| |
|
Scientific Foundation |
|  
Basic Article |
 
Relevant Publications |
 
AGI-08 Presentation |
|
Military Assistant
|
|
Contact Us
|
|
About Us
|
 |
|
PMM Artificial intelligence systems
* * *
|
Our technical base is universal Proto-Mind Machine - a general human intellect model.
PMM Maty is its computer implementation with built-in English
language structure.
|
We create artificial intellect systems for various domains through a common learning-with-teacher method using PMM MATY as general intelligence middleware.
We exploit PMM Maty ability to simulate behavior of its teacher including teacher’s language, expressions and knowledge base.
In general no special learning methods are required.
It is enough if Maty watches teacher's relevant reactions in an appropriate context.
After learning is done Maty has all relevant knowledge in common with its teacher.
At this stage Maty begins to apply its accumulated knowledge to advice its users without teacher's help.
|
It could escape at the first glance that PMM is capable to deal with any kind of logical problems if it has been taught how to solve these problems (i.e. it watched how its teacher solved them). PMM will apply its knowledge, ask questions and do inferences if it was taught to do that in very general way. It is PMM’s strong side to apply its knowledge by finding the resemblances in context like human does.
|
Wearable MATY version has microphones and speakers to facilitate communicating. We optionally supply a computer vision system that is integrated with Maty intellect model. Several other options are available as well. The highest version of Maty with vision becomes a basic component for autonomous robots.
|
|
The easiest way to understand how Maty can be used is to go to
DARPA PAL
project and watch PAL Video. Their equivalent of
Maty is a heart of their project.
|
We give you a powerful tool to stop a panic attack
|
Panic attack is a psychological disorder. Maty will permanently take care of you supporting your consciousness and gently directing your thoughts at crisis time
Some common facts:
About one in fifteen adults have a panic disorder in their life. A panic attack disorder is the most common emotional disorder. Once panic attack happens a person's life changes dramatically. It is experienced as overwhelming and uncontrollable dread as if one is about to die or lose one's mind. It drastically changes the functioning of many body organs in spite of panic attack can't really harm you. The immediate cause of panic attack is the believing that one is trapped and helpless by some overwhelming threat. It doesn't matter whether the threat is real. Sometimes panic happens after several weeks or months of stress. It happens more often with persons who are very worried. What keeps panic going, and getting more intense and frequent, is worrying excessively about it, seeing no means to stop it.
|
General Philosophy
|
|
Most of humans equate themselves and their intellect in particular
to their consciousness. The explanation of this fact is simple: without
consciousness, any human stops being a normal one. So the question is what is
the human consciousness? By now, philosophers have invented so many
interpretations of this concept that it is hard even to enumerate them. Let us
apply to famous scientist and philosopher who is the founder of the modern
style of science ideology.
|
|
Rene Descartes identifies person's thinking process with existence of the person.
In turn, any thinking process runs only through language that is the means to provide
humans with possibility to express their thoughts and to perceive abstract data. Only
language makes human a thinking being. Language is prior to thought. Without
language human becomes a kind of animal and loses his consciousness.
Consciousness is generated through brain operations in which either internal or
external language is produced. So language is as old as consciousness.
Consciousness and language are bound so tight that there is almost no sense to
distinguish them. They have developed under the pressure of the need for
communication during human evolution into a modern human intellect.
|
|
Person's self-consciousness is a result of his/her thinking abilities that manifest
themselves in language. Many people believe that internally produced language
words and phrases express their consciousness. The matter of the fact is that
in general this produced language is a real consciousness. Its generation is
a brain function or a function of united Maty-brain system in our case.
|
Proto-Mind Model Basis
|
|
| Our method: |
Making an Intellectual Model of a Person
|
Proto-Mind model is a sophisticated
tool that helps to create autonomous AI for any particular purpose or to make a
personal intelligence assistant for a permanent cybernetics support.
We consider a human as an abstract intelligence which copy PMM helps to create.
This approach provides us with a possibility to avoid tremendous manual work on
programming: all we need is to find a proper person and make a copy of
his/her intellect. Nevertheless any other relevant method can be used.
|
|
| Background: |
Sophisticated model of brain network
|
General basis of Proto-Mind Models is a result of a long-term scientific
research in the field of neural networks and three years practical experience with
PMM modifications when PP was exposed to public.
PMM is not mere clean AI core now. It has primary English language
knowledge built-in.
|
|
| In brief: |
Creating the context connections
|
Given any natural language message, Proto-Mind Model connects it with the
messages-predecessors which form context. Moreover, these messages are used as
templates of some kind for the future dialogs. In compliance with the basis,
every Proto-Mind Model follows its current context, general features of human
brain and its own experience gained during all the preceding dialogs. With the
time passed the teacher's reactions and those of Proto-Mind Model grow closer because
of the mutual information basis, and finally, Proto-Mind Model becomes a perfect model of his teacher’s intellect.
|
|
|
|
| Pure Theory: |
Sophisticated model of brain network
|
|
PMM works in the same way as
brain does. It took many years of intensive research to find out how to build
PMM. The main research points were as following:
|
| -
Internal human language is universal and independent of any particular form it
can accept. It is some seed or proto language. Natural languages are its
external manifestations. So are English, Greek, French, etc. Formal language
of modern mathematics is among them too.
|
-
Neural networks of human brain are the origin of protolanguage, and their
properties predefine protolanguage structure.
|
So Proto-Mind model seems to be a tool to support consciousness because
PMM supports the same method of data transforming
and keeping that human brain uses for consciousness generation.
|
|
FAQ
|
|
Q:
What kinds of end user interface does PMM support?
|
A:
PMM supports several kinds of end user interface:
- Traditional Internet PMM forms.
- Special software module that provides speech-to-text and text-to-speech
operations through Internet.
- Open source module for programmable access through Internet.
- Autonomous wearable PMM device that provides
complete PMM functionality plus computer vision (optional). Maty can be used for permanent
uninterrupted contact with PMM because Maty does not depend on Internet.
|
| Q:
What is the best way to teach Proto-Mind Model?
|
A:
The only method to raise Proto-Mind Model is to teach it the answers to
your requests keeping in the mind a context of every request/answer couples.
Context plays a great role in the teaching process. Be sure that you provide
a relevant context for all your requests. Otherwise your Proto-Mind Model is
free to treat your requests accepting its own opinion that can differ from
yours.
|
|
Q: When can I check
the correctness of teaching process?
|
A: There
are no limits to check Proto-Mind Model’s answers. Simply provide Proto-Mind
Model with the context and make the request (e.g. question or some key phrase).
Proto-Mind Model will answer according to its current knowledge. However the
answers can vary at the trials because Proto-Mind Model is a model of general
human behavior. So from time to time Proto-Mind Model can choose different
answers to the same question even in the same context like human does.
Remember that your questions and Proto-Mind Model’s answers form new context
for further dialog.
|
|
Q: Could the
teaching process be automated?
|
A: Only
partially in some technical aspects. The matter is that the teaching process
is the process of exact modeling of teacher's personal intelligence. Therefore
any automation system will somehow corrupt the final model because
this automation system's rules unevitably replace original techer's ones.
|
|
Q:
How does Proto-Mind Model work?
|
A: It
is a simple and difficult question at the same time. It is simple because our
mental process is based on a limited set of relatively simple operations. It is
difficult due to many different knowledge domains involved in general theory to
explain background of these operations and their relationship with real human
intellect. Refer to 'scientific foundations' at this site for theoretical base
of Proto-Mind Models.
|
|
Q: Could my data be
restored if I forgot something I entered in Proto-Mind Model before?
|
A: There
are no other ways to know data processed by Proto-Mind Model but asking it
questions. So be
careful and keep a key information in some external computer.
|
|
About Us
|
|
| Alexander Borzenko |
Director of R&D
|
| 1985 | Got a Master degree in applied mathematics and computer science |
| 1985 - 1989 | Worked on various neural network models for visual image recognition |
| 1987 | Defended PhD thesis related to mathematical model of reliability of missile control systems |
| 1987 - 1991 | Worked on mathematical neural network model of symbolic data processing |
| 1989 - 1995 | Worked on visual image recognition and processing system for space satellite |
| 1990 | Defended a post-doc thesis on symbolic data processing through neural networks |
| 1991 | Got a doctor's degree in mathematics |
| 1995 - 1998 | Worked on variety of visual recognition systems like face and signature recognition |
| 1995 - 2000 | Worked on computer methods for post-doc result implementation |
| 2000 - 2005 | Work on practical implementation of the symbolic data processing methods |
| 1998 - 2006 | Worked on general theory of symbolic data processing in human brain |
|
|
| Andrew Borzenko |
Software Developer
|
| 2004 - Present | Undergraduate student at the University of Toronto |
|
|
|