1) How does a PROLOG interpreter know whether the cut used in a PROLOG program

is a Green cut or Red cut?

Ans: How to know if a cut is Green or Red? Check these points:

1. Position:
a. If the cut is added after checking something or at the end of a rule, it’s
probably a Green cut.
2. Effect:
a. If removing the cut doesn’t change the answers, it’s a Green cut.
b. If removing the cut gives different answers, it’s a Red cut.
3. Variable values:
a. If the cut doesn’t stop Prolog from finding correct values for variables, it’s a
Green cut.
b. If the cut blocks correct values, it’s a Red cut.
4. Backtracking:
a. If the cut stops Prolog from doing extra, unnecessary work, it’s a Green cut.
b. If it stops Prolog from finding correct solutions, it’s a Red cut.

2) Is there a mechanism for a creating a loop/getting impact of loop in PROLOG?

Justify

Ans: PROLOG is considered a 5th-generation programming language because of its focus

on logic, declarative programming, AI applications, and automatic problem-solving, which
align with the goals of 5GLs. Other 5GLs like OPS5, Mercury, and KL-1 share similar
characteristics, focusing on reasoning and inference rather than procedural programming.

3) Given two wffs (A→ B) and (B), for any expression A and B, if it is consistent to
assume A, do so." Is above statement Logical? Justify your answer

Ans : Simple Example:

• A : "It is raining."
• B : "The ground is wet."
• A→B : "If it is raining, the ground is wet."
• Fact: The ground is already wet (B is true).
If it doesn’t cause any contradictions, you can assume A (that it is raining). Even if it wasn’t
raining, the ground being wet would still be fine.

4) Can there be need of backtracking when Resolution is used for answer extraction?
Justify your answer.

Ans : In standard Resolution-based answer extraction, backtracking is not necessary due

to the deterministic nature of the inference process and the uniqueness of the refutation
proof.

5) What are the arguments favoring use of heuristic?

Ans : Efficiency and Speed, Limited Information and Uncertainty, Cognitive Biases and
Human Decision-Making, Optimization and Satisficing, Real-World Applications

6) Describe Lexical Semantic Ambiguity problem in NLP with appropriate example.

Ans : Lexical Semantic Ambiguity in NLP (Natural Language Processing) happens when a
word has more than one meaning, and it's unclear which meaning is being used based on
the context.

Some words have multiple meanings depending on how they are used in a sentence. The
problem is that, without enough context, it's hard for a computer to figure out which
meaning of the word is intended.

Take the word "bank":

• Bank can mean a financial institution (like a place where you keep your money).
• Bank can also mean the side of a river (where the land meets the water).
Now, look at these two sentences:

1. "I went to the bank to get some money."

2. "The fish swam near the bank of the river."

In the first sentence, "bank" clearly means a financial institution. But in the second
sentence, "bank" refers to the side of the river. The meaning of the word changes based
on the context.
 7) Difference between Best First Search and Steepest ascent hill climbing.

Ans : Here’s the key difference in simpler terms:

• Best First Search looks at the entire journey ahead and picks the path that seems
to lead to the best overall outcome, not just the next step.
• Steepest Ascent Hill Climbing only looks at the next step and chooses the one
that seems best for getting higher right now, without worrying about how it affects
the whole journey.

8) List the capabilities, which are required by machines to pass the Turing's test. Write
the purpose of each capability listed.

Ans : Natural Language Processing (NLP) Capabilities:

1. Text Understanding:
a. The machine needs to understand human language, including the meaning
and context of what is being said.
b. Purpose: This helps the machine answer questions or respond to
statements in a meaningful way.
2. Language Generation:
a. The machine should be able to speak or write like a human, producing text
or speech that sounds natural.
b. Purpose: This allows the machine to express itself clearly and respond like a
human.
3. Conversational Dialogue Management:
a. The machine should be able to carry on a conversation by keeping track of
the context and making its responses fit into the discussion.
b. Purpose: This ensures that the conversation flows smoothly and feels
natural.

Reasoning and Problem-Solving Capabilities:

1. Logical Reasoning:
a. The machine should be able to think logically, drawing conclusions based
on the information it has.
 b. Purpose: This shows the machine can make good decisions, just like
humans.
2. Common Sense:
a. The machine should have a general understanding of the world to make
practical, sensible decisions.
b. Purpose: This ensures the machine’s answers make sense in everyday
situations.
3. Problem-Solving:
a. The machine should be able to solve problems that might not have a clear
or easy answer.
b. Purpose: This shows the machine can think creatively and find ways to deal
with challenges.

Knowledge Representation and Learning Capabilities:

1. Knowledge Acquisition:
a. The machine should be able to learn from its experiences or from the
information it’s given.
b. Purpose: This lets the machine improve over time and adapt to new
situations.
2. Knowledge Representation:
a. The machine should have a way to organize and store information in a
useful way.
b. Purpose: This helps the machine use its knowledge efficiently when needed.
3. Memory and Forgetting:
a. The machine should be able to remember important things and forget
things it no longer needs.
b. Purpose: This mimics how humans remember things and forget irrelevant
details.

Social and Emotional Intelligence Capabilities:

1. Emotional Understanding:
a. The machine should be able to recognize emotions and respond in a way
that feels compassionate or appropriate.
b. Purpose: This helps the machine connect with people emotionally, just like
humans do.
2. Social Awareness:
 a. The machine needs to understand social rules and how people interact.
b. Purpose: This ensures that the machine acts appropriately in different social
situations.
3. Personality and Humor:
a. The machine should have a consistent personality and be able to use
humor to make conversations more engaging.
b. Purpose: This makes the machine feel more human and enjoyable to talk to.

9) Do as directed.
(i) Alpha-beta pruning can alter the computed minimax value of the root of a game search
tree. True or False?
(ii) State the condition under which alpha-beta pruning will maximize the number of
branches pruned
Ans : (i) Alpha-beta pruning can change the minimax value at the root of the game tree.
True or False?

• Answer: False.

Why? Alpha-beta pruning helps speed up the decision-making in a game by cutting out
parts of the tree that don’t need to be explored. However, it does not change the final
answer (minimax value) at the root. The root's value stays the same, but the process of
finding it becomes faster.

(ii) When does alpha-beta pruning cut off (prune) the most branches?

• Answer: Alpha-beta pruning cuts off the most branches when the best moves are
checked first.

Why? If you explore the best moves first, the pruning can happen much earlier in the
process. This means more branches of the tree get cut off quickly, making the process
faster. The order in which you explore the moves matters—if the good moves are checked
first, the pruning will work better.

9) 'A good control strategy requires that it be systematic.' Justify this statement with a
proper example.

Ans : a good control strategy that is systematic makes sure the process is organized and
effective, avoiding random decisions that could lead to failure or confusion
 10) Is there a mechanism for a creating a loop/getting impact of loop in PROLOG?
Justify your answer

Ans : Yes, Prolog can create loops, but it does so in a special way, not like in regular
programming languages.

parent(X, Y).

parent(Y, Z).

If you ask Prolog to find out who is a grandparent (e.g., grandparent(X, Z)), Prolog will
keep looking for parents of parents, and might end up checking the same relationships
again and again, creating a loop.

Why it works like this:

• Prolog doesn’t use regular loops (like for or while), but it repeats actions by
backtracking when it needs to try other options.
• If the rules are set in a way that makes Prolog keep going back to the same places
without an end, it creates a loop.

11) Compare Propositional Logic with Predicate Logic.

• Ans : Propositional Logic is about simple, true/false statements.
• Predicate Logic is more detailed and can express relationships between things,
involving variables and specific details.

12) What are the advantages & Limitations of Semantic Nets

Ans: Advantages of Semantic Nets:

• Easy to Understand: They use simple diagrams to show relationships between

concepts.
• Clear Relationships: They clearly show how different things are connected.
• Good for Organizing Knowledge: They help store and represent knowledge.
• Supports Inheritance: Concepts can inherit properties from broader categories.
Limitations of Semantic Nets:

• Lack of Detail: They don’t explain concepts in depth.

• Can Get Messy: They become complex and hard to manage with many concepts.
• Ambiguity Issues: They struggle with words that have multiple meanings.
• Limited for Actions: They are not good at showing actions or processes.

13) Under which situation it will be preferred to use Forward reasoning against
Backward reasoning?

Ans : Forward reasoning is preferred when:

• You have a lot of known facts or starting information and want to build
conclusions or outcomes from these facts.
• The goal is unclear or general, and you are exploring various possibilities step by
step.