Artificial Intelligence MCQs - Knowledge Representation and Logic (Set 5)

Artificial Intelligence MCQs - Knowledge Representation and Logic (Set 5) - CodingHumans

This set of Artificial Intelligence MCQs covers how agents represent and reason with knowledge. It tests propositional logic equivalences, CNF conversion, resolution refutation, translating English into first-order logic, Skolemisation, unification with the occurs check, Horn clauses, forward and backward chaining, and structured representations such as semantic networks and frames. The set suits GATE and UGC NET candidates, university semester exams and interviews that cover logic-based AI.

How to practise: read each question and tap the option you think is correct. You will see right away whether you got it, along with a short explanation, and your score is tracked at the bottom of the screen. This set has 18 questions.

Artificial Intelligence MCQ series: Set 1 | Set 2 | Set 3 | Set 4 | Set 5 | Set 6

Topics covered in this set

  • Validity, satisfiability and entailment
  • Propositional equivalences and CNF
  • Resolution and refutation proofs
  • First-order logic translation and Skolemisation
  • Unification and the occurs check
  • Forward and backward chaining, semantic nets and frames

Artificial Intelligence MCQs with answers and explanations

Q1. A propositional sentence that is true in every possible model is called:

a) satisfiable

b) valid (a tautology)

c) contingent

d) unsatisfiable

Q2. Which inference rule derives Q from the premises P and P => Q?

a) Modus tollens

b) And-elimination

c) Resolution with a unit clause only

d) Modus ponens

Q3. The implication P => Q is logically equivalent to which sentence?

a) not P or Q

b) P or not Q

c) not P and Q

d) Q => P

Q4. Which sentence is the contrapositive of P => Q, and therefore equivalent to it?

a) Q => P

b) not P => not Q

c) not Q => not P

d) P => not Q

Q5. What is the conjunctive normal form (CNF) of the biconditional A <=> B?

a) (A or B) and (not A or not B)

b) (not A or B) and (not B or A)

c) (A and B) or (not A and not B)

d) (not A or B) and (A or B)

Q6. What is the resolvent of the clauses (P or Q) and (not Q or R)?

a) P or Q or R

b) P and R

c) Q or R

d) P or R

Q7. To prove KB entails alpha by resolution refutation, one shows that:

a) KB and not alpha is unsatisfiable, deriving the empty clause

b) KB and alpha is satisfiable

c) alpha alone is valid

d) not KB and alpha is unsatisfiable

Q8. What is the most general unifier of Knows(John, x) and Knows(y, Mother(y))?

a) {x/John, y/Mother(John)}

b) {y/x, x/John}

c) {y/John, x/Mother(John)}

d) The terms do not unify

Q9. Why do P(x, x) and P(Alice, Bob) fail to unify, where Alice and Bob are distinct constants?

a) Constants can never appear in a unifier

b) x would need to equal both Alice and Bob at once

c) The predicates have different arities

d) The occurs check fails

Q10. Unifying the variable x with the term F(x) fails because of the:

a) frame problem

b) closed-world assumption

c) qualification problem

d) occurs check

Q11. Which statement correctly contrasts forward and backward chaining?

a) Forward chaining is data-driven, starting from known facts; backward chaining is goal-driven, working back from the query

b) Forward chaining works only in first-order logic; backward chaining only in propositional logic

c) Both are goal-driven but differ in rule ordering

d) Forward chaining is incomplete for Horn clauses while backward chaining is complete

Q12. A Horn clause is a disjunction of literals with:

a) exactly two positive literals

b) no negative literals

c) at most one positive literal

d) at least one positive literal

Q13. Which first-order sentence correctly expresses 'Every student likes some course'?

a) Exists y for all x: Student(x) and Course(y) and Likes(x, y)

b) For all x: Student(x) => (Exists y: Course(y) and Likes(x, y))

c) For all x: Student(x) and (Exists y: Course(y) => Likes(x, y))

d) Exists x: Student(x) => (For all y: Course(y) and Likes(x, y))

Q14. Which translation of 'Some king is wise' is correct?

a) For all x: King(x) => Wise(x)

b) Exists x: King(x) => Wise(x)

c) For all x: King(x) and Wise(x)

d) Exists x: King(x) and Wise(x)

Q15. What is the Skolemised form of For all x Exists y Loves(x, y)?

a) For all x Loves(x, F(x))

b) For all x Loves(x, C)

c) Loves(C, F(C))

d) Exists y Loves(y, y)

Q16. In a semantic network, which kind of link mainly supports inheritance of properties from a category to its members and subcategories?

a) Part-of link

b) Causal link

c) IS-A (subset/member) link

d) Temporal link

Q17. In Minsky's frame-based representation, knowledge about a stereotyped object is stored as:

a) a set of clauses in conjunctive normal form

b) slots with fillers, which may include default values and attached procedures

c) a probability table for each attribute

d) a sequence of production rules only

Q18. Over the propositional symbols A, B and C, how many models satisfy (A or B) and not C?

a) 3

b) 2

c) 4

d) 6

Quick revision notes

  • Valid = true in all models; satisfiable = true in some; KB entails alpha iff KB and not alpha is unsatisfiable.
  • P => Q == not P or Q == not Q => not P (contrapositive).
  • Resolution: (A or B), (not B or C) gives (A or C); refutation derives the empty clause.
  • Universal quantifier pairs with =>, existential pairs with and.
  • Skolemisation: existential inside universal becomes a function of the universal variables.
  • Unification needs the occurs check; MGU = most general unifier.
  • Horn clause: at most one positive literal; forward chaining data-driven, backward chaining goal-driven.
  • Semantic nets use IS-A links for inheritance; frames use slots, fillers and defaults.

Frequently asked questions

What is the difference between validity and satisfiability?

A valid sentence is true in every model, while a satisfiable sentence is true in at least one model. A sentence is valid exactly when its negation is unsatisfiable.

Why is Skolemisation needed before resolution in first-order logic?

Resolution works on clauses with only implicitly universal variables. Skolemisation removes existential quantifiers by replacing them with new constants or functions, keeping the sentence satisfiable whenever the original was.

When should backward chaining be preferred over forward chaining?

Backward chaining suits answering a specific query, since it only touches facts and rules relevant to that goal. Forward chaining suits situations where new data arrives and all consequences should be derived.

Comments

Answered 0/0Correct 0