Tipalo AI - a thinking entity on its own

"Thought is the labor of the intellect, reverie its pleasure." Victor Hugo, Les Misérables

This paper is pointed to everybody, which is related in some kind with AI. This document was thought to provide an incentive, on the one hand to discover the human thinking independently of themselves, to explore themselves and ALSO to accept, even if one does not like what one found, while on the other hand to have the certainty that it is still possible to reach all the goals of AI, gradually. How can one find such a viable solution? Well, to find a solution is always very hard, but the solution itself is really simple.

Tipalo - the biological approach to AI

The new concept and its software implementation requires a different way of thinking in order to to describe and analyze tasks and further to design, realize, distribute, maintain, deploy, maintain and use applications. Why is a different mindset needed? Well, the most common problems of the current concepts start with the questions:

1. what is an object?

2. how do we represent an object in a computer?

3. what actions should an object have?

Unfortunately, these three questions are answered differently in each concept.

The new concept has a new definition regarding objects. Each object is considered as a network of base units. The base unit is used to create composite objects of a higher organizational form. Base units and objects are components.

The base unit requires:

1. space, to exist

2. time flow, to act and react with others

3. behavior, how and when could this unit act

Let's take two objects from the real world as an example: a stone and living organism. When we look at these two objects from the point of view of the new Tipalo concept, then we have the following object definitions:

1. a stone, is a composite object, a network of atoms, the atom is the basic unit, where each atom occupies a particular place in space and acts and reacts with other objects during the passage of time by using its own behavior. Consequently, the whole stone occupies a place in space, the sum of all its atoms, and has a composite behavior, the interaction of all its atoms.

2. a living organism, is a composite object, a network of cells, the cell is the basic unit, where each cell occupies a particular place in space and acts and reacts with other objects during the course of time by using its own behavior. Consequently, the entire organism occupies a place in space, the sum of all its of its cells, and has a composite behavior, the interaction of all its cells

The software implementation

Base units, time flow and space are the linchpin in the analysis of tasks and further in the development of applications. The principle is: "Many small things, that constantly interact with each other during time flow in a space create a composite object."

The software package allows simulations of the real world with the following features:

1. a space, which contains all the basic units

2. a time flow, for all basic units in the space

3. a standard definition for basic units, which are used to create objects of a higher form of organization.

The principle as a task description and solution approach

The first example: a house is an object for all of us. But at the same time, it is also a set of bricks in a particular arrangement and of course these said bricks are constantly reacting with each other. If we then also need a financial estimate for the construction of this house, the whole principle becomes very clear: it is much more accurate to multiply the price of a single brick by the number of all bricks needed instead of giving any sum for the house, i.e. "bottom-up instead of top-down".

The second example is the behavior of water. I just remember the television appearance of a famous French Nobel Prize winner in chemistry and his explanation. Water consists of water molecules, as 2 atoms of hydrogen and 1 atom of oxygen, i.e. H2O.

Water can freeze or evaporate, but a single water molecule does not have this behavior. Why is that so? Because water is a set of many individual water molecules, that react with each other in space and time.

Aristotle said long ago: "The whole is more than the sum of its parts." Why is that? Because the parts react in time with themselves as well as with the outside world, therefore having a different behavior/outcome even if we do not realize it immediately, such as changing its overall borders with the effect it could change its external shape or the internal structure changes at once, with the result of altering some or all properties.

The third example is the biological brain. The amount of neurons in a brain can vary, from 1-2 millions for an insect, via 1-10 billion for a mammal to 86 billion for a human, letting aside the fact that other animals have much larger brains then we do, such as elephants, wales or others. However, each of them evolves in time, by starting with a pre-defined amount of cells and ties between them, which is the genetic knowledge inherited from their ancestors and created during its inception, e.g. breeding, pregnancy, etc. But the amount of neurons alone does not represent a brain in its totality, as every intelligence consists of many and specialized regions, each of them capable of evolving and indispensable for the knowledge and behavior of the respective organism.

The huge number of building blocks for a simulation

In this type of application, something becomes very clear, and that is the number of building blocks involved. To put numbers to the whole, let's take the following example: The brain as well as the size ratios for building blocks of a cell.

The human brain consists of about 86 billion neurons. Nobody knows that exactly yet. Each neuron has connections with other neurons. How many connections? Well, at birth, every single neuron already has 5,000 ! connections with other neurons. During our lifetime, this number is increased to about 20,000 ! connections. This also means the following: in the 9 months of pregnancy the number of new neuronal connections is about 2 million per second; during our life on average 50.000 new connections per second are created.

All this requires a huge main memory and the corresponding computing power. Not to forget the fact that everything has to be processed simultaneously, in parallel and of course networked and coordinated.

Summary

Tipalo AI has a holistic approach, integrating many areas, such as philosophy, physics, chemistry, cytology, histology, genetics, embryology, biology, anatomy, physiology, neurology, psychology, sociology, business management, economics, ethics, history,?mythology, linguistics, pedagogy, semiconductor technologies, databases, programming languages, operating systems, software development and especially logic.

The Tipalo AI is a simulation of brain regions, while connected to a body hardware. While the software runs in real-time, all the time and standalone, allowing similar reaction times when compared to biological brain, e.g. 1 ms for all the spiking neurons, it enables "living machines" with a digital brain working without external intervention.

The Tipalo AI model is very clear and each element has a fixed scope, it allows hereby the interactive tracking of each active neuron. This approach allows quantifying and hereby explaining the architecture of each neural network, so that an extension is possible at any time, within the predefined amount of neurons that an ANS allows, which on its side, depends on the level of intelligence as well as of the corresponding body structure with the connected devices, as sensors, actors and internal organs.

Furthermore, a predefined genetic knowledge of the corresponding ANS, Artificial Nervous System, in the form of predefined connections between the individual cells, is both given and desirable. Thus, further knowledge can be easily added on the basis of personal experience, including the current context, so that different areas of knowledge are updated independently of each other. This is enabled by an own embedded SLM, Self-Learning Mechanism, using many and scalable SAM, Self-Associative Memory.

By remembering the accumulated knowledge, the ANS can easily explain why a certain decision or action was made, at any given time. The black-box problem of all current/conventional neural networks is thus solved, by the exact localization of each neuron and also the meaning of each connection to other neurons. Although Tipalo PNN, Programmable Neural Networks, are spiking and thus neuromorphic in design, they do not use biases, instead they make use of many and different feedback types during the self-learning process. A fully digitally implementation was chosen as SPL, Self Programming Logic, in order to simplify and replace the analogue template of the biological brain regarding connectivity logic between cells, which are based on genetic features. Neurons are primarily cells, even AFTER becoming specialized logic cells.

As a final word here are some thoughts of Sun Tzu. The Chinese thinker, which started as an unnoticed philosopher in his homeland and ended as a successful warlord abroad, describes in the forth chapter, "Tactics", of his only book, "The art of war", among many other things, the evaluation of the thinking abilities of a man, see paragraph 8 - 12.

?Lifting a spider web is no proof of big power; seeing the sun and the moon is no proof of a sharp eye; hearing the noise of thunder is no proof of a good ear.

It is no proof for outstanding performance, if you see the victory, only if it is seen by everybody else.

It is no proof for outstanding performance, when you fight and win and the whole kingdom says: Well done.

A proof for outstanding performance is winning WITHOUT any fighting.

There is no appreciation for outstanding performance, neither for wisdom, because the circumstances of the victory were not brought to light, nor for courage, because no single drop of blood was shed."