Module 5 – Important 10 Mark Questions
& Answers
1. Explain Reinforcement Learning with an example.
Answer:
Reinforcement Learning (RL) is a type of machine learning where an agent learns by
interacting with an environment to maximize cumulative reward.
Key Components:
Agent – Learner/decision maker
Environment – External system
State (S) – Current situation
Action (A) – Choices available
Reward (R) – Feedback signal
Working:
1. Agent observes state
2. Takes action
3. Receives reward
4. Updates policy
5.
Example – Getting Lost Problem:
A person tries to find the way out of a maze
Correct path → Positive reward
Wrong path → Negative reward
Over time, learns the best path
Conclusion:
RL is widely used in robotics, gaming, and AI decision-making systems.
2. Explain Markov Decision Process (MDP).
Answer:
An MDP is a mathematical framework for modeling decision-making.
Components:
S – Set of states
A – Set of actions
P(s'|s,a) – Transition probability
R(s,a) – Reward function
γ (gamma) – Discount factor
Properties:
Follows Markov Property
Future depends only on present state
Objective:
Find optimal policy π* that maximizes reward.
Applications:
Robotics
Game AI
Autonomous vehicles
3. Explain Markov Chain Monte Carlo (MCMC)
Methods.
Answer:
MCMC is used to sample from complex probability distributions.
Concept:
Construct a Markov Chain
Stationary distribution = Target distribution
Steps:
1. Start with initial value
2. Generate next sample based on current state
3. Repeat to form chain
Advantages:
Works for high-dimensional data
Used in Bayesian inference
Applications:
Machine learning
Physics simulations
4. Explain Proposal Distribution in MCMC.
Answer:
A Proposal Distribution suggests the next candidate sample.
Working:
Current state → propose new state
Accept or reject using probability
Common Types:
Gaussian distribution
Uniform distribution
Importance:
Affects convergence speed
Better proposal → faster results
5. Explain Bayesian Networks with structure and
properties.
Answer:
A Bayesian Network (BN) is a directed acyclic graph (DAG) representing probabilistic
relationships.
Components:
Nodes → Random variables
Edges → Dependencies
Properties:
Conditional independence
Joint probability representation
Formula:
Applications:
6. Explain Markov Random Fields (MRF).
Answer:
MRF is an undirected graphical model.
Features:
Nodes represent variables
Edges represent dependencies
Properties:
Local Markov property
Uses cliques
Applications:
Image processing
Computer vision
7. Explain Hidden Markov Model (HMM).
Answer:
An HMM is a statistical model where states are hidden.
Components:
Hidden states
Observations
Transition probabilities
Emission probabilities
Problems in HMM:
1. Evaluation
2. Decoding
3. Learning
Applications:
Speech recognition
NLP
8. Explain Sampling Methods in Probabilistic Learning.
Answer:
Sampling methods are used to approximate distributions.
Types:
Random Sampling
Importance Sampling
Gibbs Sampling
Advantages:
Handles large datasets
Reduces complexity
9. Explain Tracking Methods in Probabilistic Models.
Answer:
Tracking methods estimate state over time.
Types:
Kalman Filter
Particle Filter
Applications:
Object tracking
Navigation systems
10. Differentiate Bayesian Networks and Markov Random
Fields.
Answer:
Feature Bayesian Network Markov Random Field
Graph Type Directed Undirected
Structure DAG Cyclic allowed
Probability Conditional Joint
Example Medical diagnosis Image processing