Back ... ... ...
... ... ... ... ... ... ... ... On Stream ... ... ... ... ... ...
... ... ... ... ... Forward
OntologyStream Inc.
Copyright:
2001
Value added
technology and methodology
Section 1: Intellectual Property (IP) Deployment and
Evaluation Model
OSI has algorithms and
architecture for establishing an associative mapping between
1)
a
representation of the evolution of IP citation in the Patent and Trademarks
Office database
2)
a
representation of the evolution of scholarly literature, and
3)
a
representation of the evolution of markets.
Within the architecture, users
may create a graph representation of IP and a graph representation of the
markets that use IP. The graphs are
derived from standard OSI knowledge representation methodology.
A simple neural network
makes associations between the two graph representations.
Multiple associative
"memories" are created using specific neural network architecture and
a "training set" for each memory.
A memory might be between:
"the X domain of IP" and "in the Y domain of market
development (MD)"
In such a memory, a feature
subset of the set of features of the IP is differentially linked via the
association to a feature subset of the set of features of the MD. Inferences can then be set up over the set
of associative memories, such as
"will IP with feature set {a, b, ..., x} have success in the Y
domain of MD?”
A training set is actually
what controls the inference. The
specification of the training set is left to the skill of the user. Thus, the use of the product has a reduced
liability for the contributing vendors.
Section 2: Vetting Process for Knowledge Technologies
Vetting is a process that
moves information from one level of organization to another level. We can talk of two classes of examples.
Vetting of private knowledge into a public form. Vetting occurs when national security
investigation occurs as part of background checks. Private to public vetting also occurs in the process of everyday
human interaction with other humans within communities.
The vetting process begins
with human subjective experience.
Knowledge sharing occurs within communities of practice and other types
of social units. The management of
human knowledge processes has to come to address a number of subtle issues
within the new disciplines of knowledge science. These issues are also involved in a proper understanding of the
vetting of private human experience into a public setting.
Knowledge Process Management
is built based on selected principles from cognitive neuroscience, research on human
memory and anticipation, linguistics and systems theory. The methodology descriptively enumerates the
elements (topics) of a universe of discourse and then makes these elements the
controlling construct for elicitation and sharing of human knowledge.
Vetting of technology
innovation into a market place. Vetting
occurs when venture capital groups fund the development and marketing of some
innovation or set of innovations. The
process of creating a company has a complex dimension in that the company
stakeholders have subjective views and strategies that are expressed
collaboratively within a market ecosystem.
Technology innovation can be
considered as an evolutionary process.
In this process, a selection of basic innovation (a meme variation) is
subjected to random psychological and social pressures. The expression of a meme variation as a
market-adopted technology is similar to the expression of a gene variation as
an animal living within a biological ecosystem.
Modern theories of evolution
have been applied to modeling technological innovation. These models indicate that company formation
and the adoption of innovation proceeds through a series of specific
steps. These steps are also seen in the
morphology of human and group self-image.
The linkage of the company to value chains within the market is
increased as stakeholder collaboration meets with success criterion.
OSI has a process model for soft
controlling the process of adoption of knowledge sharing technology. A standard consulting product is available
for evaluation of the technology when implemented.
Section 3: Situational
Semantic Algebra
Academic work provides a
formal means to cite other scholarly work from within the text of a
document. Citation practice is an
integral part of the academic and professional culture. The citation provides one means to link
together the elements of a virtual document.
Hypertext markup using HTML and XML naturally provides a means to
navigate within a virtual collection of documents.
The notion of a hypertext
document is currently being extended into the notion of an information portal
where the anchors of the hypertext do not need to be specified in advance. A knowledge portal provides for knowledge
claim and knowledge validation services in real time.
A fixed citation list can be
instantiated using link analysis. The
links form a topology. In this topology, the notion of nearness is defined by
the degree to which one node is cited by another node. Patent and trademark citations are used to
produce a link analysis of intellectual property. This is a structured analysis using links that are defined by the
strict rules for Patent declaration.
The structured analysis of
Patent links is a dynamic reflection of the evolution of the Intellectual
Property adoption processes. As new
Patents are applied for, the link structure within the Patent and Trademark
Office’s (PTO) database is modified.
The modifications themselves provide a generalized derivative and
trending information. This information is useful in a number of ways.
The adaptive association of
PTO link topology to a similar link structure for technology adoption produces
an inference engine for the evaluation of future technology adoption and for
the performance of existing adopted technologies.
Standard, but little used,
linguistic and semantic theory easily provides a situational linkage between
elements of a text collection.
Traditional and newly innovative semantic link analysis, such as Latent
Semantic Indexing, provides an additional value to a fixed link analysis based
on PTO and academic citation links.
Data mining technologies
provide fundamental advances in establishing very fast and complex retrieval
from distributed heterogeneous data sources.
These solutions are consistent with certain essential requirements of
knowledge technology. For example, real
advances in speed alter what can be done in a machine / human action-perception
cycle. New data mining and data base
technology enable adaptive portals at the fingertip, rather than a simple
hyper-linked virtual document.
Section 4: In-Memory database for analytic retrieval
Speed of processing adds an
additional dimension to problems that knowledge technology must solve. In traditional data aggregation and mining
processes, the necessary algorithmic processes cannot be or are not achieved in
the time required to produce positive results.
A “time-to-delivery” problem exists.
Due to the falling prices of
computer memory chips, In-Memory databases have great value. If one can assume that all necessary
information is available in cashed memory, then the fetch-execute cycle between
processor and cashed memory can be optimized.
The optimization has several dimensions:
1)
the
number of elementary function calls can be reduced to a small number, thus
making the developer’s Application Programmer’s Interface (API) conceptual easy
to work with.
2)
a
serialization of complex processes allows the developer to write simplified
code that can be benchmarked and refined as operating system independent data
engines.
3)
Scatter
– gather methodology and evolutional programming techniques can be applied
without slower disk access processes.
Various innovations are
addressing the various aspects of the time-to-delivery. Of greatest importance, and perhaps the last
to be solved technically, is the issue of correctness of the delivery.