Information Technology Reference
In-Depth Information
weight, which means that the stronger the link between two concept representa-
tions, the greater the weight assigned to the path that links them. This allows an
easier mutual activation in the data processing operation. The duality of semantic
relations occurring between concepts has been taken into account in building
UBMSS systems. These relations comprise:
•
relations built on positive links, and
•
relations built based on negative paths.
Semantic relations identified based on negative paths are significantly benefi-
cial in the process of taking strategic decisions, because they demonstrate that the
greater the semantic distance between concept representations, the faster the deci-
sion can be taken that the sentence which positively links the represented concepts
is false.
The presented models of semantic networks found in UBMSS systems allow
semantic relations occurring for defined concept representations to be identified.
These models can also be used to develop the next class of cognitive categori-
sation systems - UBPAS personal authentication systems - which include biomet-
ric identification systems as an instance. UBPAS systems are founded on two
basic groups of memory examination methods, which include direct and indirect
methods.
In direct methods, the system retrieves from its memory everything that it
knows (so this is a knowledge retrieval method) or it recognises elements occur-
ring in the previously applied or learned material within the set presented to it (this
is a recognition method). In the recognition method, another possibility is as-
sumed, namely that the system should classify the analysed data, should interpret
it and reason about it. This approach to analysis processes in UBPAS systems
means that the system executes a set of actions which cannot be correctly com-
pleted without previously gained (acquired) knowledge. So UBPAS systems oper-
ate using two sets of knowledge - overt and secret knowledge. This is because the
system analyses not only the sets known to it before, but also elements that meet
definition criteria not known to it previously, without disclosing the definition it-
self, but at the same time classifying and independently identifying elements that
fulfil the definition (or do not fulfil it). This method of operation of UBPAS sys-
tems confirms to some extent the presence of secret knowledge based on two pri-
mary analysis criteria:
•
the exclusivity criterion which means that the given analysis method allows
only the knowledge that is used for executing a given task and necessary for it
to be gained: in this case, the processes of data analysis and interpretation are
limited only and solely to those of them which are to some extent independent
of other elements not connected with completing the specific task.
•
the sensitivity criterion which means that the type of analysis conducted is ex-
haustive enough to reveal the entire information held in the system knowledge
base and of which the system is in a sense 'aware': the essence of this approach
is that if the system holds only a part of the overt knowledge about the analysed
