AIML 2 marks

1. Application of AI:
• Natural Language Processing (NLP): Used in chatbots and virtual assistants (e.g.,
Google Assistant, Siri). NLP enables machines to understand and generate human
language.
• Computer Vision: Employed in facial recognition systems (e.g., Apple's Face ID) and
object detection (e.g., autonomous vehicle navigation).
• Robotics: Applied in autonomous robots (e.g., Roomba vacuum cleaners, Boston
Dynamics’ Spot robot) for tasks like cleaning and exploration.
• Healthcare: Utilized for diagnostics (e.g., IBM Watson for oncology), personalized
medicine, and predictive analytics for disease outbreaks.
• Finance: Implemented in fraud detection systems (e.g., PayPal's fraud prevention)
and algorithmic trading.
2. PEAS for Taxi Driver Agent:
• Performance measure: Profit maximization, passenger safety, high customer
satisfaction, fuel cost efficiency.
• Environment: Roads, traffic conditions, passengers, other vehicles, weather.
• Actuators: Steering wheel, accelerator, brake, GPS, fare meter.
• Sensors: Cameras (for traffic and passenger monitoring), GPS (for location tracking),
odometer (distance measurement), speedometer (speed monitoring), microphone
(for voice commands).
3. Compare Problem and Goal Formulation:
• Problem Formulation: Outlines the initial state, possible actions, state transitions,
and goal states. For example, in a puzzle game, it includes the puzzle's starting
configuration, the possible moves, and the goal configuration.
• Goal Formulation: Focuses on defining what the end state should look like. For
example, in a maze-solving problem, the goal formulation specifies reaching the exit.
4. Model-Based Reflex Agent:
A model-based reflex agent uses an internal representation of the world to make
decisions. For example, a smart thermostat adjusts room temperature based on past
usage patterns and current weather conditions, even if it doesn't directly sense every
change.
5. Problem Formulation for 8-Puzzle Game:
 • Initial State: Any shuffled configuration of the 8 puzzle tiles.
• Goal State: Tiles arranged sequentially from 1 to 8 with the empty space in the last
position.
• State Space: All possible configurations of the tiles (factorial complexity due to
permutations).
• Actions: Sliding a tile into the adjacent empty space.
• Goal Test: Checking if the current configuration matches the goal configuration.
6. Genetic Algorithm:
A genetic algorithm mimics biological evolution to find optimal solutions. For
example, it can optimize scheduling by evolving a population of schedules, combining
successful schedules (crossover), and introducing random changes (mutation) to
explore new possibilities.
7. CSP (Constraint Satisfaction Problem):
In a CSP, constraints define the relationships between variables. For example, in a
Sudoku puzzle:
• Variables: The cells in the grid.
• Domains: Numbers 1 through 9 for each cell.
• Constraints: Numbers must be unique in each row, column, and 3x3 sub-grid.
8. Informed Search Technique:
Informed search techniques use heuristics to improve search efficiency. For instance,
A search* combines the cost to reach a node and the estimated cost to the goal
(using heuristics) to find the shortest path more efficiently than uninformed
methods.
10. Bayes' Theorem in AI:
P(A∣B)=P(B∣A)⋅P(A)P(B)P(A|B) = \frac{P(B|A) \cdot P(A)}{P(B)}P(A∣B)=P(B)P(B∣A)⋅P(A)
Example: If A is having a disease and B is observing symptoms, Bayes' theorem helps update
the probability of having the disease given the symptoms.
11. Probabilistic Reasoning:
Used to handle uncertainty by modeling possible outcomes. For instance, probabilistic
reasoning is used in weather forecasting to predict the likelihood of rain based on historical
data and current conditions.
12. Likelihood Weighting:
A method used in Bayesian networks to estimate the distribution of variables given
evidence. For example, in a network modeling medical conditions, samples representing
different symptoms are weighted according to how well they fit the observed evidence.
13. Causal Network:
A graphical model that depicts causal relationships. For example, in a network modeling the
effect of smoking on lung cancer, nodes represent variables like smoking status and cancer
presence, and edges represent causal links.
14. Gibbs Sampling:
A Markov Chain Monte Carlo (MCMC) method used to approximate distributions by
iteratively sampling from conditional distributions. For example, in image denoising, Gibbs
sampling helps estimate pixel values based on the values of neighboring pixels.
15. Application of Bayesian Networks:
Used in various domains for reasoning under uncertainty. Examples include medical
diagnosis (predicting disease risk based on symptoms), decision support systems (evaluating
risk), and predictive modeling (forecasting stock prices).
16. Differentiate Between Forward Chaining & Backward Chaining:
• Forward Chaining: Starts with known facts and applies rules to infer new facts (e.g.,
an expert system for troubleshooting car issues begins with symptoms and applies
rules to identify possible problems).
• Backward Chaining: Starts with a goal and works backward to determine if known
facts can satisfy it (e.g., in Prolog, proving a theorem by working backward from the
conclusion).
17."All Pomperians were Romans" -> Convert into FOL:
∀x(Pomperian(x)→Roman(x))\forall x (Pomperian(x) \rightarrow
Roman(x))∀x(Pomperian(x)→Roman(x))
This formalizes that every Pomperian is also a Roman.
18. "Not all students like both Mathematics and English" -> Convert into FOL:
¬∀x(Student(x)→(Likes(x,Mathematics)∧Likes(x,English)))\neg \forall x (Student(x)
\rightarrow (Likes(x, Mathematics) \land Likes(x,
English)))¬∀x(Student(x)→(Likes(x,Mathematics)∧Likes(x,English)))
This indicates that there is at least one student who does not like both Mathematics and
English.
19. Difference Between Propositional and First Order Logic:
• Propositional Logic: Deals with simple propositions that are either true or false (e.g.,
"It is raining" is a proposition with no internal structure).
• First Order Logic (FOL): Extends propositional logic with quantifiers and predicates,
allowing for more complex statements about objects and their relationships (e.g.,
"All humans are mortal").
20. Basic Elements of FOL:
 • Constants: Specific objects (e.g., John).
• Variables: Placeholders for objects (e.g., x).
• Predicates: Represent properties or relationships (e.g., Likes(John, IceCream)).
• Functions: Mappings from objects to objects (e.g., FatherOf(x)).
• Quantifiers: Universal (∀\forall∀) and existential (∃\exists∃) to indicate the scope of
statements.
21. Knows(Richard, x) and Knows(Richard, John) Give Unifier:
The unifier is x=Johnx = Johnx=John, which ensures that the statement Knows(Richard, x)
matches Knows(Richard, John).
22. Advantage and Disadvantage of Forward Chaining:
• Advantages:
1. Complete Exploration: Considers all possible conclusions based on known
facts.
2. Simplicity: Straightforward to implement and understand.
• Disadvantages:
1. Scalability Issues: Can become inefficient with a large number of rules and
facts.
2. Memory Usage: Requires significant memory for storing all possible
conclusions.
23. Advantage and Disadvantage of Backward Chaining:
• Advantages:
1. Efficiency: Focuses on relevant rules related to the goal, often more efficient
for goal-oriented tasks.
2. Relevant Focus: Only searches for facts that directly contribute to achieving
the goal.
• Disadvantages:
1. Incomplete Knowledge: Dependent on a complete and accurate knowledge
base.
2. Potential Infinite Loops: May encounter infinite loops if the goal cannot be
resolved with the given knowledge.
24. Define Prolog and Give Its Application:
Prolog is a logic programming language used for symbolic reasoning and knowledge
representation. Applications include:
 • Natural Language Processing (NLP): Parsing and understanding sentences (e.g.,
grammatical analysis).
• Theorem Proving: Automated proving of mathematical theorems.
• Expert Systems: Systems that provide expert-level decisions or advice (e.g., medical
diagnosis systems).

3. Comparison of Problem and Goal Formulation:

Aspect Problem Formulation Goal Formulation

Defines state space, actions, state Specifies the desired outcome or

Definition
transitions, and goal states. final state.

Outlines how to reach the goal from any Focuses on what the final goal
Focus
given initial state. should be.

In a puzzle game, describes possible In a puzzle game, specifies the final

Example
moves from the starting configuration. tile arrangement as the goal.

9. Comparison of Uninformed and Heuristic Search:

Aspect Uninformed Search Heuristic Search

Definition Searches without domain-specific Uses additional information

knowledge. (heuristics) to guide the search.

Examples Breadth-First Search (BFS), Depth-First A* search, Greedy Best-First

Search (DFS) Search

Efficiency Can be less efficient, exploring all More efficient, as heuristics help
possibilities without prioritization. prioritize promising paths.

Exploration May explore a large number of Focuses on paths that are more
irrelevant paths. likely to lead to the goal.

16. Comparison of Forward Chaining and Backward Chaining:

Aspect Forward Chaining Backward Chaining

Definition Data-driven approach that starts with Goal-driven approach that starts with
known facts and applies rules to infer the goal and works backward to
new facts. verify it.
 Usage Used in systems where the facts and Used in systems where the goal is
rules are well-defined and the goal is defined, and the focus is on finding
not initially known. supporting facts.

Example Fire alarm systems detecting smoke Proving mathematical theorems by

and triggering alarms. working backward from the theorem
to known axioms.

Efficiency Can explore many possible More efficient for goal-oriented tasks
conclusions, which might not always but depends on a complete
be relevant. knowledge base.

19. Comparison of Propositional and First Order Logic:

Aspect Propositional Logic First Order Logic (FOL)

Definition Deals with simple Extends propositional logic with quantifiers and
propositions that predicates for more complex statements about
are either true or objects and their relationships.
false.

Basic Units Propositions (e.g., Predicates, constants, variables, functions, and

PPP, QQQ) quantifiers (e.g., Likes(John,IceCream)Likes(John,
IceCream)Likes(John,IceCream))

Expressiveness Limited to Can express relationships and properties about

expressing basic objects (e.g., "All humans are mortal").
true/false
conditions.

Examples "It is raining" "For all xxx, if xxx is a human, then xxx is mortal"
(proposition). (∀x(Human(x)→Mortal(x))\forall x (Human(x)
\rightarrow Mortal(x))∀x(Human(x)→Mortal(x))).