Information Technology Reference
In-Depth Information
In our opinion, this is concerned with human cognition, which leads to signifi-
cant differences between human and machine in problem solving. Firstly, human
can use relevant prior knowledge subconsciously in problem solving, but machine
can't. For example, we notice that everyone can solve a given problem easily from
his/her familiar fields at different levels of granularities. But (s)he even cannot
solve a problem from his/her unfamiliar field at single level, not to mention at
multiple levels. Secondly, human can use their relevant prior knowledge to gen-
erate a good 'structure' or problem representation in problem solving. As we all
know that the famous story of young Gauss who gave the answer to the sum of
all the numbers from 1 to 100 very quickly, not by very fast mental arithmetic
but by noticing a pattern in the number sequence. Namely, that the numbers
form pairs (1+100=101, 2+99=101,
, 50+51=101). This example indicates
that a good structuring, or representation of the problem helps considerably. In
essence, the so called 'good' structure of a problem is a hierarchical structure
induced from it. Thirdly, J.Hawkins and S.Blakeslee [6] pointed out that there
have fundamentally different mechanisms between human brain and machine.
The mechanism of human brain is that it retrieves the answers stored in mem-
ory a long time ago, but not “compute” the answers to a problem as machine.
This indicates that human usually search a relevant or similar answer to the
solved problem from his memory in problem solving, or we can say that it de-
pends on relevant prior knowledge to solve problem for human. Maybe these can
be used to interpret why human can focus on different levels during the process
of problem solving, but machine cannot.
In this paper, the importance of hierarchical structured prior knowledge in
granular computing is stressed firstly, where prior knowledge is extended to a
broader sense, which includes domain knowledge. A nested and hierarchical orga-
nization of prior knowledge based on multidimensional data model is proposed,
it is the basis of structured thinking [7]. Secondly, multidimensional data model
was introduced into granular computing, it can represent information from mul-
tiview and multilevel, and can be used as a representation model of granular
structures. Finally, the relation between structured prior knowledge and granu-
lar structures is analyzed.
···
2P orKnow edge
In this section, we will give a brief introduction to prior knowledge, and propose
an organization of prior knowledge based on multidimensional data model. We
also point out that structured prior knowledge is the basis of structured thinking.
The main role of structured prior knowledge is to provides humans with a much
greater control over the solved problem.
2.1
An Introduction to Prior Knowledge
Prior knowledge is all the knowledge you've acquired in your lifetime. It in-
cludes knowledge gained from formal and informal instruction. Prior knowledge
