The History of HRMNY AI

Abstract

Since 2018, the HRMNY AI team has been developing a knowledge system to understand the nuances of human personality functioning, as well as tools and products built on this knowledge system.

The mission of the company is to maximize the creative potential of humanity.

The main objectives of the company are:

1. To create an artificial intelligence system that understands a person better than they understand themselves.
2. To implement technological solutions that personalize everything related to the individual.
3. To elevate the system for assessing the level of human abilities and evaluating potential to a new level.
4. To integrate everywhere a perfect “translation” system that translates what one person says to another (meaning the sense of told).

Over the years, the company has developed the HRMNY Framework, which today includes the most advanced tools for understanding people, predicting their reactions and behaviors, and enhancing their productivity.

The foundation of the company's developments is based on the scientific research of its founder and ideological inspirer, Evgeniy Pecheniy-Shcherbanskiy, who has been engaged in theoretical and practical studies in the field of explaining and prediction of human behavior for over 15 years.

The products and solutions created based on the HRMNY Framework open a new level of depth and effectiveness in everything related to working with people, understanding people, influencing people, and managing personal growth.

At the beginning of the document, the first steps in the development of HRMNY are described in details, including research of similar works, the assessment of their possible applicability in our practice (to prevent "reinvention of a wheel"), an in-depth analysis of underlying principles and methodologies of existing external services, and as a conclusion ‒ justification of their complete inadequacy for practical use.

Next, the authors of the HRMNY Framework outline their vision for how an AI system for recognizing cognitive patterns of human personality should be constructed and operate. Each provision is explained in detail.

Separate sections describe the foundations of the HRMNY Framework methodology in comparison with existing methodologies. A detailed comparison of the methodologies is presented, along with a justification for the absolute (as of now) superiority of HRMNY.

The dataset used by HRMNY for training and evaluating its neural networks is described, along with the process of its creation and annotation.

Practical approaches to implementing AI, its evaluation, and the results of its performance are discussed.

The document concludes with a description of current work and future plans.

How It All Began

In 2019, we began to consider the implementation of AI (artificial intelligence) in our products for the first time. The goal was to automate the complex tasks of identifying cognitive patterns of personality. Those require significant time resources and, most important, highly qualified experts. At the same time, we believed that such a task (defining cognitive patterns of personality using AI) had long been solved because:

• first, well-known personality typing methodologies, such as Big-5, DISC, and MBTI, have existed for a long time. Therefore, personality typing services based on these methodologies must have already been created, operate effectively, and we could either use them or leverage their experiences and technologies as a foundation for developing our own service;
• second, at that time, the market already featured and widely recognized IBM Watson, a service that identifies a range of cognitive personality patterns based on analyzing a person's resume text. It was widely claimed that such a service solved the task of personality typing quickly, efficiently, and inexpensively;
• later, we discovered several other online services which use machine learning models for text analysis to determine Big-5, DISC, and MBTI parameters when typing individuals ‒ namely, Crystal, Symanto, and Humantic. These companies were positioned as well-known personality typing services with a high volume of sales and large, recognizable clients. It was openly stated that all of them were built on the foundation of IBM Watson, albeit with customized enhancements. The reported sales levels of these companies, their client lists, and the scientific approach for presenting the relevance of their services and methods of implementation appeared convincing enough to draw the conclusion that it would be possible to collaborate with them to study their models and products and evaluate their accuracy.