
Answer: (A) Markov networks
Q#43: Parameter tying reduces:
(A) Number of independent parameters in the model
(B) BFS only
(C) DFS only
(D) Random guesses
Answer: (A) Number of independent parameters in the model
Q#44: Learning with hidden variables can:
(A) Improve modeling of complex dependencies
(B) BFS only
(C) DFS only
(D) Random guesses
Answer: (A) Improve modeling of complex dependencies
Q#45: Knowledge in probabilistic models can be:
(A) Encoded as priors or constraints
(B) BFS only
(C) DFS only
(D) Random guesses
Answer: (A) Encoded as priors or constraints
Q#46: Bayesian parameter learning balances:
(A) Prior beliefs and observed data
(B) BFS only
(C) DFS only
(D) Random guesses
Answer: (A) Prior beliefs and observed data
Q#47: Probabilistic inference answers:
(A) Queries about variables given evidence
(B) BFS only
(C) DFS only
(D) Random guesses
Answer: (A) Queries about variables given evidence
Q#48: Exact inference is:
(A) Often intractable for large networks
(B) Always fast
(C) BFS only
(D) DFS only
Answer: (A) Often intractable for large networks
Q#49: Approximate inference includes:
(A) Sampling, loopy belief propagation, and variational methods
(B) BFS only
(C) DFS only
(D) Random guesses
Answer: (A) Sampling, loopy belief propagation, and variational methods
Q#50: The main goal of learning probabilistic models is:
(A) Model uncertainty and dependencies to make predictions under uncertainty
(B) Memorize data only
(C) BFS only
(D) DFS only
Answer: (A) Model uncertainty and dependencies to make predictions under uncertainty