Files
G4G0-2/AI & Data Mining/Week 23/Week 23 - Deductive Proofs.md
2025-03-16 18:59:42 +00:00

5.0 KiB
Raw Blame History

Slide 3: Recap on Logical Implication (Entailment) |-|=-

  • Entailment notation: p |=q if and only if the implication p$\implies$q is a tautology.
  • Example:
    • p \land q |=q
    • Truth table for p \implies q:
p q p \land q p \implies q
T T T T
T F F F
F T F T
F F F T

Slide 4: (r \implies s) \land (r \implies $\lnot$s) |-|=-

  • Intuitively, if r implies both s and $\lnot$s, then r must be false.
  • Truth table for (r \implies s) \land (r \implies $\lnot$s):
r s $\lnot$s (r \implies s) \land (r \implies $\lnot$s)
T T F F
T F F F
F T T T
F F T T

Slide 5: p \vdash q

  • Notation: p \vdash q means q is provable from p using inference rules.
  • Example:
    • A \implies B, $\lnot$A, therefore $\lnot$B

Slide 6: Differences Between |-|=- and $\vdash$

  • |= indicates semantic entailment (truth conditions).
  • \vdash represents syntactic derivation (inference rules).

Slide 7: Recap on Inference Rules

  • Example inference rules:
    • Modus Ponens ($\implies$Elim):

      p \implies q, p \vdash q

    • Conjunction Introduction ($\land$Intro):

      p \vdash q, p \vdash r \vdash p \land q

    • Conditional Proof ($\implies$Intro):

      p \vdash r, p \vdash s \vdash p \implies (r \land s)

Slide 8: Layout of an Inference Rule

  • Premises above the line, conclusion below the line.

  • Example inference rule ($\implies$Intro):

    p \vdash r, p \vdash s p \implies (r \land s)

Slide 9: Presentation of Proofs

  • Steps:
    • Number each step.
    • Justify each step with previous line(s) and inference rule used.

Slide 10: Deriving $\lnot$p \implies r From (p \land q) \lor r

  • Example proof:

    (p \land q) \lor r, $\lnot$E … $\lnot$p \implies r

Slide 11: Two Special Inference Rules

  • Deductive Theorem ($\implies$Intro):

    p \vdash r, p \vdash s p \implies (r \land s)

  • Reductio ad absurdum ($\lnot$Intro):

    p \vdash r, p \vdash $\lnot$s p \vdash $\lnot$r

Slide 12: Conditional Proofs

  • Strategy: Assume p, deduce q if possible, discharge assumption.

  • Example:

    (p \land q) \lor r … $\lnot$p \implies r

Slide 13: Indirect Proofs

  • Strategy: Assume negation of goal, deduce contradiction.

  • Example:

    (p \land q) \lor r … $\lnot$p \implies r

Slide 14: Solution to Exercise

Given argument: A (You eat carefully) ⇒ B (You have a healthy digestive system) C (You exercise regularly) ⇒ D (You are very fit) B D ⇒ E (You live to a ripe old age) ¬E Therefore, ¬A ∧ ¬C

Proof:

Line Formula Justification
1 A ⇒ B Premise
2 C ⇒ D Premise
3 B D ⇒ E Premise
4 ¬E Premise
5 ¬(B D) Modus Tollens (3, 4)
6 ¬B ∧ ¬D De Morgan's Law (5)
7 ¬B ∧Elim (6)
8 ¬A Modus Tollens (1, 7)
9 ¬D ∧Elim (6)
10 ¬C Modus Tollens (2, 9)
11 ¬A ∧ ¬C ∧Intro (8, 10)

Conclusion: We have proven that ¬A ∧ ¬C, i.e., you did not eat carefully and you did not exercise regularly.

Slide 15: Two Special Inference Rules (continued)

  • Deductive Theorem:

    p \vdash r, p \vdash s p \implies (r \land s)

  • Reductio ad absurdum:

    p \vdash r, p \vdash $\lnot$s p \vdash $\lnot$r

Slide 16: Soundness and Completeness

  • Sound: Valid argument with true premises.
  • Complete: Derives any sentence entailed by premises.

Slide 17: Formal Proofs of Natural Language Arguments

  • Steps:
    • Identify atomic propositions.
    • Formalize argument in logic.
    • Check for invalidity.
    • Attempt proof.

Slide 18: Example - Travel

  • Argument:

    … Therefore, if my neighbours claim to be impressed then they are just pretending.

Slide 19: Example - Travel (continued)

  • Formalize argument:

    p \implies q, $\lnot$p \implies $\lnot$r, $\lnot$q … $\lnot$r

  • Proof:

    … $\lnot$p \implies r

Slide 20: Example - Nutrition

  • Argument:

    … Therefore, you did not eat carefully and you did not exercise regularly.

Slide 21: Example - Nutrition (continued)

  • Formalize argument: A \implies B, C \implies D, B \lor D \implies E, $\lnot$E … $\lnot$A \land $\lnot$C
  • Proof: … $\lnot$A \land $\lnot$C

Slide 22: Application to Software Engineering

  • Questions about software specifications and claims are arguments.

Slide 23: Reading and References

  • Russell and Norvig, Artificial Intelligence (4th Edition)
  • Nissanke, Introductory Logic and Sets for Computer Scientists
  • Gray, Logic, Algebra and Databases