Propositional Logic: Introduction To Artificial Intelligence

Posts and Telecommunication Institute of Technology
Faculty of Information Technology 1
Introduction to Artificial Intelligence
Propositional logic
Outline
} Knowledge representation and reasoning
} Propositional logic
} Inference in propositional logic
2 http://www.ptit.edu.vn
The need of knowledge and reasoning
} Humans live in the environment
o Perceive the world through the senses (ears, eyes, ...)
o Collected information will be accumulated into knowledge
o Use accumulated knowledge and reasoning/inference ability to
make reasonable actions
} An intelligent system needs to have the ability to use
knowledge and make inference
o High flexibility
§ The combination of known knowledge and inference allows to create
new knowledge
o Can work in incomplete information cases
o Convenience for system building
§ Just change the knowledge base; keep the same inference procedure
Knowledge representation language
Knowledge representation language
= Syntax + Semantics + reasoning mechanism
} Syntax
o Includes symbols and rules for linking symbols (syntactic rules) to
form sentences (formulas) in the language
} Semantics
o Allows to determine the meaning of sentences in a certain domain
of the real world
} Reasoning mechanism
o A computational process
o Input: a set of formulas (formal representation of known
knowledge)
o Output: a set of new formulas (formal representation of new
knowledge)
A good knowledge
representation language
} Good representation ability
o Allows to represent all necessary knowledge of the problem
} Efficient
o Concise knowledge representation
o Inference procedure requires little computation time and little
memory space
} Close to natural language

o Convenient to describe knowledge
Outline
o Syntax
o Semantics
Syntax of propositional logic (1/2)
} Symbols
o Truth symbols (logical constants): True (𝑇) and False (𝐹)
o Propositional symbols (Propositional variables): 𝑃, 𝑄, …
o Logical connectives: ∧,∨, ¬, ⇒, ⇔
o Brackets ( and )
} Syntactic rules
o Truth symbols and propositional variables are formulas
o If 𝐴 and 𝐵 are formulas
§ (𝐴 ∧ 𝐵): “A and B” (conjunction)
§ (𝐴 ∨ 𝐵): “A or B” (disjunction)
§ (¬𝐴) : “not A” (negation)
§ (𝐴 ⇒ 𝐵): “if A then B” (implication)
§ (𝐴 ⇔ 𝐵): “A if and only if B” (equivalence)
Are formulas
Syntax of propositional logic (2/2)
} Unnecessary parentheses can be removed
o Example: ((𝐴 ∨ 𝐵) ∧ 𝐶) can be written as (𝐴 ∨ 𝐵) ∧ 𝐶
} The precedence of logical connectives
o ¬, ∧, ∨, ⇒, ⇔
} Propositional symbols are atomic sentences
o Examples: 𝑃, 𝑄
} If 𝑃 is a propositional symbol then 𝑃 and ¬𝑃 are literal
o 𝑃 is a positive literal, ¬𝑃 is a negative literal
} Complex sentences of the form 𝐴* ∨ 𝐴+ ∨. . .∨ 𝐴, , where
𝐴- are literals, are called clausal sentences
Semantics of propositional logic (1/2)
} Each propositional symbol corresponds to a propositional
statement
o 𝑃 = “Paris is the capital of France”
o 𝑄 = “Pi constant is an integer number”
} A propositional statement is either True (T) or False (F)
o 𝑃 True, 𝑄 False
} An interpretation is a way of assigning each propositional
variable a truth value (True or False)
𝐴 𝐵 ¬𝐴 𝐴∧𝐵 𝐴∨𝐵 𝐴⇒𝐵 𝐴⇔𝐵
True True False True True True True
True False False False True False False
False True True False True True False
False False True False False True True
Semantics of propositional logic (2/2)
} A formula is satisfiable if there exists an interpretation
that makes the formula true
o (𝑃 ∧ 𝑄) ∨ ¬𝑅
} A formula whose truth table contains only false in any
interpretation is called unsatisfiable
o 𝑃 ∧ ¬𝑃
} A formula is valid if it is true in every interpretation
o 𝑃 ∨ ¬𝑃
} A model of a formula is an interpretation that makes the
formula true
o {𝑃 ⟵ 𝐹𝑎𝑙𝑠𝑒, 𝑄 ⟵ 𝑇𝑟𝑢𝑒, 𝑅 ⟵ 𝐹𝑎𝑙𝑠𝑒}
Logical equivalences (1/2)
} Formulas 𝐴 and 𝐵 are said to be logical equivalent if they
have the same truth value in every interpretation
o Notation: 𝐴 ≡ 𝐵
} Basic logical equivalences

o 𝐴 ⇒ 𝐵 ≡ ¬𝐴 ∨ 𝐵 (implication law)
o 𝐴 ⇔ 𝐵 ≡ (𝐴 ⇒ 𝐵) ∧ (𝐵 ⇒ 𝐴) (equivalent law)
o ¬(¬𝐴) ≡ 𝐴 (double negation law)
} De Morgan’s laws
o ¬(𝐴 ∨ 𝐵) ≡ ¬𝐴 ∧ ¬𝐵
o ¬(𝐴 ∧ 𝐵) ≡ ¬𝐴 ∨ ¬𝐵
Logical equivalences (2/2)
} Commutative laws
o 𝐴∨𝐵 ≡𝐵∨𝐴
o 𝐴∧𝐵 ≡𝐵∧𝐴
} Associative laws
o 𝐴∨𝐵 ∨C≡𝐴∨ 𝐵∨C
o 𝐴∧𝐵 ∧C≡𝐴∧ 𝐵∧C
} Distributive laws
o 𝐴∧ 𝐵∨C ≡ 𝐴∧𝐵 ∨ 𝐴∧𝐶
o 𝐴∨ 𝐵∧C ≡ 𝐴∨𝐵 ∧ 𝐴∨𝐶
Conjunctive normal form (1/2)
} A disjunctive clause (or simply clause) is a disjunction of
literals
o A clause has the form 𝑃1 ∨ 𝑃2 ∨. . .∨ 𝑃𝑛 , where 𝑃𝑖 are literals
} A conjunctive normal form (CNF) formula is a conjunction

of disjunctive clauses
o (𝐴 ∨ 𝐸 ∨ 𝐹 ∨ 𝐺) ∧ 𝐵 ∨ C ∨ 𝐷
Conjunctive normal form (2/2)
} We can convert any formula into an equivalent formula
that is in CNF:
o Eliminate equivalences: 𝐴 ⇔ 𝐵 ≡ (𝐴 ⇒ 𝐵) ∧ (𝐵 ⇒ 𝐴)
o Eliminate implications: 𝐴 ⇒ 𝐵 ≡ ¬𝐴 ∨ 𝐵
o Apply De Morgan’s laws to move negations close to propositional
variables
o Eliminate double negations: ¬(¬𝐴) ≡ 𝐴
o Apply distributive laws: 𝐴 ∨ 𝐵 ∧ C ≡ 𝐴 ∨ 𝐵 ∧ 𝐴 ∨ 𝐶
Exercise 1
} Prove the following basic logical equivalences using truth
tables
1. 𝐴 ⇒ 𝐵 ≡ ¬𝐴 ∨ 𝐵 (Implication law)
2. 𝐴 ⇔ 𝐵 ≡ (𝐴 ⇒ 𝐵) ∧ 𝐵 ⇒ 𝐴 (Equivalent law)
3. ¬(¬𝐴) ≡ 𝐴 (Double negation law)
4. ¬(𝐴 ∨ 𝐵) ≡ ¬𝐴 ∧ ¬𝐵 (De Morgan’s law)
5. ¬(𝐴 ∧ 𝐵) ≡ ¬𝐴 ∨ ¬𝐵 (De Morgan’s law)
6. 𝐴 ∧ 𝐵 ∨ C ≡ 𝐴 ∧ 𝐵 ∨ 𝐴 ∧ 𝐶 (Distributive law)
7. 𝐴 ∨ 𝐵 ∧ C ≡ 𝐴 ∨ 𝐵 ∧ 𝐴 ∨ 𝐶 (Distributive law)
Exercise 2
} Prove that the following formulas are valid (using basic
logical equivalences and laws)
a) (P Ù Q ) Þ P g) ¬𝑃 ∧ 𝑃 ∨ 𝑄 ⇒ 𝑄
b) P Þ ( P Ú Q)
h) (𝑃 ⇒ 𝑄) ∧ 𝑄 ⇒ 𝑅 ⇒ 𝑃 ⇒ 𝑅
c) ¬P Þ ( P Þ Q )
d ) ( P Ù Q) Þ ( P Þ Q) i) (𝑃 ∧ (𝑃 ⇒ 𝑄)) ⇒ 𝑄
e) ¬( P Þ Q ) Þ P j) ( 𝑃 ∨ 𝑄 ∧ 𝑃 ⇒ 𝑅 ∧ 𝑄 ⇒ 𝑅 ) ⇒ 𝑅
f) ¬( P Þ Q ) Þ ¬Q
Exercise 3
} Prove the following logical equivalences (using basic
logical equivalences and laws)
1) 𝑃 ⇔ 𝑄 ≡ 𝑃 ∧ 𝑄 ∨ ¬𝑃 ∧ ¬𝑄
2) ¬𝑃 ⇔ 𝑄 ≡ 𝑃 ⇔ ¬𝑄
3) ¬(𝑃 ⇔ 𝑄) ≡ ¬𝑃 ⇔ 𝑄
Exercise 4
} Convert the following formula into CNF
(𝑃 ⇒ 𝑄) ∨ ¬(𝑅 ∨ ¬𝑆)
Outline
o Logical inference
o Inference using truth tables
o Inference rules
Logical inference
} A formula 𝐻 is said to be a logical consequence of a set
of formulas 𝐺 = {𝐺*, … , 𝐺, } if in any interpretation that 𝐺
is true then 𝐻 is also true
} An inference procedure consists of a set of premises and
a conclusion
𝑠𝑒𝑡 𝑜𝑓 𝑝𝑟𝑒𝑚𝑖𝑠𝑒𝑠
𝑐𝑜𝑛𝑐𝑙𝑢𝑠𝑖𝑜𝑛
o Soundness: if conclusion is a logical consequence of the set of
premises
o Completeness: if can find every logical consequence of the set of
premises
} Notations
o KB : Knowledge Base, set of known formulas
o KB├𝛼: 𝛼 is a logical consequence of KB
Inference using truth tables
} Using truth tables can determine whether a formula is a
logical consequence of a set of formulas in KB or not
o Example: 𝐾𝐵 = {𝐴 ∨ 𝐶, 𝐵 ∨ ¬𝐶}, 𝛼 = 𝐴 ∨ 𝐵
} Properties
o Soundness?
§ Yes
o Completeness?
§ Yes
o Computational complexity
§ High
Inference rules (1/2)
} Modus Ponens
𝛼 ⇒ 𝛽, 𝛼
𝛽
} Modus Tollens
𝛼 ⇒ 𝛽, ¬𝛽
¬𝛼
} And-Elimination
𝛼) ∧. . .∧ 𝛼* ∧. . .∧ 𝛼+
𝛼*
} And-Introduction
𝛼) , . . . , 𝛼* , . . . , 𝛼+
𝛼) ∧. . .∧ 𝛼* ∧. . .∧ 𝛼+
𝛼, 𝛽, 𝛼% are formulas
Inference rules (2/2)
} Or-Introduction
𝛼*
𝛼) ∨. . .∨ 𝛼* ∨. . .∨ 𝛼+
} Double-Negation Elimination
¬(¬𝛼)
𝛼
} Transitivity
𝛼 ⇒ 𝛽, 𝛽 ⇒ 𝛾 𝛼, 𝛽, 𝛾, 𝛼% are formulas
𝛼⇒𝛾
} Unit Resolution
𝛼 ∨ 𝛽, ¬𝛽
𝛼
} Resolution
𝛼 ∨ 𝛽, ¬𝛽 ∨ 𝛾
𝛼∨𝛾
Exercise 1
} Prove the following logical consequences using truth
tables
1. {𝐴 ⇒ 𝐵, 𝐴} ├ 𝐵
2. 𝐴 ⇒ 𝐵, ¬𝐵 ├ ¬𝐴
3. {𝐴 ⇒ 𝐵, 𝐵 ⇒ 𝐶} ├ 𝐴 ⇒ 𝐶
4. 𝐴 ∨ 𝐵, ¬𝐵 ├ 𝐴
Exercise 2
} Given a 𝐾𝐵:
𝑄∧𝑆 ⇒𝐺∧𝐻 (1)
𝑃⇒𝑄 2
𝑅⇒𝑆 3
𝑃 (4)
𝑅 (5)
Prove the following logical consequence using inference

rules: KB├ 𝐺

Propositional Logic: Introduction To Artificial Intelligence

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Propositional Logic: Introduction To Artificial Intelligence

Uploaded by

Copyright:

Available Formats

Posts and Telecommunication Institute of Technology

Faculty of Information Technology 1

Introduction to Artificial Intelligence

} Close to natural language

} Basic logical equivalences

} A conjunctive normal form (CNF) formula is a conjunction

Prove the following logical consequence using inference

You might also like