MCQs

Disclaimer:

Educational Use Only:

The website links provided are to be used strictly for educational and knowledge-gaining purposes only.

No Misuse:

Do not misuse the links or the information contained therein for any commercial, illegal, or unethical activities.

Respect for Websites and Businesses:

Respect the intellectual property and business interests of the websites and their owners. Do not attempt to alter, copy, or redistribute content without proper authorization.

Acknowledge Value of Knowledge:

Recognize and value the knowledge and resources shared by these websites. Use the information responsibly and ethically.

No Endorsement:

The provision of these links does not imply endorsement of the linked sites, their content, or their products and services. Users are responsible for evaluating the suitability and accuracy of the information.

User Responsibility:

By accessing these links, you agree to use the information at your own risk and to comply with all applicable laws and website terms of use.

Overall MCQs for Reference

Which structure do Bayesian Networks use to represent variables and their conditional dependencies?
a. Undirected Graphs
b. Directed Acyclic Graphs
c. Complete Graphs
d. Bipartite Graphs
Answer: b. Directed Acyclic Graphsfiveable+2
In Markov Random Fields, edges represent:
a. Conditional independence
b. Direct probabilistic influence
c. Symmetric dependencies
d. Temporal dependencies
Answer: c. Symmetric dependenciesdzone+2
What property allows Probabilistic Graphical Models to reduce computational complexity?
a. Sparsity
b. Conditional independence
c. Maximum likelihood
d. Clustering
Answer: b. Conditional independencemcb111+2
The main advantage of using Probabilistic Graphical Models is:
a. Handling only deterministic data
b. Efficiently representing joint distributions
c. Preventing overfitting
d. Faster gradient descent
Answer: b. Efficiently representing joint distributionscse.iitm+2
A Hidden Markov Model specializes in modeling:
a. Spatial data
b. Temporal sequences
c. Outliers
d. Euclidean distances
Answer: b. Temporal sequencesoak.quizalize+2
In Probabilistic Graphical Models, a node typically represents:
a. A probability distribution
b. A random variable
c. An observation operator
d. A cost function
Answer: b. A random variableafrica.engineering.cmu+2
The local Markov property in an undirected graphical model states that a node is conditionally independent of:
a. Its parent nodes
b. All other nodes given its neighbors
c. Observed variables
d. Its ancestors
Answer: b. All other nodes given its neighborsfiveable+2
Which graphical model is best suited for causal reasoning?
a. Markov Random Field
b. Bayesian Network
c. Factor Graph
d. Conditional Random Field
Answer: b. Bayesian Networkcse.iitd+2
Factor graphs incorporate which feature in their structure?
a. Variables only
b. Both variables and factor nodes
c. Only factors
d. Directed cycles
Answer: b. Both variables and factor nodesarxiv+2
What is the principal use of inference in Probabilistic Graphical Models?
a. Data compression
b. Parameter tuning
c. Calculating probabilities of unknowns given knowns
d. Preprocessing data
Answer: c. Calculating probabilities of unknowns given knownsarizona+2
Parameter learning in PGMs refers to:
a. Discovering the graph structure
b. Estimating the parameters of probability distributions
c. Data normalization
d. Decoding output sequences
Answer: b. Estimating the parameters of probability distributionsmcb111+2
The “moralization” step in PGMs converts:
a. A factor graph to a Bayesian Network
b. A Bayesian Network to an undirected graph
c. A Markov Network to a tree
d. A chain graph to a clique
Answer: b. Bayesian Network to an undirected grapharizona+1
Which of the following can be learned from data in a PGM?
a. Graph structure and parameters
b. Activation functions
c. Training loss history
d. Feature importance ranking only
Answer: a. Graph structure and parameterscse.iitd+2
Which PGM is defined by a bipartite structure?
a. Markov Random Field
b. Factor graph
c. Bayesian Network
d. Dynamic Bayesian Network
Answer: b. Factor graphfiveable+1
Conditional Random Fields are typically used for:
a. Semi-supervised learning
b. Sequence labeling
c. Dimensionality reduction
d. Outlier detection
Answer: b. Sequence labelingarxiv+2
In a Markov model, the “Markov assumption” states that:
a. Each variable is independent of all others
b. Each variable only depends on its immediate predecessor
c. Only terminal nodes are dependent
d. Variables are jointly independent
Answer: b. Each variable only depends on its immediate predecessortelnyx+2
The joint probability in a Bayesian Network is factorized according to:
a. Induced cliques
b. Node degree
c. Parents of each node
d. Tree height
Answer: c. Parents of each nodecse.iitm+2
In graphical terms, d-separation refers to:
a. Strengthening graph connectivity
b. Determining conditional independence
c. Minimizing joint entropy
d. Maximizing compatibility
Answer: b. Determining conditional independencecse.iitm+2
Variational inference in PGMs is mainly used for:
a. Graph structure estimation
b. Approximating complex posterior distributions
c. Distance metric learning
d. Data normalization
Answer: b. Approximating complex posterior distributionspmc.ncbi.nlm.nih+2
Which type of PGM is most useful in image analysis tasks?
a. Bayesian Networks
b. Markov Random Fields
c. Factor Graphs
d. Decision Trees
Answer: b. Markov Random Fieldsdzone+2
Belief propagation is a technique for:
a. Sampling from posterior distributions
b. Exact or approximate inference
c. Learning model parameters
d. Structure learning
Answer: b. Exact or approximate inferencemcb111+2
Which model allows for cycles in their undirected graphical structure?
a. Bayesian Network
b. Markov Random Field
c. Hidden Markov Model
d. Naive Bayes Classifier
Answer: b. Markov Random Fieldsciencedirect+2
The Hammersley-Clifford theorem connects Markov Random Fields with:
a. Maximum entropy principle
b. Factorization of probability distributions
c. D-separation criterion
d. Expectation-maximization
Answer: b. Factorization of probability distributionsafrica.engineering.cmu+2
Dynamic Bayesian Networks extend Bayesian Networks for:
a. Multi-class classification
b. Time-series modeling
c. Unsupervised clustering
d. Nonlinear regression
Answer: b. Time-series modelingarxiv+2
Which inference method is most commonly used in tree-structured Bayesian Networks?
a. Markov Chain Monte Carlo
b. Forward-backward algorithm
c. Belief propagation
d. Gibbs sampling
Answer: c. Belief propagationsciencedirect+2
The clique tree algorithm is used for:
a. Infinite mixture models
b. Efficient inference in undirected graphical models
c. Structure learning in chain graphs
d. Learning activation functions
Answer: b. Efficient inference in undirected graphical modelssciencedirect+2
Maximum likelihood learning in PGMs requires:
a. Labeled data only
b. Estimating parameters to maximize joint probability
c. Hidden variables to be observed
d. Complete independence between features
Answer: b. Estimating parameters to maximize joint probabilityfiveable+2
In EM (Expectation-Maximization), the E-step involves:
a. Sampling latent variables
b. Computing expected values given current parameters
c. Maximizing likelihood
d. Fitting a decision tree
Answer: b. Computing expected values given current parameterscmu+2
Which term describes the graphical model representing dependencies as a directed acyclic graph?
a. Factor graph
b. Bayesian Network
c. Restricted Boltzmann Machine
d. Markov Random Field
Answer: b. Bayesian Networkcse.iitm+2
In a conditional independence graph, an edge between two variables implies:
a. They are marginally independent
b. They are conditionally dependent unless d-separated
c. They are causally unrelated
d. No information shared
Answer: b. They are conditionally dependent unless d-separatedcse.iitd+2
The main inference challenge in large PGMs comes from:
a. Too many observed variables
b. High-dimensional marginalization
c. Lack of labeled data
d. Excessive regularization
Answer: b. High-dimensional marginalizationdzone+2
Latent variables in PGMs are typically:
a. Directly observed
b. Unobserved contributors to the model
c. Output labels
d. Measurement errors
Answer: b. Unobserved contributors to the modeltelnyx+2
Structure learning in PGMs aims to:
a. Estimate hidden parameters
b. Discover optimal graph connectivity from data
c. Normalize data distributions
d. Cluster categorical features
Answer: b. Discover optimal graph connectivity from datamcb111+2
A Markov blanket of a node in a Bayesian Network is:
a. The complete graph
b. The set comprising its parents, children, and co-parents
c. Only its direct parents
d. All leaf nodes
Answer: b. The set comprising its parents, children, and co-parentscse.iitd+2
The partition function in a Markov Random Field serves to:
a. Normalize the joint probability
b. Compute gradients
c. Reduce overfitting
d. Select hyperparameters
Answer: a. Normalize the joint probabilitydzone+2
The process of “triangulating” a graph is mainly used for:
a. Ensuring tree-structured inference
b. Converting cyclical graphs to acyclic
c. Facilitating efficient inference algorithms
d. Removing all cycles
Answer: c. Facilitating efficient inference algorithmssciencedirect+2
Which approach approximates inference by generating samples from the distribution?
a. Message passing
b. Sampling-based inference (e.g., MCMC)
c. Structure learning
d. Belief propagation
Answer: b. Sampling-based inference (e.g., MCMC)arxiv+2
Chain graphs allow the combination of:
a. Only directed edges
b. Directed and undirected edges
c. Only undirected edges
d. Weighted edges only
Answer: b. Directed and undirected edgescse.iitm+2
The no-cloning theorem in PGMs is relevant to:
a. Quantum graphical models
b. Structure learning
c. Inference in classical Bayesian networks
d. Parameter estimation
Answer: a. Quantum graphical modelsdzone+1
Belief propagation works exactly in:
a. Loopy graphs
b. Tree-structured graphs
c. Bipartite cliques
d. Fully connected graphs
Answer: b. Tree-structured graphsmcb111+2
Learning in PGMs with hidden variables is typically done using:
a. Naive Bayes
b. Expectation-Maximization
c. Principal Components Analysis
d. Convolutional networks
Answer: b. Expectation-Maximizationcmu+2
Which model is commonly used for image denoising?
a. Bayesian Network
b. Markov Random Field
c. Factor graph
d. Hidden Markov Model
Answer: b. Markov Random Fieldtelnyx+2
In a directed graphical model, an edge from X to Y indicates:
a. X is conditionally independent of Y
b. X directly influences Y
c. Y is always observed
d. X and Y are unrelated
Answer: b. X directly influences Yfiveable+2
The concept of “plates” in PGMs is used for:
a. Representing repeated independent structures
b. Marking observed data
c. Highlighting hidden variables
d. Showing causal loops
Answer: a. Representing repeated independent structurescse.iitm+1
For structure learning in Bayesian Networks, which scoring function is commonly used?
a. Ridge regression
b. Bayesian Information Criterion (BIC)
c. Mean squared error
d. Partition function
Answer: b. Bayesian Information Criterion (BIC)mcb111+1
Node potentials are associated with which type of graphical model?
a. Bayesian Network
b. Markov Random Field
c. Decision Trees
d. Naive Bayes
Answer: b. Markov Random Fieldafrica.engineering.cmu+2
Inference in a loopy graph typically requires:
a. Tree decomposition
b. Loopy belief propagation
c. Factor elimination
d. K-means clustering
Answer: b. Loopy belief propagationarxiv+1
The EM algorithm alternately optimizes:
a. Graph structure and evidence
b. Latent variable expectations and parameter estimates
c. Node potentials and observation weights
d. Marginal probabilities and decision boundaries
Answer: b. Latent variable expectations and parameter estimatescmu+2
The sum-product algorithm is primarily used for:
a. Parameter estimation
b. Inference on factor graphs
c. Regularization
d. Loss minimization
Answer: b. Inference on factor graphsarxiv+1
Probabilistic Graphical Models are widely used for all EXCEPT:
a. Machine translation
b. Speech recognition
c. Reinforcement learning policy updating
d. Deterministic rule-based systems
Answer: d. Deterministic rule-based systemstelnyx+2

Blog

Systemethic Technologies

Leave a Reply Cancel reply

Useful links

Other Terms & Policies

Contact Info