You are on page 1of 20

HIGHER ORDER LOGICS (Handling

Uncertainty)
FIRST ORDER PREDICATE LOGIC
- Not expressive enough
 CANNOT HANDLE
 INCOMPLETE
 UNCERTAIN
 VAGUE
 INCONSISTENT KNOWLEDGE

SITUATIONAL LOGIC

 PROPOSITIONAL & PREDICATE LOGIC


 STATIC RELATIONL STRUCTURE
 CURRENT APPLICATIONS
 KNOWLEDGE ABOUT CHANGING UNIVERSE OF DISCOURSE
 EXTRA ARGUMENT
EXAMPLE
on (b1, b2, s1)

b1

b2

7on (b1, b2, s2)


event - transfer

b1 b2
NON MONOTONIC LOGIC
 FORMAL SYSTEM
DEDUCTION IS POSSIBLE
( FROM WFFS BY SYNTACTIC OPERATION )
 THEORY
formal system
+
a set of wffs that are true in some set of intended
interpretations
Theorem (of a theory)
- A formula which has a formal proof in that theory.
(T F, F IS A THEOREM OF THEORY T)
In a ‘monotonic’ logic if T P, T T’ THEN T” P
( T + proper axiom  T’)
In a non – monotonic logic addition of assertion to T may invalidate
previous conclusions.
• CIRCUMSTANCES

 INCOMPLETE KNOWLEDGE
 CHANGE IN UNIVERSE OF DISCOURSE
 IN PROBLEM SOLVING WHERE TEMPORARY ASSUMPTION

• DEFAULT REASONING (DR)

HUMANS USE DR WHEN INCOMPLETE KNOWLEDGE

EXAMPLE

OZZIE IS A PET BIRD SUBSEQUENTLY,


WINDOW OPENED
OZZIE ESCAPED
• CONCLUSION
OZZIE FLEW AWAY
NOW : CLIPPED WINGS
REVISE THE BELIEF IN HOW HE ESCAPED

• PROOF AS FOLLOWS
a) INITIAL BELIEF SET
OZZIE IS A BIRD
X: [ X IS A BIRD ^ NOT KNOWN X CANNOT FLY ]
X CAN FLY
OZZIE CAN FLY
b) LATER
OZZIE IS A BIRD
X: [ X IS A BIRD ^ NOT KNOWN X CANNOT FLY ]
X CAN FLY OZZIE CAN FLY
OZZIE CANNOT FLY
 USE OF TEMPORARY ASSUMPTION
 ARRANGE A MEETING WITH COLLEAGUES
 MONDAY
 ROOM
 EXTENDING A CLASSICAL LOGIC
 MC DERMOTT & DOYLE – NMLOGIC – DERIVED FROM
CLASSICAL LOGIC –ADDING M – MODEL OPERATOR

 MP MEANS
 P IS CONSISTENT WITH EVERYTHING THAT IS BELIEVED
EXAMPLE
(a) X: [ (X.€.BIRDS) ^ M (X. HAS ABILITY FLYING)
(X. HAS
ABILITY FLYING) ]
(b) ( OZZIE .E.BIRDS )
FROM (a) & (b) DEDUCE
OZZIE CAN FLY
BY ADDING THE FACT

OZZIE CANNOT FLY


- INCONSISTENT & HENCE RETRACT DEDUCED
BELIEF

FIRST FORMULA
FOR ALL X, IF X IS A BIRD AND IF ( X CAN FLY ) IS
CONSISTENT WITH EVERYTHING THAT WE BELIEVE THEN
WE CAN CONCLUDE THAT X CAN FLY
TMS

 Doyle – TMS maintains consistency among


a set of belief statements of beliefs –
NODES

 True – IN
 OUT- false
Example
Status Justi-1 Justi-2
IN OUT IN OUT
1 Ozzie is a bird IN
2 Ozzie can fly OUT 1 3
3 Ozzie cannot fly IN 4 5
4 Ozzie has clipped wings OUT
5 Ozzie is a Penguin IN
1. Assertion
2. Belief which is not justified
3. Belief justified by 5
4. Previous assertion Retracted
5. Assertion

If 5 were retracted then TMS retracts 3;=> 2 to


become IN.
Fuzzy Logic
 Accommodating Commonsense
Knowledge Uncertainty
 Eg. If a car which is offered for sale is
cheap and old, then it is probably not in
good shape.
 Test-score semantics
 Used to assign a meaning to a proposition
in fuzzy logic. In test score semantics-
proposition => a collection of constraints
Example - Proposition
1. Mary is brunette (woman with dark brown hair)
- Constraint on colour of Mary’s
hair
2. Most fat men are not very agile
-Constraint on number of fat men who
are agile
Procedure to give meaning
a. Identifying the variables X1,….,Xn Whose
values are constrained by the proposition
Eg. X1= colour of Mary’s hair
b. Identify constraints C1,C2,….Cn
c. Get score for each constraint
d. Aggregate partial test scores.

Eg. Proposition
Usually snow is white.
Explanatory Database
n samples S1,S2,….,Sn
WHITE USUALLY
Snow sample Degree of Proportion Degree to which
whiteness represents
‘usually’
S1 0.7
0.1 0.0
S2 0.8
0.2 0.0
. .
0.3 0.1
. .
0.6 0.2
Sn .
. .
a) =Σti b) Compute usually for x
n
MODAL LOGIC
Allows to reason with Subjunctive moods rather than indicative
mood.

Subjective statements assert


- what must be
- ought to be
- might be
- is believed to be
- hoped to be
- will be in the future
…….
…….
Temporal logic
Time-varying relations
Approaches
1) Extension of first order logic
2) Modal temporal logic
First method :
uses predicate logic to reason about time.
Second :
new logic
Lundberg
Time points as like other entities. Related to each other by
predicates.
Eg. et(t1, t2) and ss(t1, t2)

Properties using formulas


Eg.
a) x y[ et (x, y)  tpt (x)Λ tpt (y)]

b) x y [et (x, y) v (x=y) v


et(y,x) v 7tpt(x) v 7tpt(y)]
c) x y z [et(x,y) Λet(y,z)  et (x, z) ]

d) x y[et(x,y)7et(y,x)]
Age=20
Age Never Decreases Year = 1990
v w x y z [age(x,y,z) Λ
25 1995
age(x,w,v) Λ et(z,v) 
7lessthan(w,y)]
n predicates
25,20
(n+1)tharg – time

y age = 20
z year = 1990
 Linear temporal Logic ·pas ·present · future
 Branching Temporal Logic t

past now
· ·
(1) Always q 
.future
Sometimes q
(2) q->sometimes q (it rains)-> it is raining
p (it rains) -> it rained
pp (it rains) -> it had rained
F (it rains) -> it will rain
FP (it rains) -> it will have rained
•.Syntax – propositions, connections, operations.
•.Semantics -> rules (inference)
Eg: operations
x - Every next
x - Some next
Logical Axiom Schemas (Derived Rules)
Makes Formulas
Others:
• Many valued logics
• Epistemic logic
• Lambda calculus
• Derivation of Database query languages
• Reasoning methoda(eg: Abstraction)
• Closed world assumption
Conclusions:
• Incomplete knowledge
• Expressive
• Deduction while query answering

You might also like