NEW
Font size
WorksheetsASSIGNMENT 5-UNIT V-AI-AL23311
Total questions: 20
Worksheet time: 10mins
Why must AI agents handle uncertainty?
Deterministic environments
Complete observability
Partial observability and nondeterminism
Logical consistency
What is a 'belief state'?
The physical state of the environment
The agent's internal program
The set of all possible world states the agent might be in
The probability of an event occurring
Logical agents fail in uncertain environments because:
They lack sensors
They must consider every possible explanation
They do not have actuators
They ignore probabilities
Which theorem updates beliefs based on new evidence?
Shannon's Law
Bayes' Theorem
Boolean Logic
Regression Law
Bayes' theorem is:
P(A|B) = P(A) + P(B)
P(A|B) = P(B|A) × P(A) / P(B)
P(A|B) = P(A) × P(B)
P(A|B) = P(A) - P(B)
'Prior probability' means:
Before evidence is observed
After evidence is observed
Joint probability
Posterior probability
Naïve Bayes assumes features are:
Correlated
Conditionally independent given the class
Dependent on each other
Unrelated to the class variable
Common application of Naïve Bayes:
Image generation
Spam filtering
Object tracking
Path planning
Limitation of Naïve Bayes:
Slow execution
Assumes feature independence
Needs huge datasets
Ignores probability values
Probabilistic reasoning combines:
Logic and probability theory
Deep learning and reasoning
Fuzzy logic and constraints
Deterministic rules only
Which is NOT a probabilistic model?
Bayesian network
Markov model
Decision tree
Hidden Markov model
A probability distribution assigning values to all atomic events is:
Conditional probability
Joint probability distribution
Marginal probability
Prior probability
Bayesian networks represent:
Logical statements
Probabilistic causal relationships
Linear models
Deterministic data
In Bayesian networks, nodes and edges represent:
Variables and causal dependencies
Inputs and outputs
Equations and constants
Agents and actions
Exact inference means:
Estimating random samples
Computing precise posterior probabilities
Ignoring priors
Using approximations
Approximate inference is used when:
Network is small
Exact inference is too expensive
There is no uncertainty
All variables are known
A common approximate inference method:
Monte Carlo sampling
BFS algorithm
A* search
Constraint propagation
Causal networks are an extension of:
Neural networks
Bayesian networks
Markov chains
Expert systems
Causal reasoning helps in:
Explaining cause-effect relationships
Reducing computation
Improving hardware performance
Encoding syntax rules
Probabilistic reasoning primarily deals with:
Certainty
Heuristics
Quantifying uncertainty
Symbolic logic
