Upload
nhoc-buong-binh
View
218
Download
0
Embed Size (px)
Citation preview
7/30/2019 2Loeng
1/28
Enn Tyugu 1
Algorithms of ArtificialIntelligence
Lecture 2: Knowledge
E. Tyugu
Spring 2003
7/30/2019 2Loeng
2/28
Enn Tyugu 2
PrologProlog is a logic-based programming language, i.e. a language forlogic
programming. Itsstatements are Horn clauses.
Examples:
A program: ancestor(X,Z):-parent(X,Z).
ancestor(X,Z):-parent(Y,Z),ancestor(X,Y).
A program:
state(1,0).
state(S,T):-state(S1,pred(T)), nextstate(S,S1,pred(T)).
nextstate(X+1,X,T).
A goal: ?- state(X,3).
7/30/2019 2Loeng
3/28
Enn Tyugu 3
Prolog interpreter
prog- program to be executed;
goals - list of goals, which initially contains the goal
given by a user;
unifier(x,y) - produces the most general unifier ofx
and y, or nil;
apply(x,L) - applies a unifierx to each element of alist L producing a new list.
7/30/2019 2Loeng
4/28
Enn Tyugu 4
Prolog interpreter
A.1.3: exec(prog,goals,success)=
i fempty(goals) thensuccess( )
elsegoal:= head(goals);
goals:= tail(goals);
L: {rest:= prog;
wh i le notempty(rest) doU:=unifier(goal,head(head(rest));i fUn il
thengoals1:=(apply(U,tail(head(rest)));
exec(prog,goals1,success);
i fsuccess thenexit L f if i ;rest:=tail(rest);
od ;failure( )
};exec(prog,goals, succes)
f i
7/30/2019 2Loeng
5/28
Enn Tyugu 5
Semantic networks
Linguists noticed long ago that the structure of a sentence can
be represented as a network. Words of the sentence are nodes,
and they are bound by arcs expressing relations between the
words. The network as a whole represents in this way a
meaning of the sentence in terms of meanings of words and
relations between the words. This meaning is an approximation
of the meaning people can assign to the sentence, analogous in
a way to other approximate representations of the meaning, for
instance, how floating point numbers represent the approximate
meaning of real numbers.
7/30/2019 2Loeng
6/28
Enn Tyugu 6
Example
give
pick up
have a meeting
before
after
after
morning lunch
at the time
who?
John
who?
who?
what?
report
to whom?
me
whos?
his
John must pick up his report in the morning and have
a meeting after lunch. After the meeting he will give the report to me.
7/30/2019 2Loeng
7/28
Enn Tyugu 7
Example continued
Inferences can be made, depending on the properties of the
relations of a semantic network. Let us consider only time
relations of the network in our example, and encode the time
relations by atomic formulas as follows:
before(lunch,morning) = general knowledge
after(morning,lunch)
after(lunch,have a meeting) = specific knowledge
after(have a meeting,give)
at-the-time(morning,pick up)
7/30/2019 2Loeng
8/28
Enn Tyugu 8
Example continuedInference rules:
before(x,y) before(y,z)before(x,z)
after(x,y)before(y,x)
at-the-time(x,z) before(y,z)before(y,x)
Applying these rules, we can infere
after(lunch,have a meeting)before(have a meeting,lunch) and
at-the-time(pick up,morning) before(lunch,morning)before(lunch,pick up) etc.
7/30/2019 2Loeng
9/28
Enn Tyugu 9
Frames
1. The essence of the frame is that it is a module of knowledgeabout something which we can call a concept. This can be asituation, an object, a phenomenon, a relation.
2. Frames contain smaller pieces of knowledge: components,attributes, actions which can be (or must be) taken whenconditions for taking an action occur.
3. Frames contain slots which are places to put pieces ofknowledge in. These pieces may be just concrete values of
attributes, more complicated objects, or even other frames. Aslot is being filled in when a frame is applied to represent aparticular situation, object or phenomenon.
7/30/2019 2Loeng
10/28
Enn Tyugu 10
Inheritance
ideas
even ts states things
actions abstract things
polygons
triangles quadrangles
parallelograms
rectangles rhombuses
An essential idea developed in connection with frames was inheritance. Inheritance is a conveni
way of reusing existing knowledge in describing new frames. Knowing a frame f, one can describ
a new frame as a kind off, meaning that the new frame inherits the properties off, i.e. it will hav
these properties in addition to newly described properties described. Inheritance relation express
very precisely the relation between super- and subconcepts.
7/30/2019 2Loeng
11/28
Enn Tyugu 11
Default theories
A default has the following form
A:B1, ... , Bk--------------
C
where the formulaA is a premise, the formula C is a
conclusion and the formulas B1, ..., Bkarejustifications.
Conclusion of the default can be derived from its premise, if
there is no negation of any justification derived.
7/30/2019 2Loeng
12/28
Enn Tyugu 12
Examples
1. bird(x): flies(x)---------------------
flies(x)
2. Closed world assumption (CWA):
:not F------not F
7/30/2019 2Loeng
13/28
Enn Tyugu 13
Derivation step with a default.
A - premise of a default
C- conclusion of a default
J- justifications of default
A1.4 Default(A,C,J):
fo rB Jdo
ifderrivable(B) thenfailure f i
od;
success( )
7/30/2019 2Loeng
14/28
Enn Tyugu 14
Rules
Rules are a well-known form of knowledge which is easy to use.
A rule is a pair
(condition, action)
which has the meaning: "If the condition is satisfied, then the
action can be taken." Also other modalities for performing the
action are possible - "must be taken", for instance.
7/30/2019 2Loeng
15/28
Enn Tyugu 15
Using rules
Let us have a set of rules called rules and functions cond(p) and
act(p) which select the condition part and action part of a given
rulep and present them in the executable form. The following
is a simple algorithm for problem solving with rules:
A.1.5
wh i le notgooddo
found := false;
fo rp
rules doi fcond(p) thenact(p); found:=true f i
od ;
i f notfoundthenfailure f i
od
7/30/2019 2Loeng
16/28
Enn Tyugu 16
Decision trees
A simple way to represent rules is decision tree: a tree with
nodes for attributes and arcs for attribute values. Example:
legstwo four
handsno
yes
furryno yes
table animalfurry bird
no
monkey man
yes
7/30/2019 2Loeng
17/28
Enn Tyugu 17
Rete algorithm
Rete algorithm uses a data structure that enables fast search ofapplicable rules. We shall consider it in two parts:
knowledge representation,
knowledge management (i.e. introduction of changes into theknowledge base).
Any rule that is reachable in the Rete graph (see below) via nonemptyrelation nodes can be fired.
Rete algorithm is used in JESS (Java Expert System Shell) and itspredecessor CLIPS (both developed in NASA.)
7/30/2019 2Loeng
18/28
Enn Tyugu 18
Rete algorithm continued
Knowledge includes:
1. facts, e.g.(goal e1 simplify), (goal e2 simplify), (goal e3 simplify),
(expr e1 0 + 3), (expr e2 0 + 5), (expr e3 0 * 2),...
2. patterns, e.g.(goal ?x simplify)
(expr ?y 0 ?op ?a2)(parent ?x ?y)
...
3. and rules, e.g.(R1 (goal ?x simplify) (expr ?x 0 + ?y) => (expr ?x ?y))
(R2 (goal ?x simplify) (expr ?x 0 * ?y) => (expr ?x 0))
...
7/30/2019 2Loeng
19/28
Enn Tyugu 19
Rete algorithm continued
Knowledge is represented in the form of an acyclic graph. It is for the
presented example as follows:
root
goal expr
goal expr * expr + ...
*** ***
R2 R1
xe1e2e3
y
2
y
3
5
x ye3 2 x ye1 3
e2 5
7/30/2019 2Loeng
20/28
Enn Tyugu 20
Rete algorithm continued
The overall structure of the Rete graph is the following:
root
predicate names layer
patterns layer - alpha nodes (with one input)
rules layer - one node for every rule
beta-nodes(with two inputs)
7/30/2019 2Loeng
21/28
Enn Tyugu 21
Adding facts to Rete graph
When a fact arrives then
1. Select the predicate2. Select the pattern
3. For every relation depending on the selected pattern update
the relation (add a new line to the relation).
7/30/2019 2Loeng
22/28
Enn Tyugu 22
Rete algorithm continued
The Rete graph is built, updated and used as follows:
1. One level down from the root are placed all predicate names.
2. The next level down contains alpha-nodes for all patterns of all rules assuccessors of their predicate names.
3. Beta-nodes of the following levels down (with two inputs each) includerelations that unify with the patterns along the path from the root to thenode.
4. The paths lead finally to nodes representing rules.
5. When a new knowledge item arrives, it is placed into the correct places.Finding the places is simple and straightforward, because it is guided
by a relation in every node.6. When a goal is given, the search is simple and straightforward, because
it is guided by a relation in every node.
7/30/2019 2Loeng
23/28
Enn Tyugu 23
Rules with plausibilities
Rules can be extended by adding plausibility values to them.
Let us associate with each rule p a plausibility value c(p) of
application of the rule. These values can be in the range from 0to 1. We shall consider as satisfactory only the results of
application of a sequence of rules p, ..., q for which the
plausibilities c(p) ,..., c(q) satisfy the condition c(p) * ... * c(q) >
cm, where cm is the minimal satisfactory plausibility of the
result. When selecting a new applicable rule, it is reasonablenow to select a rule with the highest value of plausibility
7/30/2019 2Loeng
24/28
Enn Tyugu 24
Plausibilities
A.1.6
c:=1;
whi leno tgooddo
x:=cm;
fo rp rules doifcond(p) andc(p) > x
thena:=act(p);
x:=c(p)
f i
od ;
c:=c*x;i fc > cm thena elsefailure f i
od ;
success
7/30/2019 2Loeng
25/28
Enn Tyugu 25
Using a plausibility function
A.1.7
c:=1;
whi leno tgooddo
x:=cm;
fo rp rules doifcond(p) andplausibility(c(p),c) > x
thena:=act(p);
x:=plausibility(c(p),c)
f i
od ;
c:=x;ifc > cm thena elsefailure( ) f i
od ;
success( )
7/30/2019 2Loeng
26/28
Enn Tyugu 26
Classification of knowledge systems
KNOWLEDGE SYSTEMS
Symbolic
(derivability,
soundness,
completeness)
Rules
(effcient
computabiliy)
Semantic networks
(eloquence,
simplicity)
Frames
(modularity,
inheritance)
7/30/2019 2Loeng
27/28
Enn Tyugu 27
Exercise
Facts: parent(pam,bob).parent(tom,bob).
parent(tom,liz).
parent(bob,ann).
parent(bob,pat).
parent(pat,jim).
Questions and answers:
?-parent(bob,pat). yes
?- parent(liz,pat). no
?- parent(X,liz). X = tom?- parent(bob,X). X = ann
; X = pat
; no
7/30/2019 2Loeng
28/28
Enn Tyugu 28
Bibliography
Bratko, I. (2001)Prolog Programming for Artificial Intelligence.
Addison Wesley.
http://herzberg.ca.sandia.gov/jess/docs/ (Jess ja rete algoritm)
Genesereth, M., Nilsson, N. (1986) Logical Foundations of
Artificial Intelligence. Morgan Kauffmann.
http://herzberg.ca.sandia.gov/jess/docs/http://herzberg.ca.sandia.gov/jess/docs/