Dealing with Adaptive Systems and Solving the Biases of Data-based AI
Data-based AI is functional in systemic environments but tends to generate biased responses when it is applied to complex adaptive environments. To avoid biases, data-based AI needs to be complemented when dealing with complex adaptive environments that are feedback dependent, have open boundaries and in which there is a consequent objective impossibility of existence of observers.
Complex adaptive systems are composed by elements that have biunivocal relationships and are integrated by conjunctions without the existence of exclusive disjunctions. This approach is managed using the double dialectical logic that allows dealing with adaptive environments. Typical application fields in business are: marketing, strategy building and root cause management in adaptive business environments.
It has to be considered that data-based AI is managed using inductive and deductive approaches without considering the abductive approach that allows dealing with systems that have open boundaries. Therefore, the use of data-based AI in complex adaptive environments might generate biased responses because in these environments small changes may produce significant consequences (the butterfly effect) and apparently big changes may produce no consequences.
One must consider, that the basic schooling systems are based on teaching inductive reasoning and mainly deductive (analytical) reasoning, disregarding the use of the abductive reasoning approach (Charles S.. Peirce 1839-1914) to manage complex adaptive environments. https://en.wikipedia.org/wiki/Abductive_reasoning
The abductive approach implies managing the concepts and fundamentals of things which has been reinstalled in the business world by Elon Musk in the “First Principle Method” https://www.youtube.com/watch?v=NV3sBlRgzTI
The unicist artificial intelligence is part of the unicist evolutionary approach that was developed to manage complex adaptive environments in the fields of social, institutional and individual behavior.
It integrates Data-based AI with Fundamentals-based AI to generate reliable knowledge for decision making, where the subjective biases and the biases resulting from the adaptive processes of the environments have been eliminated or minimized.
About the Unicist Artificial Intelligence
The Unicist Artificial Intelligence Monitor is an intelligent interface that allows organizing adaptive systems and environments and finding the root causes of their functionality. It is based on the use of the ontogenetic maps of business functions that have been researched. This monitor can work as an artificial substitute for the mental emulation of processes when dealing with adaptive environments. It is also a facilitator of the mental emulation of processes when human management is needed.
The Unicist AI Monitor is driven by a Unicist Inference Engine that is based on the rules of the Unicist Double Dialectical Logic that allows dealing with the evolution of complex adaptive systems and environments.
The double dialectical logic is an emulation of the ontogenetic intelligence of nature that drives the functionality and evolution of complex adaptive systems and environments.
Unicist Artificial Intelligence emulates the human reflection process to apprehend the concepts of complex adaptive systems and environments.
The Unicist AI Monitor allows developing solutions and learning from the pilot tests of their implementation until their functionality has been confirmed. Its intelligence allows emulating solutions based on the unicist ontological structure of business functions using the rules of the unicist double dialectical logic that allow managing the dynamics and evolution of complex adaptive systems and environments.
The monitor is an intelligent interface that defines the concepts and fundamentals of a business function as objects and establishes their relations and functionality. The system defines the value the objects produce and establishes a pilot test system to learn from the feedback until the goal of the business function has been achieved.
The solutions are based on the ontogenetic maps of business functions that define their concepts and fundamentals. This information allows defining what is needed and comparing it with the actual state to define the actions that are necessary to achieve the established goals.
The unicist artificial intelligence allows emulating the solutions of a complex adaptive system to build structural adaptive solutions. The use of the monitor requires managing the unicist strategy model, that emulates the intelligence of nature, to build maximal strategies to grow and minimum strategies to ensure results.
The main business applications are: Strategy Building – Business Intelligence – Business Process Management – Root Cause Management – Market Laboratories – Conceptual Design – IT Architecture Design – People Management – Business Scenario Building – Future Scenario Building – Business Education.
Fundamentals-based AI and Data-based AI
The unicist artificial Intelligence is integrated by fundamentals-based AI, data-based AI and a fundamentals knowledge-management system to provide functional knowledge in adaptive environments.
The fundamentals-based AI is based on the use of the ontogenetic maps of the concepts and fundamentals that drive the functionality of the process involved and the use of pilot tests to learn from the environment.
Fundamentals-based AI uses predictors to monitor the functionality of processes and as an input to the inference engine that allows defining the actions that are tested to confirm their functionality.
The data-based AI is supported by the fundamentals-based AI to avoid having subjective biases. When the quantity of data does not suffice, data-based AI is replaced by the use of non-destructive testing that provides analogous information.
Unicist artificial intelligence provides the necessary functional knowledge to manage the functionality and evolution of adaptive environments.
Unicist Artificial Intelligence in Adaptive Environments
Unicist Artificial Intelligence allows developing different types of functionalities according to what is needed. There are basically 4 types of solutions that are homologous to human decision-making processes which are being emulated:
- Descriptive function – Driven by “how” things work
- Diagnostics function – Driven by “what” is being done
- Predictive function – Driven by the “what for” of actions
- Prescriptive function – Driven by “why” things work
This function describes the knowledge that has been inferred from data, using an analogical inference model based on the inductive approach used by data-based AI. A typical application of this is the use of neural networks to define the segments of buyers of products or services.
This descriptive function produces reliable results when the fundamentals (why) of the buying processes are known and there is a knowledge of the objective of the process (what for).
It implies the integration of the descriptive function with the prescriptive and predictive functions that are driven by the fundamentals of the processes. This data-based AI approach is integrated with the fundamentals-based AI approach in order to be reliable.
This function defines the diagnostics of what is happening based on the use of analogical inferences of data, benchmarks and experiences. It is based on the inductive-deductive approach used by data-based AI. A typical application is the diagnostics of internal or external human/social problems of an organization.
This diagnostics function produces reliable diagnoses when the possible objectives of the processes are known (what for) and there are alternative solutions (why) available that depend on the results of the diagnoses.
It implies the integration of the diagnostics function with the predictive and prescriptive functions that are driven by the fundamentals of what is possible to be achieved. This data-based AI approach is integrated with the fundamentals-based AI approach in order to be reliable.
This function establishes the possible evolution based on the functionality of what is being done based on fundamental knowledge and the use of homological inferences.
It is based on an abductive process that defines the hypotheses, an inductive approach to validate their functionality and a deductive approach to transform these hypotheses into possible solutions. A typical use is its application in business strategy building.
The predictive function generates forecasts within the possibilities that can be achieved.
It requires being integrated with the diagnostics (what) function and with the descriptive function that are based on the available data managed using analogical inferences. This fundamentals-based AI approach is integrated with the data-based AI approach in order to be reliable.
This function establishes the actions that allow achieving the goals establishes within the boundaries of actual possibilities. A typical application is the solution of complex problems in adaptive environments. It is based on developing homological inferences that allow integrating the functions that need to be established with the objects that provide the solutions. To do so it is needed to access the fundamental knowledge bank to find the solution.
It uses an abductive approach to define the hypothetical objects to be used, an inductive approach to monitor the pilot test of the solutions and a deductive approach to validate that the objects provide a structural solution.
It requires being integrated with the descriptive function (How) to confirm the functionality and with the diagnostics (what) to confirm that the problems that have been diagnosed have been solved. This fundamentals-based approach is integrated by the data-based AI approach in order to be reliable.
Unicist Artificial Intelligence emulates the Human Intelligence
The purpose of conscious intelligence is the development of functional actions that allow adapting to an environment.
The use of conscious intelligence in complex adaptive environments requires using unicist reflection to apprehend the unified field of a system including its restricted and wide contexts.
This requires an action-reflection-action process to apprehend the functionality of the system considering that the individual is part of it.
Unicist Artificial Intelligence emulates the reflection process of human intelligence requiring two functions to make this possible: The learning function and the decision function.
The learning function allows confirming the functionality of actions based on the feedback of pilot tests.
Learning begins when the functionality fails. The learning function drives the resilience of the system and expands its boundaries towards a better adaption to the environment.
This learning function uses the application of the ontogenetic maps and the evolution rules to define and monitor the functionality of hypothetical actions, which are monitored through pilot testing that provides the learning of the system.
The unicist artificial intelligence based learning allows building intelligent knowledge systems to manage and monitor complex adaptive environments.
The decision-making function of a UAI approach to reality, allows making automated decisions that work as conscious decisions based on the recycling though the learning function.
An Integration of Analogical and Homological Approaches
The goal of human conscious intelligence is to deal with the root causes of things to adapt to the environment. Conscious intelligence has two possible approaches: an analogical approach or a homological approach.
The Analogical Approach
The analogical approach to reality deals with the observable facts and actions that are the consequences of the underlying concepts and fundamentals that define the root causes of their functionality.
The artificial intelligence approaches that deal with observable facts and actions can only learn empirical knowledge and make analogical decisions. This is an emulation of the human reactive intelligence approach that is functional when the goal is to develop operational reactions.
The Homological Approach
The unicist artificial intelligence approach is needed to manage homologies when the goal is to influence complex adaptive environments. Two entities are homologous when they share the underlying concept.
This homological approach deals with the concepts of things, which define the ontogenetic maps of the unified fields and establish the rules of their dynamics and evolution. The unicist homological approach includes the analogical approach but not vice versa.
The unicist artificial intelligence is based on using a homological approach that allows defining the necessary actions to influence a complex adaptive environment and measure the consequences of these actions to learn from them.
The UAI approach includes both learning and deciding actions based on a homological approach that deals with the essential concepts of functions that have been transformed into operational solutions. The feedback from the environment defines the functionality of actions or drives the learning of the system until the results become functional.
Unicist Artificial Intelligence: The Use of Predictors
Predictors are signs that can be read to anticipate the future. They are ambiguous signs that require to be read considering the conditions of the restricted and wide contexts. Predictors are observable events that make the fundamentals of specific aspects of reality observable. The fundamentals of a specific reality are able to define a concept if there is a catalyst or a gravitational force that is influencing it.
Everyone uses predictors to interpret actions. For example, a smile is a predictor of what can be expected. Non-verbal communication necessarily includes the observation of “predicting signs” in order to act or react.
The rational use of predictors requires being aware of the structure of fundamentals that rule a given reality and the external forces of the restricted and wide contexts that influence it.
It is necessary to use predictors to deal with complex adaptive aspects of reality. The unicist algorithms and the unicist ontogenetic maps provide the structure of predictors to be observed and measured to anticipate the future in order to react or exert influence to make things happen.