The
specific categorical Modelling System is the first step in the third stage in
the first phase for the construction of Specific Artificial Intelligences forArtificial Research by Application, whose first stage or comprehension stage
consists of the database of categories (a complete list or taxonomy of specific
conceptual categories regarding to some specific science, discipline,
activity), the second stage or explanation stage consists of the attributional process of categories to real objects, and the third stage or decision stage
will consist of all those process destined to make decisions according to the attribution
of categories to real objects.
This
first stage of any Specific Artificial Intelligence, by Application or by
Deduction, is in essence an application stage or comprehension stage, and the
first stage in the Specific Artificial Intelligence by Application consists of
a database of categories, as a list or taxonomy of concepts regarding to an
specific science, discipline, or activity, like if it were the encyclopedia of
that specific science, discipline or activity, for instance, the first stage or
application stage in botany should include all categories of vegetables, the first
stage or application stage in mineralogy should include all categories of
minerals. The first stage is called the application stage or comprehension
stage because the database works as if it were the application itself, and the
main objective of the application is to serve as a tool to comprehend the
world.
The
second stage of any Specific Artificial Intelligence, by Application or by
Deduction, is the replication stage or explanation stage, and the second stage
in the Specific Artificial Intelligence by Application consists of the
replication in artificial psychology of all those human psychological skills
responsible for the attribution of categories to real objects. For instance,
given any plant, the attribution of the right category within the database of
categories in the first stage to any plant in the real world, matching categories
of plants and real plants to know what type of plant, according to the specific
database of categories on botany, is every plant in the real world. This
process matching categories of botany and real plants is an explanation process
to explain what type of plant is every plant in the reality.
The
third stage of any Specific Artificial Intelligence, by Application or by
Deduction, is the auto-replication stage or decision stage, and the possible
auto-replications could be subdivided in objective auto –replications and
subjective auto-replications. In turn, objective auto-replications could be
sub-divided between real objective auto-replications and knowledge objective
auto-replications. And subjective auto-replications could be subdivided between
artificial psychological subjective auto-replications and robotic subjective
auto-replications. In any case the concept of auto-replication means self-improvement
or self-enhancement, of the reality (real), the knowledge (factual or catetorical),
the artificial psychology (software), the robotic system (hardware).
For
real objective auto-replication is understood the capacity of an Specific or
Global Artificial Intelligence to improve the reality, the real world, by
itself, without human assistance, so an Artificial Intelligence, Specific or
Global, could be able to improve a particular, specific, or global, reality,
the improvement of the conditions of a particular thing or being, an specific
field of a science, a discipline, or activity, or the real improvement of the
whole world by artificial means only.
For
instance, given an Specific Artificial Intelligence by Application specialised
in farming, given any land in any place (from the Earth or any other planet) to
decide what to culture, having an exhaustive database of agriculture this
intelligence as first stage, should be able to match what kind of plant is
better for that land according to the qualities of that land and the weather in that area, matching this
land and the weather with the most suitable seeds to grow up in that land with
that weather.
In
this last case, the specific database of categories of that Specific Artificial
Intelligence by Application specialised in farming, having included a database
of categories of plants, for every plant the category should have included a
very exhaustive description in mathematical terms of every type of plant
including a very objective description in quantitative terms of what
atmospheric and geological conditions are the most suitable for each plant in
the specific database of categories of plants in farming.
For
instance, according to the chemical composition of the farmland and the
weather, what seeds are more suitable to grow up under such chemical and atmospheric
conditions. This attributional process
matching land and seeds, is an attribution of the second stage of replication
or explanation in that Specific Artificial Intelligence by Application
responsible for the farming in that area.
Once
the Specific Artificial Intelligence by Application in the second stage has
matched the land and the seeds, the third stage consists of all the range of
decisions and intructions to make possible to seed that land according to the
attribution made on the second stage.
The
third stage as an auto-replication or decision stage, as real objective
auto-replication has as purpose to improve the reality, in this case, in order
to improve the cultivation on the farm, once the second stage has attributed
what kind of seed is most suitable for a land, the third stage must carry out
all the range of decisions and instructions to make it possible.
For
this purpose the third stage as real objective auto-replication to improve the
farming system on the farm, will carry out the process through four steps: the
first step the specific categorical Modelling System, the second step the
specific categorical Decisional System, the third step the specific categorical
Application System, the four step the specific categorical Learning System.
The
inner organization of every step, is though the organization of every step
according to the three stages, first stage of application or comprehension as a
database, second stage of replication or explanation as all artificial
psychological skills replicated from the human psychology, and third stage of
auto-replication or decision stage to carry out the decisions according to the
explanation of the reality given in the second stage.
In
the case of the specific categorical Modelling System, as first stage a
conceptual scheme, as second stage the conceptual sets, models, maps, what
means that the first and second stages are going to integrate the deep
artificial comprehension system, and as third stage the decision according to
the deep artificial comprehension system, having as a psychological hypothesis that
in order to explain something firstly is necessary to comprehend something, is
you do not comprehend something, you cannot explain it, and once you can
explain it, you can decide on that matter.
In
general the sequence in human knowledge is: comprehension, explanation,
decision; what in artificial psychology is translatable to: application,
replication, auto-replication as self-improvement of the real world, the
artificial knowledge (categorical or deductive), the hard/soft-ware,
self-improvement made by the artificial psychology itself without human assistance.
While
the first and second stage of any intelligence, global or specific, by
Application and by Deduction, consists only of one only stage, what means that
the first stage consists only of a database as first stage, and the second stage
consists only of the attributional process as replication a human psychological
skill: by Application the attribution of real objects to categories within the
database, by Deduction the attribution of sets of data from the specific,
global, or particular matrix, to a pure reason (equation).
In both cases, first stage and second stage,
as comprehension and explanation stages, by Application or by Deduction, consists
only of single stages: first stage the database as comprehension stage, second
stage the attributional process as replica of human psychology.
Instead
the way in which this sequence of comprehension-explanation-decision, as
application-replication-autoreplication will be carried out in the third stage
of any intelligence, global or specific, by Application or by Deduction, is
through the sub-division of the whole third stage in three inner stages or
sub-stages. Having an addition sub-stage more as an evaluation stage, the
Learning System is in essence an evaluation stage to improve the whole
intelligence.
While
the first and second stage of any intelligence are single stages, as
comprehension-database, explanation-attribution, the third stage is sub-divided
in three inner sub-stages plus the evaluation stage of the whole intelligence.
For
that reason the third stage in any intelligence, global or specific, by
Application or by Deduction, is sub-divided in: first sub-stage or first step
as Modelling System (the models), second sub-stage or second step as Decisional
System (the projects), third sub-stage or third step as Application System (the
implementation).
In
fact in the third stage, by Application or By Deduction, categorical or
deductive, specific, particular, or global, in the end the sequence of the
third stage is: models, projects, implementation. Models as comprehension or application,
projects as explanation or replication, implementation as auto-replication or decision. But regardless of the name, Modelling, Decisional, Application
systems, or model, project, implementation, in fact the three stages of the
human psychology are still there, but now translated into artificial psychology.
Having
as fourth addition sub-stage, the evaluation as Learning System, at some point,
when you evaluate you are assessing what you need to learn, not discarding at
any time other possible sources of learning without previous evaluation as for
instance learning by error and trial, another possible way of learning to
replicate in artificial psychology to be implemented within the specific and
global intelligences, in fact a learning that could be included in the Artificial General Intelligence.
Having
in mind that the organization of an Specific Artificial Intelligence for
Artificial Research by Application consists of first stage as database of
categories of an specific science, discipline, activity, the second stage
consists of the attribution of real objects to a category within that database
of categories, the third stage consists of all the range of decisions and
instructions to be applied according to the attribution of that real object to
that category.
If an
Specific Artificial Intelligence by Application specialised in farming, has a whole catalogue of different species of seeds and plants for farming within
the database of categories, and as second stage has to attribute what plant or
seed is better for every farmland, the third stage should consist of the
plantation of every plant and seed attributed to every farmland in the second
stage.
The
way to carry out the third stage is through the three steps as three sub-stages
within the third stage to model, project, and implement all the decisions
regarding to the plantation, plus another sub-stage for the whole evaluation of
the process. In total these three sub-stages plus the evaluation one, are the
four steps in the third stage: Modelling System (models), Decisional System
(projects), Application System (implementation), and Learning System
(evaluation).
Because
the Modelling, Decisional, Application, Learning Systems in by Application work
in a different way respect to the Modelling, Decisional, Application, Learning
Systems in by deduction, to distinguish the four steps in by Application respect
to by Deduction, the surname of every
step by Application will be “categorical”, while the surname of every step by
Deduction “deductive”, for that reason it is necessary to distinguish between: specific
categorical Modelling System (by Application), specific deductive Modelling
System (by Deduction), specific categorical Decisional System (by Application),
specific deductive Decisional System (by Deduction), specific categorical
Application System (by Application), specific deductive Application System (by
Deduction), specific categorical Learning System (by Application), specific
deductive Learning System (by Deduction).
Understanding
that the specific intelligences, by Application or by Deduction, correspond to the first phase, in the following phases, second of collaboration, third of
standardization, fourth of unification, fifth of particular programs or
particular applications, is necessary to highlight the distinction between
categorical or deductive, distinction that disappears in the third step in the sixth phase when
both models of intelligence, by Application and by Deduction are integrated
finally in the integrated Application System, according to my proposal, where the plan as product to synthesised the categorical/deductive model/Project in only "The plan" wil be postpone to the sevent phase, the reason itself.
Having
in mind that the third stage in by Application or by Deduction is not a single
stage, and can be sub-divided in sub-stages: Modelling System, Decisional
System, Application System, Learning System; in turn every sub-stage could be
subdivided in three stages, so every system is organized in three stages.
In
the specific case of the specific categorical Modelling System, the three
stages are the first stage as conceptual scheme, second stage as: conceptual
sets, models, maps; the third stage as that one to attribute decisions to real
objects according to the category attributed in the second stage by
Application.
In
a farm, the Specific Artificial Intelligence by Application for farming, in the
first stage has a very complete database of categories about farming, the
second stage the possibility to attribute categories to real objects, for
instance, the attribution of what kind of seed is more suitable for every
farmland, and the third stage should be the plantation process.
In
the third stage for the plantation firstly is necessary to have accurate models,
to make accurate projects, to plant the seeds, assessing in the end the
accuracy of the whole process.
In
this process the first part is to make models, whose responsible is, in this
case, the specific categorical Modelling System.
The
way in which the specific categorical Modelling System works is, once the
farmland has been labelled, attributed, to the right plant or seed, as first
stage of the Modelling System as deep artificial comprehension system, firstly includes
that farmland within the conceptual scheme, being the conceptual scheme the
synthesis of the database of categories organised like a conceptual scheme
where all the categories are linked each other by vectors according to logical
relations based on conceptual sets, and within every category are included
every real object attributed to that category in the previous second stage by
Application. Secondly, once the real object has been file in the right category/place
within the conceptual scheme, the first categorical check will ensure that
there is no contradiction between the place/category where the real object has
been included in the scheme and any other aspect of this place/category within
the conceptual scheme.
The
responsible to file the real object in the right category/place in the
conceptual scheme is the second stage by Application as soon has attributed the
right category, within the database of categories to that real object, once the
second stage by Application has attributed the right category, according to the
database of categories as first stage by Application, to that real object, then
the second stage by Application files that real object in the right
category/place within the conceptual scheme as first stage for the categorical
Modelling System, filing the real object in the right category/place within the
conceptual scheme according to the category attributed to that real object in
the second stage by Application.
The
different between the category/place (or place/category, the order does not
change the meaning of the sense) in the conceptual scheme as first stage for
the categorical Modelling System respect the category within the database of
categories within the first stage by Application, is the fact that the first
stage by Application as database of categories is only a list of categories or
taxonomy of categories, where every category is described in quantitative
terms, in order that the second stage by Application once has received the
measures from a real object the only
thing to do is to compare the measurements from a real object with the
list or taxonomy of categories, in quantitative terms, to match what quantitative
description of what category has more percentage of similarities with the
measurements from the real object, making the attribution between real object
and category as soon the level of similarity between the measurements of a real
object, and the quantitative description of a real object, the level of
similarity is equal to or greater than a critical reason.
If,
in the second stage by Application, the level of similarity between the
measurements of a real object, and the quantitative description of a category,
is equal to or greater than a critical reason, is considered that this category
corresponds to that real object by Application.
If,
in the second stage by Application, there are two or more categories whose similarity
respect to a real object is equal to or greater than a critical reason, then
the category to be chosen to make the attribution by Application is that one
which the greatest level of similarity.
If,
in the second stage by Application, there is no category whose level of
similarity with the real object is equal to or greater than a critical reason,
there are two options: 1) first option and the most rational the consideration
of that real object as a new category, taking the sample of measurements as a
quantitative description of this new category, 2) second option and the most
utilitarian accepting a greater margin of error than the margin of error
normally accepted for this kind of statistical decisions, the acceptation of
that category with the less margin of error respect to that real object,
although not enough to reach the critical reason, so that margin of error is
greater than the margin of error normally accepted according to the critical
reason set up for this decision on regular basis, understanding this solution
increasing the margin of error only as an exception for this situations in
which there is no category within the rational margin of error according to the
critical reason.
As
a result of the first option, the consideration of this real object not
matching within the normal margin of error, critical reason, respect to any
category within the database of categories, as if this real object were a new
category, then this new real object as new category should be included as a new
category within the database of categories as first stage by Application, and
should be included in the conceptual scheme as first stage in the categorical
Modelling System placing this new category in the conceptual scheme according
to those conceptual logical sets in which this new real object had got any
resemblance or similarity when matching this new real object in the second
stage.
If
in space exploration, a spaceship landing in another moon or planet, collects a
geological sample with some minerals which do not match respect to any other
category in the taxonomy of minerals already known, so this new real object is
treated as if it was a new mineral, and the sample of measurements of this new
mineral is treated as the quantitative description of this new mineral as new
category to be included in the database of categories as first stage by
Application, and placed in the right place/category in the conceptual scheme as
first stage of the categorical Modelling System, the way to place this new
category in the conceptual scheme according to the logical conceptual sets in
which this new mineral could be placed, is analysing when the second stage by
Application studied the similarities between this new mineral and the
categories in the database, what level of similarity (although not enough high
as to match with any category) this new mineral had with other categories, in
order to analyse what kind of logical relations, as logical sets, to be treated
as vectors, this new mineral got respect to other categories, in order to place
the new category in the conceptual scheme according to the potential logical
relations, conceptual sets, vectors, that this new mineral has respect to other
ones within the categories already placed in the conceptual scheme.
What
I have described in the paragraph above it should be the most rational option,
the first option, when a real object does not match in the second stage by
Application with any category in the second stage by Application, the
consideration of this real object as a new category including its sample of
measurements as the quantitative description of the new category in which this
new object could be catalogued, placing this new category in the conceptual
scheme as first stage of the Modelling System according to the logical
relations, conceptual sets, that this new category could have respect other
categories according to its level of similarity analysed in the second stage by
Application, although not reaching the matching level, but level of similarity
which not reaching the matching level can give information about what qualities
of this new category have similarities with other categories, although not
reaching the matching level, but enough as to understand this similarities as
part of the logical network or conceptual network to place this new category within the conceptual scheme.
For
instance if a new mineral found in another planet has a chemical composition
not identical to other minerals already included in the taxonomy, but the
chemical composition has some similarities, although not reaching the matching
level, respect to other categories, this similarities although not reaching the
matching level are going to be considered as the framework where to place the
new category in the taxonomy as first stage by Application and in the
conceptual scheme as first stage in the categorical Modelling System.
If
a new mineral is found with high percentage of some chemical, and there other
minerals with high percentage of that chemical, even being a new mineral, the
place to file the new mineral in the conceptual scheme will be that place where
are placed all the minerals which have in common a high percentage of that chemical.
This
first option, the creation of a new category based on the measurements of a new
real object when not matching with the existing categories, could be the most
rational, especially when it is applied to situations where the most important aim
is to get knowledge about the real world.
But
when the purpose of an specific intelligence by Application is not just
knowledge, but how to apply the knowledge that it already has, gathered in the
database of categories as first stage by Application and the conceptual scheme
as first stage of the categorical Modelling System, the most useful solution is
the consideration as possible categories to be matched with that real object, although
not reaching the matching level, those categories with the higher percentage of
similarity, therefore the less margin of error, respect to that new real
object, in order to work with that real object as if it had matched with that
category, knowing before hand that it had not matched fully, being only the
category with the higher percentage of similarity so the less margin of error
but not reaching the matching level. This solution is not so rational but is
more useful, is the second option or utilitarian option.
If
an Specific Artificial Intelligence for Artificial Research by Application for
farming, has to seed a farmland and has to decide what kind of seeds are much
better according to the chemical conditions of the land and the weather in that
area, and according to that information from the land there is no any category
of seeds in the database of categories as first stage by Application able to
reach the matching level in the second stage by Application as to be considered
the most suitable seed for this farmland, in this case the second option or
utilitarian option is the consideration of that seed which, not reaching the
matching level, has the higher percentage of similarity, so the less margin of error,
knowing before hand that this seed does not match fully with this land in order
to make later on the project, the specific categorical Decisional System, as
many adjustments on the project as to save all the difficulties that it could
have to plant seeds with some margin of error on that land, analysing what
qualities of the ground or the weather did not match with this seed as to make
adjustments over this qualities, and having in mind the adjustments to make the
project, to be implemented later on the categorical Application System.
In
this second option or utilitarian option, although not reaching the matching
level, that category whose percentage of similarity with the real object is
chosen as to be attributed to that real object, and according to this
utilitarian attribution (not full attribution in this case), according to the
utilitarian category assigned to that real object, the real object is filed in
the place where this category is located in the conceptual scheme, making clear
that is a utilitarian, not full, attribution.
The
importance to distinguish between: full attribution (normal attribution when reaching
the matching point not needing new categories nor utilitarian attribution), new
category (first option when not having full attribution), utilitarian attribution (second option when
not having full attribution); is because as soon the second stage by
Application has assigned a category (full, first, utilitarian) to a real
object, according to the category the object is placed in the conceptual scheme
where the category is placed in the conceptual scheme, having as network the
vectors that this category on the conceptual scheme could have respect to other
categories according to logical relations, conceptual sets.
As
soon a real object has been assigned (fully, new, utilitarian) to a category,
the logical relations of this category, so the conceptual sets where this
category is included, are the logical relations of this real object respect to
any other objects, what means that the conceptual relations of this object respect
to any other object are the conceptual relations of this category in relation
to any other category, the framework of vectors or network of vectors where the
category is inserted is automatically the framework or network where this real
object will work on the model and the project to implement instructions, to be
later evaluated.
Once
the second stage by Application has made the (full, new, utilitarian)
attribution, assigning a (full, new, utilitarian) category to a real object,
according to the category, the second stage by Application files the real
object in the conceptual scheme as first stage of the categorical Modelling
System, placing the real object in that (full, new, utilitarian) category in which
the real object has been assigned in the second stage by Application.
As
soon the second stage by Application has placed the real object in the right
(full, new, utilitarian) category on the conceptual scheme as first stage of
the categorical Modelling System, is carried out the first categorical check,
ensuring that, the (full, new, utilitarian) attribution made in the second
stage by Application is right, checking that according to the results in the
analysis of percentages done in the attributional process, the set of
quantitative qualities of this real object are in harmony per average with the
quantitative description of the rest of real objects filed on this category, admitting
only more margin of error in utilitarian attributions, but not greater than the
margin of error, although not reaching the matching level, in which the second
stage by Application admitted this category for this real object as an
utilitarian attribution.
If
once a real object is filed in the conceptual scheme, the first categorical
check does not find any contradiction, per average and within the critical
reason excepting the utilitarian attributions but not higher the margin error accepted
by the second stage by Application, between the new real object added to that
place/category in the conceptual scheme and the rest of real objects placed in
that category in the conceptual scheme, then not having found contradictions
the solution of the first categorical check is that there is no contradiction
in that fully, new, utilitarian attribution, so the work on the real object can
move on the second stage of the categorical Modelling System, which will
consist of the analysis of the logical/conceptual sets where this real object
is placed, having option to remove some, especially in utilitarian
attributions, as to make the model of that real object to locate on the
conceptual map, and according to the model on the conceptual map the third
stage of the categorical Modelling System will distribute the range of
decisions to carry out the final purpose, if the final purpose of an Specific
Artificial Intelligence for Artificial Research by Application for farming, is
to plant and take care of the plantation, then, according to the model of
plantation and the location on the map, the distribution of all the decisions
necessary to plant and take care of the plantation on that farmland where the
second stage by Application attributed that category of seeds.
