Transcript Real-Time Input of 3D Pose and Gestures of a User`s

Knowledge
Engineering
Knowledge Engineering
 Process of acquiring knowledge from experts and
building knowledge base

Narrow perspective
 Knowledge acquisition, representation,
validation,
inference, maintenance

Broad perspective
 Process
system
of developing and maintaining intelligent
KE Process

Knowledge representation



Organized knowledge
Acquisition of knowledge

General knowledge or metaknowledge

From experts, books, documents, sensors, files
Inferences

Software designed to pass statistical sample data to generalizations

Knowledge validation and verification

Explanation and justification capabilities
Inference Process (1 of 4)
•In rule-based expert system, the domain knowledge
is represented by a set of IF-THEN production rules
and data is represented by a set of facts about the
current situation.
•The inference engine compares each rule stored in
the knowledge base with facts contained in the
database.
Inference Process (2 of 4)
Done in three stages:


match  select  execute
Match : contents of the working memory are compared to
the
facts and rules contained in the knowledge base

Select: When consistent match found the corresponding rules are
placed in the conflict set.

Execute: When all matched rules are placed in the conflict set
one of the rules is selected for execution
Inference Process (3 of 4)
Database
Fact: A is X
Fact: B is y
Match
Fire
Knowledge base
Rule: IF A is x THEN B is y
Figure : The inference engine cycles via a match-fire procedure
Inference Process (4 of 4)
 The matching of the IF parts to the facts produces
inference chains.
 The inference engine must decide when the rules
have to be fired. There are two principal ways in which
rules are executed:
 Forward Chaining
 Backward Chaining
Inference Process: Forward Chaining
It’s the data-driven reasoning.
The reasoning starts from the known data and proceeds
forward with that data.
Each time only the topmost rule is executed.
When fired, the rule adds a new fact in the database.
Any rule can be executed only once.
The match-fire cycle stops when no further rules can be
fired.
Inference Process: Forward Chaining
Let’s see an example
Inference Process: Forward Chaining
Database
A
B
C
D
Database
E
A
X
Match
Fire
Match
B
C
D
E
X
L
Fire
Knowledge-Base
Knowledge-Base
Y&D Z
Y&D Z
X&B&EY
X&B&EY
AX
AX
CL
CL
L&MN
L&MN
Cycle #1
Inference Process: Forward Chaining
Database
A
B
C
D
E
X
L
Y
Fire
Match
Knowledge-Base
Y&D Z
X&B&EY
AX
CL
L&MN
Cycle #2
Database
A
B
C
D
E
X
L
Y
Z
Fire
Match
Knowledge-Base
Y&D Z
X&B&EY
AX
CL
L&MN
Cycle #3
Inference Process: Backward Chaining
It’s the goal-driven reasoning.
Here an expert system has the goal and the inference
engine attempts to find the evidence to prove it.
First the knowledge base is searched to find rules that
might have the desired solution.
Such rules must have the goal in their THEN parts. If
such rule is found and its IF part matches data in the
database, then the rule is fired and the goal is proved.
Inference Process: Backward Chaining
Database
A
B
C
D
Database
E
A
B
C
D
E
?
Z
Y
Knowledge-Base
Y&D Z
X&B&EY
AX
CL
L&MN
Pass 1: Goal: Z
Knowledge-Base
Y&D Z
X&B&EY
AX
CL
L&MN
Pass 2: Sub-goal: y
Inference Process: Backward Chaining
Database
A
B
C
D
E
Database
A
B
C
D
E
X
?
X
Knowledge-Base
Y&D Z
X&B&EY
AX
CL
L&MN
Pass 3: Sub goal:X
Fire
Match
Knowledge-Base
Y&D Z
X&B&EY
AX
CL
L&MN
Pass 4: Sub goal:X
Inference Process: Backward Chaining
Database
A
B
C
Database
D
E
X
Y
Fire
Match
Knowledge-Base
Y&D Z
A
B
C
D
E
X
Y
Z
Fire
Match
Knowledge-Base
Y&D Z
X&B&EY
X&B&EY
AX
AX
CL
CL
L&MN
L&MN
Pass 5: Sub-goal: Y
Pass 6:Goal: Z
Inference Process: Backward Chaining
Database
A
B
X
C
D
E
Y
Z
L
Fire
Match
Knowledge-Base
Y&D Z
X&B&EY
AX
CL
L&MN
Pass 7:Goal: L
Forward vs. Backward Chaining
Forward Chaining
planning, control
data-driven
Backward Chaining
diagnosis
goal-driven (hypothesis)
bottom-up reasoning
top-down reasoning
find possible conclusions supp
orted by given facts
similar to breadth-first search
find facts that support a given
hypothesis
similar to depth-first search
antecedents (LHS) control
evaluation
consequents (RHS) control
evaluation
THANK YOU