University of Dar es Salaam

COICT
Department of Computer Science & Eng.

IS365: Artificial Intelligence

Lecture 3 – Knowledge Representation in AI

Lecturer: Nyamwihula, W.
Block B: Room B09
Mobile: 0784281242/0754588001
Email: [email protected]
Outlines

Knowledge representation in AI
Symbolic Logic
 Symbolic logic representation
 Formal system
 Propositional logic
 Predicate logic
 Theorem proving
1. Knowledge representation

 Why Symbolic logic

 Power of representation
 Formal language: syntax, semantics
 Conceptualization + representation in a
language
 Inference rules
2. Formal systems

 S =< A , F , A ,  >
O formal system is a quadruple
 A rule of inferenceR  of arity n is an
association:
R
R  F n  F , y =  y1 ,..., y n   x, x, y i  F , i = 1, n
 Immediate consequence  = {y , ... , y } E =   A
1 n 0
 Be the set of premises
E1 = E 0 {x| y  E 0n , y  x} E 2 = E1 {x| y  E1n , y  x}
n 1 n 1

E i ( i  0)
 An element
is an immediate consequence of a set of premises 
Formal systems - cont

 If E 0 = A (then
 = )the elements of Ei are called theorems

 Be a theorem; it can be obtained by successive

x  Ei
applications of i.r on the formulas in Ei
 Sequence of rules - demonstration . |S x |R x
 
If E =   A then can E deduced from 
x be

0 i
 |S x
 3. Propositional logic

 Formal language
 3.1 Syntax
 Alphabet
 A well-formed formula (wff) in propositional logic is:
(1) An atom is a wff
(2) If P is a wff, then ~P is a wff.
(3) If P and Q are wffs then PQ, PQ, PQ si PQ are wffs.
(4) The set of all wffs can be generated by repeatedly applying
rules (1)..(3).
 3.2 Semantics

 Interpretation
 Evaluation function of a formula
 Properties of wffs
 Valid / tautulogy
 Satisfiable
 Contradiction
 Equivalent formulas
 Semantics - cont
 A formula F is a logical consequence of a
formula P
 A formula F is a logical consequence of a set
of formulas P1,…Pn
 Notation of logical consequence P1,…Pn F.
 Theorem. Formula F is a logical
consequence of a set of formulas P1,…Pn if
the formula P1,…Pn F is valid.
 Teorema. Formula F is a logical
consequence of a set of formulas P1,…Pn if
the formula P1…  Pn  ~F is a contradiction.
 Equivalence rules

Idempotenta PP  P PP  P

Asociativitate (P  Q)  R  P  (Q  R) (P  Q)  R  P  (Q  R)

Comutativitate PQ  QP PQ  QP PQQP

Distributivitate P  (Q  R)  (P  Q)  (P  R) P  (Q  R)  (P  Q)  (P  R)

De Morgan ~ (P  Q)  ~ P  ~ Q ~ (P  Q)  ~ P  ~ Q
Eliminarea
implicatiei P  Q ~ P  Q
Eliminarea
implicatiei duble P  Q  (P  Q)  (Q  P)
 3.3 Obtaining new knowledge

 Conceptualization
 Reprezentation in a formal language
 Model theory
KB || x M
 Proof theory
KB |S x M
 Monotonic logics
 Non-monotonic logics
 3.4 Inference rules

P
PQ
 Modus Ponens Q
 Substitution
PQ
 Chain rule QR
PR

P
 AND introduction Q
PQ

 Transposition PQ
~ Q ~ P
Example
 Mihai has money
 The car is white
 The car is nice
 If the car is white or the car is nice and
Mihai has money then Mihai goes to the
mountain
 B
 A
 F
 (A  F)  B  C
 3. First order predicate logic

3.1 Syntax
Be D a domain of values. A term is defined as:
 (1)A constant is a term with a fixed value
belonging to D.
 (2)A variable is a term which may take values in
D.
 (3)If f is a function of n arguments and t ,..t are
1 n
terms then f(t1,..tn) is a term.
 (4)All terms are generated by the application of
rules (1)…(3).
Syntax PL - cont
 Predicates of arity n
 Atom or atomic formula.
 Literal

A well formed formula (wff) in first order predicate logic is

defined as:
(1) A atom is an wff
(2) If P[x] is a wff then ~P[x] is an wff.
(3) If P[x] and Q [x] are wffs then P[x]Q[x],
P[x] Q[x], PQ and PQ are wffs.
(4) If P[x] is an wff then x P[x], x P[x] are wffs.
(5) The set of all wffs can be generated by repeatedly
applying rules (1)..(4).
 Syntax - schematically
Constante Variabile Functii
a x f(x, a)

Termeni Predicate
P

Formule atomice negate Formule atomice

~P(a, x) P(a, x)

Cuantificatori Literali Conectori logici

  

Formule bine formate

 CNF, DNF

 Conjunctive normal form (CNF)

F1… Fn,
Fi , i=1,n
(Li1  … Lim).
 Disjunctive normal form (DNF)
F1  … Fn,
Fi , i=1,n
(Li1… Lim)
 3.2 Semantics of PL

 The interpretation of a formula F in first order

predicate logic consists of fixing a domain of
values (non empty) D and of an association of
values for every constant, function and predicate
in the formula F as follows:
 (1)Every constant has an associated value in D.
 (2)Every function f, of
n
arity n, is defined by the
correspondence D  Dwhere
D n = {(x1 ,..., x n )|x1  D,..., x n  D}
 (3)Every predicate of arity n, is defined by the
correspondence P : D n  {t, f }
3.3 Properties of wffs in PL
 Valid / tautulogy
 Satisfiable
 Contradiction
 Equivalent formulas
 A formula F is a logical consequence of a formula P
 A formula F is a logical consequence of a set of formulas
P1,…Pn
 Notation of logical consequence P1,…Pn F.
 Theorem. Formula F is a logical consequence of a
set of formulas P1,…Pn if the formula P1 …  Pn F is
valid.
 Teorema. Formula F is a logical consequence of a set of
formulas P1,…Pn if the formula P1…  Pn  ~F is a
contradiction.
 Equivalence
Echivalenta of quantifiers
cuantificatorilor

(Qx)F[x]  G  (Qx)(F[x]  G) (Qx)F[x]  G  (Qx)(F[x]  G)

~ (( x)F[x])  ( x)(~ F[x]) ~ (( x)F[x])  ( x)(~ F[x])

( x)F[x]  ( x)H[x]  ( x)(F[x]  H[x]) ( x)F[x]  ( x)H[x]  ( x)(F[x]  H[x])

(Q1x)F[x]  (Q2 x)H[x]  (Q1x)(Q2 z)(F[x]  H[z]) (Q1x)F[x]  (Q2 x)H[x]  (Q1x)(Q2 z)(F[x]  H[z])
Examples

 All apples are red

 All objects are red apples
 There is a red apple
 All packages in room 27 are smaller than any
package in room 28
 All purple mushrooms are poisonous
 x (Purple(x)  Mushroom(x))  Poisonous(x)
 x Purple(x)  (Mushroom(x)  Poisonous(x))
 x Mushroom (x)  (Purple (x)  Poisonous(x))

(x)(y) loves(x,y)
(y)(x)loves(x,y)
3.4. Inference rules in PL

 Modus Ponens

 Substitution
 Chaining
 Transpozition
 AND elimination (AE)
 AND introduction (AI)
 Universal instantiation (UI)
 Existential instantiation (EI)
 Rezolution
 Example
 Horses are faster than dogs and there is a greyhound that is
faster than every rabbit. We know that Harry is a horse and that
Ralph is a rabbit. Derive that Harry is faster than Ralph.
 Horse(x) Greyhound(y)
 Dog(y) Rabbit(z)
 Faster(y,z))

x y Horse(x)  Dog(y)  Faster(x,y)

y Greyhound(y)  (z Rabbit(z)  Faster(y,z))

Horse(Harry)

Rabbit(Ralph)
y Greyhound(y)  Dog(y)
x y z Faster(x,y)  Faster(y,z)  Faster(x,z)
 Proof example
 Theorem: Faster(Harry, Ralph) ?
  Proof using inference rules
1.  x y Horse(x)  Dog(y)  Faster(x,y)
2. y Greyhound(y)  (z Rabbit(z)  Faster(y,z))
3. y Greyhound(y)  Dog(y)
4. xyz Faster(x,y)  Faster(y,z)  Faster(x,z)
5. Horse(Harry)
6. Rabbit(Ralph)
7. Greyhound(Greg)  (z Rabbit(z)  Faster(Greg,z)) 2, EI
8. Greyhound(Greg) 7, AE
9. z Rabbit(z)  Faster(Greg,z)) 7, AE
 Proof example - cont
10.  Rabbit(Ralph)  Faster(Greg,Ralph) 9, UI
11. Faster(Greg,Ralph) 6,10, MP
12. Greyhound(Greg)  Dog(Greg) 3, UI
13. Dog(Greg) 12, 8, MP
14. Horse(Harry)  Dog(Greg)  Faster(Harry, Greg) 1, UI
15. Horse(Harry)  Dog(Greg) 5, 13, AI
16. Faster(Harry, Greg) 14, 15, MP
17. Faster(Harry, Greg)  Faster(Greg, Ralph)  Faster(Harry,Ralph)
4, UI
18. Faster(Harry, Greg)  Faster(Greg, Ralph) 16, 11, AI
19. Faster(Harry,Ralph) 17, 19, MP