I apologize for the quality of the translation, it is made with the assistance of an interpreter Google.
November 20, 2015.
I came up with my ex friend "hybrid intelligent systems" and decided to patent it. Sometimes it is necessary to obtain a patent to attract investors because the cost of international patenting burdensome. To guard against possible risks signed a preliminary agreement, which shall be certified by a notary. However, as shown by my practice it is impossible to insure against betrayal, especially when one betrays the trust. And that's how I was cheated.
On the basis of a notarized "Agreement" dated July 26, 2014 on register number 5426, certified by a notary public, HD Kazakbaev notary office number 6, Hamza district of Tashkent investor had access to information on the patent "Hybrid intelligent system."
Investor in its own name applied for registration number ¹ IAP 20140484 to the Agency for Intellectual Property of the Republic of Uzbekistan November 19, 2014, they without agreement with me was to change the name of the application with the "Hybrid intelligent system" to "artificial intuition" and it was filed without the application of a It "Agreement" dated July 26, 2014 ¹ 5426. Roster signed a new agreement to change the name of "artificial intuition" investor refused, in the light of recent developments which will be described below, it is obvious to me now that it was part of his plan to deprive me of authorship and stealing the results of my intellectual work.
In the words of investor Agency on Intellectual Property of the Republic of Uzbekistan have refused to examine the application for grant of a patent on the so-called "Artificial intuition", and about the further steps in the implementation of the agreement between the investor informed me not to.
November 7, 2015 I came across the internet site www.lifehacker.ru to comment investor to one of the articles on the topic "Artificial Intelligence" where he informs that he is the author of "artificial intuition" (!) And that they submitted to the 2015 a patent application in Russia "method of creating an object model" for the number 2015125952 of 30 June and the United States "artificial intuition" for ¹14 / 833,505 of 24 August. When I asked about the status of international patenting investor gave the answer "I'm not a writer" and submitted them to the applications I have on the list of co-sponsors, but only he and my former friend. In Russia and in Uzbekistan investor action can be qualified under Article 168 "Fraud" of the Criminal Code of Uzbekistan (Article 159 of the Criminal Code of Russia. Fraud), as in the application ¹ IAP 20140484 "artificial intuition" filed with the Intellectual Property Agency of the Republic of Uzbekistan in 2014 I was on the list of authors. Unfortunately steps an investor who does not understand anything at all on the patent, supported by the work of my former friend who helps him everywhere for a fee to advertise itself as the author of "artificial intuition", shorter than the other betrayed me for money.
So I decided to publish the text of the abstract of the patent "Hybrid intelligent system" on the Internet, to keep his authorship of those that then repainted in "artificial intuition", as well as the Agency for Intellectual Property of the Republic of Uzbekistan have refused to consider and passed a year with the application for the invention.
ABSTRACT OF THE INVENTION
The invention relates to computer models of artificial intelligence - a hybrid intelligent systems (HIS), using the work of an expert system (ES), with or without an artificial neural network (ANN), different way to get the object model.
EFFECT: - the construction of a model of reality (consistent and coherent model of the studied object) describes a set of links between elements of the object, to solve various problems of intellectual processing of information, including: approximation and interpolation, recognition and classification of images, data compression, prediction, identification management, associations, etc ..
The technical result is achieved by:
1. Get a baseline data about the object under study, which is [the object] may be of a plurality of divisible or not divisible elements, characterized by a set of data.
2. The data are subject to the procedure of "cleaning" in order to exclude non-essential data are based on the criteria of cleaning that may be adaptive changes during the iterative process of constructing a model of the studied object.
3. Next, the object under study is divided into elements of the [group] characterized by identical sets of data, groups are formed on the basis of the criteria of "grouping", which may be adaptive changes during the iterative process of constructing a model of the studied object. The criteria for "group" are used including for the evaluation and optimization of the number of connections between the elements of [Group], which for example in the procedure for paired comparisons of elements of the object are of the proportional [to the square of the number of items divided into two minus the number of elements]. Number of links generated as seen from above is dependent on the square of the number of elements, so this imposes certain limitations associated with the computing power.
4. Grouped data is checked for completeness and redundancy, and incompleteness when they are complemented with interpolated depending on the model of the object under study data, while redundancy - redundant data is discarded, thus making the procedure of normalization grouped data.
5. Normalized data is then grouped with intelligent processing of the expert system (ES), with or without an artificial neural network (ANN), which consists of the pair (or more) processing functions of homogeneous data according to the "Rules of the joint data ES and HC" which may be adaptive changes during the iterative process of constructing a model of the studied object. And if the function is one - that limited the work of ES sorting. And if a few features - it is connected to the INS, the operation of which is to collect the values ;;of different functions for different parameters and fixed data.
6. The importance of relations between the elements of the object under study are subjected to the normalization of relations normalization according to the rules that may be adaptive changes during the iterative process of constructing a model of the studied object.
7. The normalized values ;;for relations between the elements of the object being studied are in the process of analysis, the analysis criteria can be adaptively changed during the iterative process of constructing a model of the studied object. The purpose of the analysis is to identify the "Governing Relations" [dominant relations], that is, bonds, the value of which is dominant in the resulting model of the studied object, and in the interim and final model of the object under study is not the dominant communication discarded as insignificant, resulting in a "disconnect" of the respective neurons ANN.
8. After receiving the intermediate model of the studied object is its verification of conformity by comparing the predicted and actual data of the object in accordance with the rules of compliance testing that may be adaptive changes during the iterative process of constructing a model of the studied object.
9. In accordance intermediate model of the studied object validation criteria, the intermediate model becomes final [finished] model of the object under study. And for non-compliance to investigate the cause, and to the rules of "clearing", "group", "co-processing of data for the ES and the National Assembly," "normalization", "analysis" amended in accordance with the existing knowledge base of ES or expert knowledge. And the process of obtaining a model of the object is started again until the final [finished] model of the studied object.
The invention consists in the following.
The task of creating a model of the studied object, describes a set of links between elements of the object in general, it is very complicated. The satisfactory construction of a model in many cases, can not be achieved without the participation of the person who uses the construction of formalized indications of experience.
Intelligent processing of information using computer technology can facilitate a person to create a model of the studied object. Especially effective this treatment for a large number of elements of the object being studied, when they receive a lot of data about the object. For complete information about the object must be reduced to the form allowing the optimum interpret the data. In preparing the object information to facilitate the understanding of the human subject used a variety of methods, including - pattern recognition theory.
But our analysis of existing methods of recognition using ANN revealed unaccounted opportunities in intelligent processing of information, namely, the pre-treatment [preprocessing] Information supplied to the neurons of the ANN.
The principal difference between the claimed invention is the treatment of relations between the elements of the object of study in pre-processing, ie before the submission of information to the INS. This considerably simplifies the learning process and the final structure of the INS (reduces the number of layers in the ANN and the number of neurons), because for the purposes of recognition uses only those neurons that are responsible for "managing communication," in contrast to the traditional multi-layer ANN where the information about the internal connections the object is stored in the form of formalized links between the neurons of different layers.
The claimed invention is directed to the solution of various problems of intellectual processing of information, including - approximation and interpolation, recognition and classification of images, data compression, prediction, identification, management, associations, etc ..
The amount of the above features of the invention is related causal link to the technical result of the invention. The final model of the object is obtained by us from the set [synthesis] "manage connections" between the elements of the object in a way that results in a similarity with the standard within the optimum computing power user.
1 shows a block diagram of the hybrid intelligent systems claimed herein.
The proposed way of presenting the object model of intelligent hybrid system has several advantages over conventional methods.
The main criterion for the advantages of this method is not the volume of the information retrieved, and its quality. The method allows to simplify and accelerate the creation of a synthetic model of the object for interpretation (including - visual) and thus allows to focus on the interests of the observer objects hidden patterns. Synthesized by the inventive process object model is much less sensitive to the input data of the object, so in its implementation are reduced requirements for the receipt of data. It is important that the synthesized GIS object model in most cases is compact, allowing for its use in a wide range of devices having limit computing capabilities.
Claim
The method of intelligent information processing in hybrid intelligent system (HIS) with the use of the expert system (ES), with or without an artificial neural network (ANN) to build the model of reality (consistent and coherent model of the studied object), comprising the artificial generation of pair (or more) ties and subsequent filtration of links between elements or groups of elements of the object in order to identify critical relationships - the hidden laws of the object being studied. And when working together ES and ANN Electric trains ANN picking up weights in total action elements ANN and the results of the processing of the pair (or more) bonds elements of the object in order to identify critical links between elements or groups of elements of the object by means of a retrospective analysis.
PS The above-mentioned "Investor" (Kolotygin AI) and "my former friend" (Akayomov AA) received the Russian patent RU 2607977 "METHOD OF CREATING THE OBJECT MODEL" (published 11/01/2017) and US patent US 20170004401 A1 "Artificial Intuition "(published on January 5, 2017), as you can see on the site http://www1.fips.ru and through Google are not much different from the description of the" Hybrid Intelligent System "given here.
Analysis of patent formulas RU 2607977 and US 20170004401 A1 shows that the patent applicant, because of his ignorance, fully trusted his patent attorneys and in fact they patented the algorithm, and algorithms are known to be easily modified and therefore such patents do not cost anything.
For example, Ms. Olga Uskova and her company Cognitive Technologies
announced a breakthrough at the international electronics exhibition CES-2018 (Consumer Electronics Show) and received many foreign awards for the original approach - the so-called Low Level Data Fusion. It embodies the principle of intuitive understanding, which Olga Uskova explained as follows: "We break away from logical sequence and begin to build convergence randomly, having at the end the desired, maybe not optimal, but very accurate result."
If you wish, you can prove that this is nothing more than a description of the patent formula given above, but since as a result, the algorithm was patented. to sue O. Uskova is useless.
Taking into account the committed different financial weight categories Kolotygina A.I. and the companies Cognitive Technologies have completely different destinies - the first to carry on the Internet to the mass media on their own behalf the “revelations” written by A.A. Akayomov. and bubbling up from the purchased dubious glory of the pioneer of "artificial intuition", while the latter actually did this very "artificial intuition" and received all the laurels and money.
Moreover, Mrs. Olga Uskova amazes me - sometimes she publicly literally quotes “revelations” written by A.A. Akayomov. for various Internet sites and voiced by the "speaking head" Kolotygin AI, without any reference to the source, i-yi-yi.
He that lives a knave will hardly die an honest man.
PSS An appeal to my dear former friends - I know you regularly drop by and read this opus in the hope that I will say something else to you idiots. In vain, I already said everything.
https://www.youtube.com/watch?v=3Fej5orKAbU