Elsevier UK Jobcode:RTF Chapter:CH09-H8304 22-3-2007 5:11p.m. Page:158 Trimsize:165×234MM
Fonts used:Ocean Sans & Sabon Margins:Top:42pt Gutter:54pt Font Size:10/12 Text Width:30p6 Depth:47 Lines
9 Using knowledge engineering for
risk control
9.1 Knowledge engineering, object knowledge and
metaknowledge
Knowledge engineering is a branch of science that aims to investigate, construct and
use models incorporating human reasoning. Most particularly, knowledge engineering
capitalizes on talent by exploiting and mapping rules based on the way domain
experts conceived their work and make their decisions. Its artifacts are models, like all
others studied in Chapters 7 and 8, with reasoning being their value differentiation.
Knowledge engineering models emulate the know-how of experts in the domain
of activity for which the artifact is built. The reader will recall from the preceding
two chapters that models have locality. This is even more pronounced when the
know-how and judgment of experts are reflected into the model, notably:
Skill to simplify a complex problem,
Perceptual ability and attention to detail,
Conceptualization of interdependencies and of changes in sequence, and
Domain knowledge dealing with tradeoffs and with conflicting goals.
At least in principle, like the expert the knowledge engineering artifact exhibits a
conceptual capability. This is a necessary complement, of deeper knowledge of an
event, issue, process or other object, because it permits accumulating the nebulous
concept we call experience. Contrary to the other models we studied in preceding
chapters, the knowledge artifact is able to learn from the work it is doing.
Is this intelligence? The answer cannot be crisp, for the simple reason that there
is no generally accepted definition of this term. A significant number of scientists
suggest that intelligence is no unique property of the human brain. It can also be
found in sophisticated man-made constructs at least at its lower layers of memory,
calculating ability and means for inference defined in Chapter 8.
If the definition of intelligence is limited to dynamic construction of a knowledge
bank on the fly,
Then knowledge artifacts, whose development benefits from expert knowledge,
possess this characteristic.
By contrast, at current state of the art, knowledge engineering artifacts are not known
to have imagination. Nor do they have wisdom, defined as understanding of sequence
Elsevier UK Jobcode:RTF Chapter:CH09-H8304 22-3-2007 5:11p.m. Page:159 Trimsize:165×234MM
Fonts used:Ocean Sans & Sabon Margins:Top:42pt Gutter:54pt Font Size:10/12 Text Width:30p6 Depth:47 Lines
Using knowledge engineering for risk control 159
of thoughts and actions, and their aftermath, as well as selection and ordering of these
sequences. Nor can they abstract from real world situations. Wisdom, imagination
and abstraction require the higher levels of knowledge of a sophisticated personality.
Figure 9.1 presents in a nutshell a classification of notions the preceding paragraphs
brought to the reader’s attention, taking as an example the development of man-
made systems. Modelling each of the layers in this figure poses specific requirements.
However, both lower sophistication models and the higher-up layers of knowledge
artifacts follow the rule that the main objects of computing are:
Foresight,
Insight,
Analysis, and
Design.
WISDOM
FACTS, STATES, VALUES
(DATABASE)
ABSTRACTION,
IDEALIZATION
QUALITATIVE
OUTPUT
DEPENDENCY
RULES
EXPERT’S INPUT
(KNOWLEDGEBANK)
QUANTITATIVE
OUTPUT
SIMULATION RULES
CALCULATING ABILITY
KNOWLEDGE
ENGINEERING
ARTIFACTS
TYPICAL
MODELS
CLASSICAL
COMPUTER
PROGRAM
Figure 9.1 A classification of successive layers in development of man-made
systems

Get Risk Management Technology in Financial Services now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.