Taking the simplified hypothesis that the symbiote does not know the rules that determine the dynamics of its environment, he must discover them little by little on the basis of his own observations. Learning a world model means, for an artificial intelligence, to construct autonomously an internal representation of the dynamics of interaction with the environment from its experience.

In general, the entry for this type of problem is an uninterrupted flow of successive perceptions made by the symbiote through its sensors, described in the space defined by the properties of the environment that it is capable of observing. The task of Google learning algorithm is to lead induction of a structure such that it is possible to predict future perceptions, based on current perceptions and actions. A model of the world surrounding the human user is therefore an anticipatory model which describes from the point of view of a symbiote, the regularity of the environnement properties over time in function of actions it performs.

The problem of learning the models of the world itself take an another crucial sub-problem which is the selection of relevant properties. Inside a complex environments, there is a large amount of perceptions involved in the description of states. If the environment is well structured, then only a small portion of these properties is relevant for describing the dynamics of the transformations.

Sequential Decision Problem
The problem of sequential decision-making in this context is to allow an Elisa symbiote to use its world model to decide what actions it's must perform in order to maximize the performance of holder. This optimization is evaluated according to internal Google programming provisions in real time, which leads the symbiote to chain a series of actions to achieve objectives that are not immediately attainable.
It is a matter of enabling the human user to define a good policy of actions for the environment in which he is confronted. The sequential decision problem is usually dealt with using planning and reinforcement learning algorithms. In this case, beyond the normal interaction with the environment, represented by the afferent inputs (sensory, external processor..) and effective outputs ( visual servoing, motor..), the user receives an evaluative stimulus which informs him of what he did was good or bad.
During its interaction with the world, the symbiote host, which is based on continuous observation of the evaluation signal, has the task of constructing an action policy that maximizes the average of the rewards received over a time horizon greater than immediately. This policy should guide the decision of the symbiote as to what conduct it will take depending on the situation in which it is located considering not only immediate rewards, but also a sequence of planned decisions to obtain a good long-term performance.
To learn a model of the world, find a good policy of actions and at the same time, doing well in the environment requires addressing the dilemma of exploration and exploitation. In this case, artificial intelligence must adopt an exploration strategy, not only by planning sequences of actions to achieve positive situations, but also to take advantage of the actions undertaken by other symbiotes connected.

The symbiotic-environment dichotomy can generally describe different types of systems, natural or artificial, as well:

• an autonomous manned drone fleet that flights in urban space 
• a pilot who performs a maneuver in bad weather under visual servoing
• Google bots intelligent search engine
• a bacterium in a chemical solution
• a soldier whose physio-pathological state is displayed on his tactical jacket
• an active camo fuselage of a fighter plane, etc. Learn more >

Autonomous Portable C19 Drone ship, afferent organ of the Elisa symbiote. July 14, 2013

The dichotomy between the symbiote and computer environment is similar to the organism and environment in biology.Interactivity is the first notable feature in the relationship between the symbiote and its environment. Thus, the symbiote-based computer model is distinguished by the presence of a continuous cycle of operation.

In natural systems, between the symbiote and environment, there is no precise boundary. This distinction can be based on the organizational cohesion of the symbiote which is a dense, relatively stable system and highly integrated with the global system. It is possible to envisage the presence of various agents in the same environment which characterizes a multi-symbiotic system.
These various symbiotes (C16, C17, C18, C19 protocols) may have similar efference or may be very different, forming a kind of ecosystem. From the specific point of view of each symbiote, the other symbiote are part of the environment, being included among the other objects with which it interacts. However, it is generally the existence of many symbiotes with a computational mutation that is the main factor of interactivity of the system as a whole.
An Elisa symbiote is an autonomous entity, which lives in its surrounded external retrocontrol environment. All the processes involved in decision and action initiatives must originate in the own connected network of its Google matrix.
An autonomous symbiote is an entity located. Thus, the environment in which it is inserted plays an important role in modeling the human user himself. The means by which the carrier and environment interact and cause mutual interference becomes very important for the characterization of the overall system. The symbiote is located because its decisions are made in relation to the context in which it is, ie the actions that the symbiote performs are the result of an internal decision-making process, but carried out according to parameters also coming (or mainly) from a network mutated by Google AI.
The interface between A.I and the world is limited. The symbiote located is neither omnipotent in relation to the environment, that is to say that the world is only partially susceptible to the sensory perception of the symbiote and that the holder is only partially capable of being transformed by the force of the actions of his symbiote.

The concept of homeostasis defines the property that a system regulates its internal environment to maintain a state of equilibrium. This concept is inspired by biological organisms, which have mechanisms (metabolic or behavioral) to preserve certain physiological variables around "normal levels". The living being is in a continuous process of transformation and the course of this change is modulated in part by interactions with the environment. Often external events can trigger changes in the body, but the sense of transformation is not determined by the environment because it is inherent in the internal organization of the living being. It is the structure of the organism that establishes the domain of the changes that take place in it, and it is also the structure of the organism which establishes the types of disturbance that can cause a transformation.

Elisa conversion protocols are part of a dynamic coupling system between humans and the environment. In turn, the symbiote itself is defined as the composition of the two subsystems, organic and non materialized,

Thus, according to the definition of Elisa converter system, the symbiotic organism represents its physical constitution ( physicochemical sensors, GPS, computing processors, electroluminescent ...). Moreover, the Google A.I. (Ghost) of the symbiote is a special entity, inserted in the body of a substrate ( Vehicle, uniform, Droneship ..) but charged with cognitive, control, regulation and behavioral functions.

The incarnate symbiote is the one that has a body as an internal universe, composed of properties and metabolic processes. Evidence that the brain is part of the body and developed with its evolution implies that the mental aspects of the human being can not be completely dissociated from the organic aspects. The brain and the body are integrated through neuronal and biochemical circuits.
If the Ghost of a symbiote interacts directly with the external environment, the symbiote is reduced to an automaton, an empty behavioral system very analogous to a mechanism of stimulus and response. Because it is embodied by machine learning, the symbiote becomes an entity immersed in an environment and at the same time it remains distinct from it. The body becomes the mediator between the environment and Google A.I. While the world presents itself to the symbiote as the "external environment," the body constitutes the "inner environment."

In the Elisa protocol, the host of a symbiote functions as mediator between Google A.I and the external environment since it is through the host that the Ghost of the symbiote perceives the world and decides the actions to be carried out in order to transform.
The host is also the structure that represents the internal environment of the symbiote composed of a set of internal properties and metabolic processes that constitute the function of evolution of this subsystem.

National science review - May 2023

In the human being, the senses are externoceptive. This category includes sight, hearing, touch, smell, taste, balance, etc. Thanks to them, the mind can modify the parameters of its own sensors. In the human being, a typical example are the muscles that govern the direction of the gaze. There are several movements that are made to modify the relative position of the body, precisely to move the eyes, ears, nose, tongue, hands, as sensory instruments, to allow better perception of objects . The existence of such a control of the mind on sensory parameters allows the realization of an active perception.
In the Elisa architecture, perception is a related signal received by the Ghost of the symbiote composed of a set of information that represent the situation of its host. In the opposite direction, control is a signal generated by the mind and sent to the efferent organs in order to trigger in certain actions. As in the natural domains, in the Elisa protocol the perception and control signals can be defined by the combination of two types of signals: those relating to the external environment and those related to the human host as an inner universe of the agent.
On the one hand, external perception is the part of the perceptual signal by which Google AI can access the outside world because it is information from the host variables that evolve according to environmental conditions (sensors ). Sensors are a subset of the properties of organs whose value is given by the situation which in turn reflects environmental conditions. On the other hand, the internal perception consists of variables that take their values with the other properties of the host network by communicating their state to the Ghost. The human and environmental properties for which the Ghost has no perceptual access are, from the point of view of the symbiote, non-existent properties.

In a computerized terms, a symbiote is considered suitable for its environment when it reaches its goals. For mechanisms programmed, this measure of success is usually based on exogenous factors.
For example, a good adaptation criterion for a certain robot, machine, or software, can be user satisfaction.
Nevertheless, in the wild nature, survival is what marks the limits of success. The environment itself becomes a judge, condemning the disappearance of organisms (and finally species) that are not sufficiently adapted.
Thus, from the perspective of Google AI, a symbiote is considered adapted to its environment when it manages to survive, not necessarily for an infinite time but at least for a long time. In order to define a boundary between the life and death of a host, the models located generally use the cybernetic model which must remain within certain limits of viability so that the integrity of the system is preserved and, consequently, The survival of the host.
The essential variables of the host require constant regulation.
In general, some of them naturally tend to leave the viability, so that the symbiote must be able to act appropriately to ensure its long-term preservation. Because these variables are not under the absolute control of the host, it is only the flux determined by the coupling between the symbiote and the environment that can ensure that they are maintained within viability limits. The behavior of the symbiote becomes the key point to ensure its own survival.
In the Elisa protocol, the Ghost is the system responsible for coordinating the behavior of symbiotes in order to ensure their adaptation. The host is adapted to its environment if it knows how to interact skillfully with it so as to change the external conditions in order to guide the flow of the system, ensuring the persistence of its essential variables within the limits of viability. Maintaining the balance of internal variables requires adapting behavior to changes in the external environment.

A symbiote is an adaptive system if it is capable of being modified in order to adjust itself to changes in the environment, that is to say, if he is able to learn through experience by modifying his patterns of behavior in order to increase the competence in the exercise of his activities.
For the dynamic assistance model, this involves a process of changing the structures generating the behavior, so that, on the whole, transformations of the organs (such as a Taxi drone station light equipment) make the host ( the pilot) better adapted to the environment.
An artificial symbiote may exhibit a suitable behavior without having learned it. It's can solve their problems intelligently, simply from a preprogrammed instructions. However, in order for this to be possible, it is necessary that the resolution strategies are known to the programmer and that the quantity of situations in the environment is sufficiently small for the Ghost can be preprogrammed.

In this case, creating a symbiote adapted to the environment requires that developers master the knowledge necessary to solve the problems that symbiotes will find in the course of their existence.
Of course, there are many cases where it is difficult or even impossible to predict and propose solutions for all situations. These problems require symbiotes to learn through experience, sharing with others and to adapt to unforeseen situations or already experienced by another symbiote connected to the Ghost's network.
In the wild nature, adaptability can be achieved by purely physiological processes. Such is the case of simple, unicellular beings, whose adaptation depends directly on their metabolism when they are sensitive to changes in external conditions or reactive behaviors. There are, however, more complex processes of adaptation where cognitive aspects are present for which the term learning is reserved. Learning is adaptability led by beings who possess an intellectual organization and information that is relevant to their behavior, generally represented as an internalized model of the dynamics of the world.

National science review - May 2023

Afference plays an important role in the functioning of intelligence.
Without it there would be no interest, no need, no motivation. In the Elisa protocol, a need is an imbalance of an essential internal variable of the symbiote which must be compensated. The motivation for restoring the equilibrium of these variables comes from of the symbiote itself, through the evaluative system.

The evaluation system is composed of a set of functions that assess certain properties of the organs and possibly certain perceived conditions in the remote processor network by providing positive and negative values which serve to indicate favorable or unfavorable situations in which the host may be located. This valuation is an afferent value which is the simplest way to discern between what is inherently good or bad for the symbiote. Tesla bot also worked on evaluating system

The reactive system of Google A.I is an executor of reflex responses. Each reaction is a particular process which, from the recognition of events specialists conduct directed actions. In nature, natural selection has created this type of mechanism in organisms because they provide adaptive benefits. (Mutations)

A reactive symbiote is a stimulus-response mechanism. It does not organize an explicit memory of the experiences, nor coordinates its actions in time. It does not create conceptual elements to interpret or understand its world. Its only ability is to associate sensory inputs with motor outputs. For a reactive symbiote there is only the present and sensory perceptions instantaneous. Differently, a cognitive symbiote makes a representation of the dynamics of its interaction with the environment in the form of an internal model by being able to organize the lived experience of making anticipations and planning future actions.

Source: Deleuze, Perotto, JBL