The Knowledge
Sharing Foundation
General Systems AS-IS Model
of federal procurement of advanced Information Technology
Authored by
January 1st, 2003
(See also)
General
systems theory can be applied to the process of evaluating and deploying
technology in support of intelligence operations.
In many competitive markets the open competition
between suppliers of product works well.
But traditional commercial non-disclosure agreements and the sensitivity
of real-time data limit the competition to a few trusted partners.
To understand the knowledge sciences it is
necessary to understand what is possible from computer based analysis and
synthesis of structured and semi-structured data. This type of innovation has had difficulty in transfer into
national defense information technology because the procurement cycle is simply
too long in duration.
The
procurement situation is exasperated by the nature of the transfer of basic
innovation from academic centers, and from individual innovators, into
intellectual property. In traditional
business models, vendors require IP and non-disclosure (trade secrets) as a
re-requisite to creating commercial product.
But conflicting patents, and patents on algorithms (mathematics) that
have significant prior published history make objective evaluation difficult.
Mapping
the patent space surrounding computational intelligence and comparing this map
to a map of the academic literatures reveals the nature of transfer
issues.
One
can see that a few issues dominate all others in slowing the pace and reducing
the quality of innovation transfer.
One can, and should, observe that a class of specific types of values is lost in the transfer of innovation to a product. General systems analysis of the transfer is needed in order to understand what this class of values is, and how to adjust the procurement process so that the government may reduce the loss.
The present rules of procurement take too long and are too rigid. This process is not allowing the most relevant
methods to be identified.
Educational material is not developed by academia so that users can
hope to understand these methods.
Core methods are not deployed within a software infrastructure where
the core systems are available in an agile and flexible fashion.
An
mismatch has grown between intelligence communities use of vendor software and what
is possible given both educational processes and software functionality that is
essentially based on pure mathematics (neural networks, genetic algorithms),
machine representation of taxonomy and ontology, and formal logics.
This
is a natural consequence of vendor control of core mathematics and
methods.
General
systems models of the procurement process reveals that the procurement process
has created a number of distinct communities.
Each of these acts to maximize narrowly defined group interest. Most often the groups’ interests are
expressed as financial compensation.
Most often groups avoid actions that benefit other groups because the
capacity to compete on future proposals might be enhanced and might then
restrict financial compensation from these future proposals. This is
understandable and is often healthy, but also one needs to understand some of
the consequences of locally expressed self-interest, with respect to the global
interests.
Thus,
something like the stratified economic theory of Nash can be found to be absent
in the overall process that governs the behavior of these communities. The global value that might develop from
high quality transfer of innovation into adapted technology does not occur, due
to the local interests of each competing community. Sometimes, the correct products are not developed and when
deployed do not work.
No one wants to see our intelligence community
fail again in managing real time information about the threats from
asymmetrical warfare. But it is not
clear how to adjust community behavior in an environment where competition
overwhelms collaboration almost immediately.
Stability
has settled around a specific process that accounts well for the self-interest
of the various communities, but accounts poorly for global needs of the
intelligence community. The global need
is too difficult to meet, and there is too much money in the system. The system stability has a greater
propensity to maintain current behaviors than to make needed adjustments.
The
problems, that must be better addressed, are related to informational awareness
of the real time processes that are occurring in support of asymmetrical
threats. We have to see and understand
the activity of an enemy that is using every means possible to remain
stealthy.
The global need is for clear, agile and interoperable information systems where advanced computational intelligence (data mining) tools are un-encumbered by vendor business models and users have a liberal understanding of the limitations and features of a set of core tools.
Part
of elevating the process comes by understanding the way problems are
entrenched, and finding a by-pass that changes the system in a positive
fashion.
We
have the possibility of a different type of deployment model that has two
components
1) An university based educational component that provides a liberal arts understanding of the history and principles of those areas of computer science, cognitive science and general systems theory that one might suppose must inform the decision about technology evaluation and deployment. Scientists would develop a curriculum.
2) The core tools that this curriculum reveals will
be made available within the deployment
The Knowledge Sharing
Foundation