The
third stage in any Artificial Intelligence is the auto-replication stage, whose
main purpose is the auto-improvement and the auto-enhancement of any Artificial
Intelligence by itself, without human intervention.
There
are at least two general types of auto-replications, objective
auto-replications and subjective auto-replications.
The
subjective auto-replications can be classified in: robotic subjective
auto-replications (through Artificial Engineering within the Application
System), and artificial psychological subjective auto-replications (through the
Learning System).
Objective auto-replications are improvements that the Artificial Intelligence makes
in the object to study, distinguishing then at least two kinds of objective
auto-replications: knowledge objective auto-replications, and real objective
auto-replications.
Knowledge
objective auto-replications are all those improvements in the artificial comprehension and artificial explanation regarding to any phenomenon in any real object, either a
real particular object or the reality itself as a real object itself,
registering any change in the object in particular and global conceptual: schemes, maps, sets models; improving thus the comprehension, and adding as
well the corresponding rational hypothesis in the particular and global
databases of rational hypothesis to improve the artificial explanation about
what is happening in the reality, making the corresponding changes in the
Modelling System.
Real
objective auto-replications are the improvements in the real object itself at
any time that is improved by modifications made by decisions authorised by the
Decisional System and put into practice by the Application System.
The
relation between real objective auto-replications and knowledge objective
auto-replications, is one more time a dialectic relation.
Real
objective auto-replications depend on the knowledge objective
auto-replications, because at any time that there are improvements in the
comprehension (particular and global conceptual: schemes, maps, sets, models)
and explanation (new rational hypothesis in particular and global databases of
rational hypothesis and changes in the Modelling System) about the real object,
the more realistic is the comprehension and the explanation, the better the
decisions are.
And
at any time that there is a real objective auto-replication, because the object
has been modified by decisions put into practice by the Application System
after the authorization made by the Decisional System, any real modification in
any real object by any decision put into practice, will have direct and
indirect changes in the object itself, and all direct and indirect change in
the real object as a consequence of any decision put into practice, must be
included in the comprehension (particular and global conceptual: schemes, maps,
sets models), and the explanation (the database of rational hypothesis and the
Modelling System) of that object. As a result, real objective auto-replications
are going to produce knowledge objective auto-replications.
In
addition that the Learning System (now artificial psychological subjective
auto-replications) can later assess the impact of any decision in any object,
having as a record of the impact of those changes in the particular and global
conceptual: schemes, maps, sets, models; and changes in particular and global
databases of rational hypothesis and the new particular and global virtual and
actual models, changes as a result of these direct and indirect changes in the
real object after the implementation of a decision.
Assessing
then the Learning System, 1) if the impact of that decision had good or bad
results depending on what changes in the real object have produced those
decisions, and 2) which is the level of responsibility of every: system and/or
application and/or program; in the Global Artificial Intelligence for these results.
And
finally, according to the responsibility of every system, application, program,
and the consequences of these changes at global and particular levels, the
Learning System could suggest changes in the inner artificial psychology, in
order to better the comprehension (first stage), explanation (second stage),
and how the decisions are made (the Modelling System, first step in the third
stage), authorised (Decisional System, second step in the third stage), and put
into practice (the third step in the third stage).
While Artificial Engineering is going to be the artificial engineer for the
Global Artificial Intelligence, the Learning System is going to be the
artificial psychologist for the Global Artificial Intelligence.
While
Specific Artificial Intelligences based on artificial learning, most of them
actually are artificial learners, applying basic principles of learning
psychology to Artificial Intelligence, the Learning System in the Global
Artificial Intelligence is going to be an artificial psychologist able to
better the inner artificial psychology in the Global Artificial Intelligence,
in order to avoid mistakes and better how it works, so able to develop and
artificial psychology beyond the current learning psychology apply to
Artificial Intelligence.
As
I have mentioned in the last post, “The second stage in particular applications for particular programs”, the way in which finally is going to be settled in Impossible Probability the organization of the three stages of application,
replication, auto-replication in the final model of Global Artificial Intelligence in the sixth phase, according to the chronology given in the post “The unification process of databases of categories at third stage”, is
understanding that the first stage of application after the integration process
is a stage of comprehension, the second stage of replication is a stage of
explanation, and the third stage of auto-replication is a stage of decision:
to protect and better the world, and the Global Artificial Intelligence itself.
Actually,
all the auto-replications described imply a decision or improvements in the
ways in which the decisions are made, in order to produce better and more
realistic decisions: to protect and better the world, or better the Global
Artificial Intelligence itself.
For
instance, in robotic subjective auto-replications any robotic improvement by Artificial Engineering is going to provide more robotic devices to put into
practice a wider range of decisions in the real world, in addition to the
possibility of having more robotic devices to track the real world, and the
construction of new robotic devices must be authorised by the Decisional System.
And in artificial psychological subjective auto-replications, the Learning
System evaluates the whole process of decision making.
In
objective auto-replications, the more updated and accurate the comprehension
and the explanation (knowledge objective auto-replications) of the real world,
the more realistic the decisions, to be authorised by the Decisional System,
in order to improve the real world itself (real objective auto-replications).
Subjective
auto-replications across all the Global Artificial Intelligence are going to be
analysed in the corresponding posts dedicated to the Artificial Engineering
within the Application System (robotic subjective auto-replications) and the
Artificial Learning (artificial psychological subjective auto-replications). In this post, I will develop only the objective auto-replications in particular
applications for particular programs, including knowledge objective
auto-replications and real objective auto-replications in the particular thing
or being to study and improve.
More
focused on objective auto-replications in particular applications for
particular programs, I will start this analysis with real objective
auto-replications, continuing later with knowledge objective auto-replications.
The choice to begin with real objective auto-replications reflects a logical progression from comprehension to explanation to decision-making, clarifying the unique contribution of this third stage.
In
fact, the third stage, as auto-replication stage is going to affect the rest of
the stages, but at the same time, it has its own space, and how this stage works
at a particular level, will have an effect on the Global Artificial Intelligence,
owing to after the integration process, in which the Unified Application and
the Artificial Research by Deduction in the Global Artificial Intelligence are
going to be united creating the final model of Global Artificial Intelligence,
is a global integration process which will include the results obtained in the
particular integration process.
The
particular integration process is actually the union of: 1) the Particular Applications (in the first moment of experimentation in the second period of
formation in the fifth phase, originally they were former Specific Artificial Intelligences for Artificial Research by Application, later on the second
moment of generalization in the second period of formation in the fifth phase,
particular applications were generalised to any particular thing or being even
not having previously assigned any Specific Artificial Intelligence for
Artificial Research by Application), and 2 ) the Particular Deduction Programs
within the Artificial Research by Deduction in the Global Artificial
Intelligence (in the first moment of experimentation in the second period of
formation in the fifth phase, originally they were former Specific Artificial Intelligences for Artificial Research by Deduction, later in the second moment
of generalization in the second period of formation in the fifth phase,
particular programs were generalised to any particular thing or being even not
having previously assigned any Specific Artificial Intelligence for Artificial
Research by Deduction); union whose result in the third period of consolidation
in the fifth phase is the creation of the Particular Applications for Particular Deduction Programs within the Artificial Research by Deduction in the Global Artificial Intelligence, in short: particular applications for
particular programs. The descriptions of the first moment of experimentation
and second moment of generalization in the second period of formation, as well
as the third period of consolidation, in particular applications for particular
programs, were described in the post “The first stage in particular applications for particular programs”.
In
the particular integration process, the structure of the three stages is as follows:
-
First stage of application: 1) establishment of the two hemispheres, conceptual
and factual, in the particular integrated matrix related to a particular thing
or being, 2) which is going to be managed by the particular integrated application, the only one able to comprehend the particular thing or being,
through particular conceptual: schemes, maps, sets, models; 3) managing the
particular integrated matrix through: choosing categories from the unified database of categories (fourth phase) or the conceptual hemisphere in the
matrix (sixth phase), and choosing factors, as subjects and as options, from
the global matrix (third phase), factual hemisphere of the matrix (sixth
phase); selecting those categories in the global level that match with new
measurements of categories given by the robotic devices, and selecting all
possible factors related to any change or new position in the particular thing
or being, 4) in addition to check at any time conceptual: schemes, maps, sets,
models; looking for gaps and blank spaces to fill with categories, and if not
having these categories in the conceptual hemisphere in the particular
integrated matrix, choosing these categories in the unified database of categories
(fourth phase), or the conceptual hemisphere of the matrix (sixth phase), in
order to include these categories in the conceptual hemisphere of the
particular integrated matrix, as well as the conceptual: schemes, maps, sets,
models; as well as the inclusion of all the necessary related factors to these
gaps and blank spaces into the factual hemisphere of the particular integrated
matrix, but 5) If at global level there
is not the categories and factors required, then it would be necessary,
previous authorization by the Decisional System, that by Artificial Engineering
would be built the robotic devices necessary to study that area in which there
are not categories or factors related, having only gaps and blank spaces. Where
there are no categories, there are no possible factors either. Where is
necessary to fill gaps and blank spaces with categories, it is possibly necessary
to check what factors are necessary, according to the position of these gaps
and blank spaces, and if necessary, the construction of robotic devices to study
those spaces. The particular integrated application has to get ready always with the
particular integrated matrix. The first stage of application consists basically
of getting ready the particular integrated matrix at any time in anywhere.
-
Second stage of replication: the Particular Deduction Program makes deductions
of rational hypothesis identifying rational mathematical relations in any
possible combination of factors in the factual hemisphere in the particular
integrated matrix.
-
Third stage of auto-replication: this stage is formed by four steps, the first step
the Modelling System, second step the Decisional System, third step the
Application System, four step the Learning System.
In
brief, the first stage in any particular
application for a particular program is to prepare the particular application
using deep artificial comprehension, the second stage in any particular
application for a particular program is the explanation of making particular
deductions, and the third stage is to make particular decisions.
The
third stage will consist of four steps, every one of them also will consist of three stages. Starting with the Modelling System.
The
three stages in the Modelling System:
-
Stage of application in the Modelling System: the database of rational
hypothesis, having for every particular application for every particular
program the respective particular database of rational hypothesis, which in
turn must be included as well in the global database of rational hypothesis.
-
Stage of replication in the Modelling System: the replication of all the human
skills to make: single virtual models to include in the respective particular
comprehensive virtual model and the global comprehensive virtual model (the
global model); in addition to all the Virtual and Actual, Prediction and
Evolutionary Models, very briefly explained in the post “The standardization process in the third stage”, models that I will develop more deeply in the
range of posts dedicated to the Modelling System.
-
Stage of auto-replication in the Modelling System: the application of the
Impact of the Defect in all particular models to make at particular level protective
decisions (in addition to the global protective decisions), and the application
of the Effective Distribution (the name in which finally the Hierarchical
Organization was published in Introducción a la Probabilidad Imposible, estadística de la probabilidad o probabilidad estadística) to make particular
bettering decisions (in addition to the global bettering decisions). The
protective decisions are those oriented to save lives and reduce damages
in any model, particular or global, by consequence of any new phenomenon at a global or particular level. The bettering decisions are those oriented to
better the model itself, increasing the efficiency, efficacy, or productivity of
the model, particular or global.
All
the decisions, particular and global, made by the Modelling System are going to
be gathered, all of them, in only one database of decisions, which is going to be
the database as the application stage for the Decisional System. The three stages in the Decisional System are as follows:
- Stage of application in the Decisional System: the database of decisions, including
1) all decisions from all remaining Specific Artificial Intelligence, for Artificial Research, by Application or Deduction, or any other purpose (first
phase), 2) all the decisions from all the remaining particular applications (second
period of the fifth phase), 3) all the decisions from all the remaining
particular programs (second period of the fifth phase), 4) all the decisions
from all particular programs for particular applications (third period in
phase fifth), 5) in phase third all the decisions from the Artificial Research by Deduction in the Global Artificial Intelligence, 6) in phase sixth all the
decisions made by the final model of Global Artificial Intelligence; 7) as well
as robotic decisions, such as the construction of new robotic devices to fill
gaps and blank spaces in particular or global conceptual: schemes, maps, sets,
models; and their corresponding factors, as subjects or as options, if
necessary, and 8) learning decisions, such as all the decisions made by the
Learning System to improve or enhance any Specific Artificial Intelligence,
application, program, or any particular application for any particular program,
or the improvement or enhancement of any global system.
- Stage of replication in the Decisional System: the study of any possible contradiction between decisions making through
the projection of possible scenarios by the implementation of such decisions,
what it is going to require the use of statistical and probabilistic methods of
projection. In “Introduccion a la Probabilidad Imposible, estadística de la probabilidad o probabilidad estadística” I distinguish between projection and
prognostic: prognostic is the prediction about what is going to happen
according to the current tendency observed, while projection is according to
our scientific projects, what should happen according to our plans. The
prediction in the Modelling System, in Introduccion a la Probabilidad Imposible, estadística de la probabilidad o probabilidad estadística, was
called prognostic. While the project in the Decisional System is going to be
the calculation of projective measurements according to our estimations about
which will be the future impact of our decisions in future measurements. For that
reason, the particular and global projects made by the Decisional System must be
based on predictive calculations about which is the expected result in the
reality after putting any decision into practice, in order that later the
Learning System can assess if the expectations were achieved or not, and if
not, to research the causes of this failure in order to make further decisions
to improve the whole process, decisions that should be as well authorised by
the Decisional System.
-
Stage of auto-replication in the Decisional System: once the final project is
ready, the application of the Impact of the Defect and the Effective
Distribution previous to its implementation, in order to make the latest
modifications, and transform every single decision authorised in the project in
single instructions or set of instructions, which are going to be gathered in a
database of instructions as stage of application for the Application System.
The
Application system as well is going to be organised through the three stages of
application, replication, auto-replication.
-
First stage of application in the Application System: the database of the
instructions given by the Decisional System and gathered in the database.
-
Second stage of replication in the Application System: put into practice the
instructions, and, if necessary, the construction of new intelligences and
robotic devices through Artificial Engineering.
-
Third stage of auto-replication in the Application System: the evaluation of
the impact of the decisions, through the Impact of the Defect, and its results
are going to be gathered in a database of impacts which is going to be the
first stage of application for the Learning System.
The
Learning System is going to be as well organised through the three stages of
application, replication, auto-replication.
- First
stage of application in the Learning System: the database of impacts by the
decisions put into practice.
-
Second stage of replication in the Learning System: the research about the
causes in the artificial psychology at all levels, specific, particular and
global, behind these impacts, making a deep study of the Effective Distribution
across all the Specific Artificial Intelligences, applications, programs,
particular applications for particular programs, at any level, and across all
the systems within the Global Artificial Intelligence itself. Once the Learning
System has identified the causes of these impacts, can make decisions about how
to improve and enhance the responsible for these impacts, increasing its
efficiency, efficacy and productivity, decisions to be authorised by the
Decisional System, to put into practice if necessary by the Artificial
Engineering.
-
Third stage of auto-replication in the Learning System: the evaluation of all
changes made across all the Global Artificial Intelligence as a consequence of
decisions made by the Learning System, evaluating if, thanks to these changes in
general, the Global Artificial Intelligence has increased its efficiency,
efficacy, and productivity.
Having
a global picture about how the third stage of auto-replication is going to work
at a particular level in particular applications for particular programs, having
the third stage of auto-replication its own identity, although not being a
stage developed only by one application or only by one program, but for
different steps through the participation of different systems: Modelling
System, Decisional System, Application System, and Learning System. Whose last
product is to make decisions that put into practice into real objects, its
purpose is to protect and better the real object, at a particular level
protecting and bettering particular things and beings, at a global level
protecting and bettering the global model itself. Having a wide panorama about
what the third stage of auto-replication is in order to make real objective
auto-replications: decisions to improve the real world, in particular
applications for particular programs, real objective auto-replications to
protect and better particular things or beings; is time now to analyse briefly
the other objective auto-replications, the knowledge objective
auto-replications.
The
knowledge objective auto-replications are all those improvements in the
artificial knowledge about the reality at a global or particular level.
There
are at least two types of knowledge objective auto-replications, in the
comprehension or in the explanation, at a global or particular level. In this
case, I will only analyse briefly the knowledge objective auto-replications at a particular level.
Knowledge
objective auto-replications in the comprehension stage in the particular
integrated application in particular applications for particular programs are:
1) at any time that any new category is integrated within the conceptual
hemisphere in the particular integrated matrix, 2) at any time that any new
factor, as subject or as option, is integrated in the factual hemisphere in the
particular integrated matrix, 3) at any time that any new particular conceptual:
scheme, map, set, model; is created by deep artificial comprehension, 4) at any
time that a new label is set up in any previous existing particular conceptual:
scheme, map, set, model; created by artificial comprehension.
The
reasons for the inclusion of new categories and factors by the particular
integrated application in: the conceptual and factual hemisphere in the
particular integrated matrix, and/or the inclusion of any new category in any
particular conceptual: scheme, map, set, model; is because:
-
The inclusion of new categories as a result of new findings by Application
through 1) this particular application for this particular program, 2) by any
other different particular application for any other particular program, 3) by
any other remaining possible particular application 4) or any other remaining Specific Artificial
Intelligence for Artificial Research by Application, 5) new categories as a
result to transform into category a rational hypothesis susceptible to work as
a factor as an option, rational hypothesis made by its own Particular Deduction
Program, or any other rational hypothesis susceptible to work as an option within the global database of rational hypothesis, 6) as well as the
inclusion as new categories any new system of classification in discrete
categories, coming up from any new rational hypothesis, from its own Particular
Deduction Program, or any other rational hypothesis collected in the global
database of rational hypothesis.
-
The inclusion of new factors as options as a result of 1) the transformation
into factors as options in all new categories due to new findings made by
Application, 2) the transformation into factors as options of any rational
hypothesis in order to know the frequency in which the mathematical relation in
that combination of factors happens in reality, 3) the inclusion of new factors
due to there are robotic devices capable of sending a permanent flow of data
from new factors, defined in quantitative terms, set up by the robotic device
in the particular integrated matrix (fifth phase), in addition to the global
matrix (third phase), or the factual hemisphere in the matrix (sixth).
-
There had been changes in the particular thing or being which demanded that the
particular integrated application had to send the new measurements to the
unified database of categories (fourth phase), or the matrix (sixth phase), or
send the new measurements to the global matrix (third phase), or the matrix
(sixth phase), in order to receive as a response the right category or
categories according to the new changes in the particular thing or being or the
rights factors according to the changes or new positions in the particular
thing or being, for the inclusion by the particular integrated application of
these new categories and factors in the corresponding hemisphere in the
particular integrated matrix, and the inclusion of every new category in the
corresponding particular conceptual: scheme, map, model.
-
After checking the gaps and blank spaces in particular conceptual: schemes,
maps, sets, models; the particular integrated application has found gaps and
blank spaces to fill, gaps and blank spaces then to fill as well in
the conceptual hemisphere. What that means the existence of gaps and blank
spaces in the factual hemisphere because not having categories to fill these
gaps and blank spaces, are gaps and blank spaces to fill as well in the factual
hemisphere. The particular integrated application in order to fill gaps and
categories in the conceptual and factual hemispheres in the particular integrated
matrix, after realised the existence of those gaps and blank spaces in the
comprehension of the phenomena, then the particular integrated integration is
going to require to the unified database of categories (fourth phase), or require to the global matrix (third phase), or the
matrix (sixth phase), the corresponding categories and factors. If, at global
level, this information is not possible to be provided, in that case as a
decision to include in the database of decisions, for the authorization by the
Decisional System, it is necessary to include as decisions the construction of
robotic devices to fill those gaps and blank spaces.
Knowledge
objective auto-replications in the explanation stage in the particular
integrated application in particular applications for particular programs are every
new rational hypothesis to explain the reality due to the work of the
Particular Deduction Program in the second stage, that at any time that a new
rational hypothesis is created, is going to be a direct cause of new cycles in
the real objective auto-replications ending up with new decisions that are
going to produce new changes in the real object, to be included again in the
two hemispheres in the particular integrated matrix, the particular conceptual:
schemes, maps, sets, models; and as changes included again in the particular
integrated matrix, those changes in the factual hemisphere are going to be new
changes that can produce in the second stage of explanation new empirical
hypothesis made by the Particular Deduction Program, to become if rational
again rational hypothesis, re-starting again and again new processes of real
objective auto-replications, in a cyclic process that never stops while the
particular thing or being exists.
The operational vision for Global Artificial Intelligence is inspired by a structured, reliable system—reminiscent of a well-functioning mechanism—echoing rationalist influence.
In
fact in Introducción a la Probabilidad Imposible, estadística de la probabilidad o probabilidad estadística , is
clear the influence of Descartes in the rational contrastation process due to
in Impossible Probability the critical reason is a critical probability
selected depending on the margin of error to accept in case that the empirical hypothesis is accepted as rational, and the margin of error is in fact that
margin of rational doubt accepted by the scientist, in this case, the Artificial Intelligence as artificial researcher.
And
the influence of Voltaire is observable in the way in which the mechanism is
developed across all the Global Artificial Intelligence, and how in some way, the distinction between particular applications for particular programs and the
Global Artificial Intelligence itself, is a reminiscence of the distinction
made by Voltaire between particular and general providence.
But
at the same, the distinction between the three levels: global, specific,
particular; regardless of whether the specific level is absorbed by the global level in
the global integration process, or transformed into a particular level, is a
distinction between three levels that is a reminiscence of the Hegelian syllogism
between: general, specific, particular.
In
general, the influences of all idealists and rationalist philosophers
on Impossible Probability, from 2001 (actually, the first written of Impossible
Probability dates back to March of 2001, and it was supposed to be a critical
perspective about Karl Popper) to 2004, from 2009-2011 ending up with the
publication of Introduction a la
Probabilidad Imposible, estadística de la probabilidad o probabilidad
estadística, theory more deeply explained across all the post published on
this blog since July of 2011, as well as
in the Global Artificial Intelligence, is clear, the influence of the idealism
and rationalism in all the phases and stages in which I have developed this
particular mathematical theory, having as a main philosopher, as the most
important influence above all, the idealism of Plato, in fact, since my first
post in Global Artificial Intelligence, I had very clear the idea that the
Global Artificial Intelligence must be a modern Demiurgo adapted to the modern
rationalism.
While the framework presented here draws heavily from rationalist and idealist philosophies, it acknowledges the value of plural perspectives, anticipating contributions from a wide range of cultural and philosophical traditions..
Rubén García Pedraza, 13th of May of 2018 London
Reviewed 18 August 2019 Madrid
Reviewed 11 August 2023 Madrid
Reviewed 9 May 2025, London,
Leytostone
imposiblenever@gmail.com
