The Pilot Tests of Adaptive Systems
The results of adaptive systems are omnipotent fantasies unless they have been tested. Adaptive system testing implies testing their functionality and requires a precise design of the tests. The “trial and error” use of objects is not a pilot test.
As shown in the research, pilot tests are the drivers of the unicist reflection processes. Pilot tests have two objectives:
- Falsification of knowledge
- Validation of knowledge
This research was led by Peter Belohlavek at The Unicist Research Institute.
1) Falsification – Destructive testing
Falsification, in the field of complex problems, implies finding the limits of the validity of a given knowledge. To do so, it is necessary to develop experiences in homologous fields until the limits of validity are found.
Two elements are homologous when they have the same “nature”. A whale and a dog (an extreme example) are homologous if they are considered as mammals. A dollar and a yen are homologous considering that they are both money.
These two cases demonstrate that homology can be total or partial. When the knowledge necessary to influence a reality is falsified in a totally homologous field, then it is naturally secure knowledge. The extreme condition of this example is the homology of two identical elements.
The falsification process is a destructive test for knowledge that is applied to realities with incomplete homologies. The destruction occurs when a condition is found to demonstrate the fallacy of the knowledge.
Models to falsify knowledge using destructive testing
Destructive testing needs to be the first test when dealing with complex problems. The first step of a reflection process implies projecting one’s beliefs on the external reality. This implies needing a destructive testing approach to eliminate the subjectivism that is implicit in any projection.
Destructive testing allows defining the limits of the validity of knowledge considering that there are always, on the one hand, conceptual limits and, on the other, operational limits.
The active function of destructive tests implies finding the conceptual limits which means dealing with operational and ontological benchmarking of succedaneum solutions.
On the other hand, the energy conservation function is based on finding the operational limits considering the operational benchmarking and the ontological benchmarking of substitutes.
There are different models of destructive tests:
1) Substitute Clinics
This approach implies developing a real solution, comparing this solution with its substitutes and finding out the SWOT they both generate and the response of the market.
2) Complexity Research
It implies finding the limits of the validity of substitutes based on experiencing, using acceptable preexisting secure knowledge and comparing it with the knowledge that is being falsified.
3) Ontological Reverse Engineering
This implies using the technology of reverse engineering comparing succedaneum solutions with the solution that is being falsified.
4) Succedaneum Clinics
This is the final stage before real application. It requires developing a real solution for a real problem and allowing the market to choose between succedaneum solutions and the one that has been developed.
It implies finding the SWOT the solution generates and the response of the market.
5) Real Operation
The real operation is what defines the final limits of the knowledge that is being falsified.
2) Validation – Non-destructive testing
Validation implies the factual confirmation of the validity of knowledge. Validation is achieved when knowledge suffices to exert influence on a reality in a predictable way.
The validation process is homologous to a non-destructive test in the field of material research. Validation implies cause-effect relations. Therefore, validation can only be applied to a simplified field of a complex reality.
Validation provides a reliable knowledge to operate under controlled conditions. The knowledge is valid if the conditions of the application environment are analogous and homologous to the characteristics of the validation environment.
Models to Validate a Specific Reality
The available models to validate a reality are:
- Analogical models
- Mathematical models
- Rule based models
- Scientific-empirical models
- Conceptual models
1) Analogical Models
Analogical models are the most basic way to validate a reality. The typical expression of this level of validation is “If something worked here, why wouldn’t it work in this other similar context?”
This validation concept has so many “ifs”, that there is an extremely high probability of being fallacious. Taking others’ experiences and transferring them to other contexts without a validation framework is a “random” process.
2) Mathematical Models
Empirical foundations need mathematical models to be valid.
Statistics is one of the tools that empirical foundation uses to ensure that results are reliable. Mathematical models are the foundation of empiricism.
Without mathematics, empiricism is equivalent to an analogical approach.
3) Rule based Models
Foundations are logical when strict rules are applied.
If rules are not applied, the logical approach degrades to common sense, the outcome of which also depends on chance or pure intuition.
Rule based models are the support for the unicist logic.
4) Scientific-empirical Models
Scientific-empirical models are based on mathematical applications to validate knowledge, or on an epistemological approach to falsify foundations.
They provide certainty to causal foundations. Without validation or falsification causal foundations are fallacious.
5) Conceptual Models
Conceptual models and conceptual analysis are necessary to make conceptual foundations reliable.
The possibility of building conceptual foundations does not exist if the conceptual structures of a particular reality and its context are not available.
Conceptual foundations are based on the knowledge of the structure of concepts.
Pilot tests must include both non-destructive and destructive tests. The application of destructive tests requires being aware of the concepts of the realities where this test is applied.
Knowledge is secure when its validity and its limits were found. Exceptions to this rule are universal natural laws which are “universally homologous”.