Information Technology Reference
In-Depth Information
v1) DENSITY(p1,
d1) ŗ
WEIGHT(p1,
v1*d1)
and
therefore
yields
VOLUME(
, d1). Regressing the conjunct WEIGHT (y, w2)
through the rule ISA (p2, ENDTABLE) ŗ WEIGHT (p2, 5) and yielding ISA(y,
ENDTABLE). Finally, since no rule consequent can be unified with the conjunct
LESS (w1, w2), this conjunct is simply added to the result expression. Therefore
the concept description of SAFE-TO-STACK(
x
, v1) DENSITY(
x
) is as follows:
VOLUME(x, v1) DENSITY(x, d1) LESS (v1*d1, 5) ISA(y,
ENDTABLE) ŗ SAFE-TO-STACK (
x, y
)
This illustration satisfies operationality criterion.
x, y
SAFE-TO-STACK(x, y)
R1:
SAFE-TO-STACK(p1, p2)
{x/p1, y/p2}
LIGHTER (p1, p2)
LIGHTER (x, y)
LIGHTER(p1, p2)
R2:
{x/p1, y/p2}
WEIGHT(p1, w1)
LESS(w1, w2)
WEIGHT(p2, w2)
WEIGHT(x, w1)
LESS (w1, w2)
WEIGHT(y, w2)
R3:
WEIGHT(p1, v1*d1)
R4:
WEIGHT(p2, 5)
{x/p1, v1*d1/w1}
{y/p2, 5/w2}
ISA(p2, ENDTABLE )
VOLUME(p1, v1)
DENSITY (p1, d1)
VOLUME(x, v1)
DENSITY(x, d1)
LESS (v1*d1, 5)
ISA(y, ENDTABLE )
Figure 9.4. Generalizing from the explanation of SAFE-TO-STACK (OBJ1, OBJ2)
Explanation-based generalization does not lead to truly 'new' knowledge, but
only enables the learner to reformulate what the learner already knows implicitly.
EBG must understand the initial description of goal concept and even if the
description is correct it is inoperable. Informally speaking, the learner can not
improve its performance by means of effectively utilizing the description, that is
to say, the concept description itself is different from whether the concept can be
used. The task of EBG system is to narrow the difference according to
transformation of initial description or to make the initial description operable.
Search WWH ::




Custom Search