Our latest Research & news
Jan 24, 2018
The main functions of the human mind in AI
We finished the software development of the basic 7-year-old person in the AI. Now in the system realized follow abilities: text understanding, questions answering, the analysis of the new information for its compliance with existing data, the requesting of the missing knowledge for learning.
Aug 17, 2017
Next level of AI cognitive abilities
CyberMind Technology finished the re-creation of the cognitive process of the 7-year- old person in the AI. Thanks to that, the AI can get knowledge about the environment or the knowledge domains with the content people have written for others (for example, the AI can read and understand Simple English Wikipedia). At the same time, the AI forms a knowledge structure identical to the knowledge structure of a human. This means that when the AI operates on knowledge, it will get the same results as if a person were doing it.
Re-creating the cognitive process is part of the current project of developing the AI with the re-created mind of a 7-year- old person.
Mar 16, 2017
CyberMind Technology filed the Patent Application with the USPTO; the title was "Detecting underlying message in context of current environment".
Mar 8, 2017
CyberMind Technology filed the Patent Application with the USPTO; the title was "Context-based detection of control commands".
Dec 7, 2016
CyberMind Technology raised investments
We are pleased to announce that our team was joined by investors who share our views on the development of AI and on the global changes that technology will bring in the future.
The funds raised will be used for the implementation of the software that re-creates the mind of a 7-year- old person in the AI.
Nov 16, 2016
Our article in the Hackernoon – "The development of strong AI, by copying the structures and processes of the human psyche"
What do I want from AI? Design of protein compounds, computation of elemental reactions and processes in atmosphere are very interesting tasks, of course. But it is far from me. I would like AI to understand me rather than simply identify words and fulfil voice commands. Though I should be thankful for the latter. But what I really want is that it could take hints. I want to be understood like in “Her”... Full article available on Hackernoon
Jul 8, 2016
CyberMind Technology finished the first stage of work on the AI development
We implemented a part of the cognitive process that corresponds to a pupil aged 7 years in the AI. Thanks to these cognitive possibilities, the computer can understand a simple natural language, understand and integrate new knowledge, and use the new information.
The next stage is to re-create the cognitive processes in the AI completely, and then, re-create all the processes and structures of the mind of a 7-year- old person in the AI.
Mar 21, 2016
Started the training of the AI
On March 21, we started the first stage of training the AI. We uploaded data, which a man receives during the preverbal phase of growth (from birth till the age of 12 months), into its semantic net. These data are the descriptions of approx. one thousand notions and their interconnections, and are specifically prepared by psychologists and linguists. Samples of such notions are “good”, “brightly”, or “what contrasts the hard from the soft”.
In April 2016, we'll start the second stage of training. The purpose of this stage will be to make it possible for the AI to read texts for children aged up 7 years and communicate with our experts.
Oct 06, 2015
Development of the AI
Our company focuses on the development of AI; we use our knowledge and experience in digitizing the human mind.
We plan to integrate algorithms that describe individual processes of the human psyche in a single system. Also, we will improve the algorithms by copying more and more complex structures of the human psyche. We are confident that our approach to development, our methods, and our knowledge and experience, will enable us to become leaders in this field.
Jun 02, 2015
Testing of the algorithms describing human needs
We developed the project for Data-Centric Alliance on Big data analysis. In this project, we tested our algorithms that modeled the human needs and selected the users according to their needs. We segmented the database containing 112 M profiles; each profile contained on an average up to 800 activities during 30 calendar days.
Within the framework of the project, models that simulated the processes of satisfying the needs of each individual user were developed based on the processing of the user’s internet activity data. Then, using models, we predicted the user’s reactions to the advertising content. As a result, we managed to reduce the CPC ratio to 8.5 times.
Jan 23, 2015
Algorithms for determining the reaction to website content based on the user's emotional preferences
In this project, we verified the algorithms that describe the emerging process and the behavior of emotions in the human psyche (emotions evaluation function). For this, we developed the solution to determine a user’s reaction to the news content. The project was implemented on the basis of the web-platform tut.by.
The forecast precision of the user's emotions onto the yet non-read news content was 82%. As a result, the average site visit duration increased by 32%, the conversion rate increased by 47%, and the bounce rate reduced by 28%.
Sep 17, 2014
Test of the algorithms describing the cognitive processes in the human psyche
Our goal was to test the basic algorithms on a large amount of data. For this, we used the data of Yandex—one of the largest internet companies in Europe. We conducted an experiment in which 1.1 M search requests data were processed.
We identified the users' characteristics of cognitive processes based on the search history. Then, we ranked websites in the search results based on the dominant cognitive processes of a specific user. As result, we managed to enhance the search relevance figure by 16%.
The experiment confirmed the correct operation of the algorithms, including the modeling of human cognitive processes using indirect data.
Feb 20, 2014
The human mentality basic architecture
Our team finished the development of the human mentality basic architecture in Feb 2014. In particular, we had developed the:
- Determination of module functions and character of module correlation;
- Algorithms reflecting basic level of cognitive processes;
- Algorithms reflecting the basic level of emotional behavior.
The results obtained will be used in the development of personalization systems for the Internet content.