“DNA Resolution” will expand cognitive abilities

“DNA for decision-making is a metaphor. It is a bio-inspired method of knowledge management. It is an attempt to recreate what natural DNA does in artificial systems in a stunningly perfect and unprecedented way” – says the Professor. Eduard Scherbeke of the School of Management and Economics at Gdansk University of Technology and the Australian University of Newcastle, leads the international team of scientists who have translated the most important functions of the human genetic code into the language of algorithms.

The source of knowledge is experience

Decision making DNA is an algorithm that allows you to make optimal decisions based on experience. It is drawn from the past as well as from the research and achievements of scholars and experts. Added to this are the endless resources of the Internet and data from many sensors that, thanks to the Internet of Things, will record billions of parameters – from the temperature of the water in the ocean, to the heart rate during the interview, to the wear and tear of oilfield drill bits. He quotes Professor Albert Einstein, who said that “experience is the only source of knowledge”. The challenge, he adds, is “only” to skillfully use the millions of data we have.

“Natural intelligence is the ability to act appropriately in a new situation. We want to support it with past experiences collected by people and other things. We can already see applications of DNA for decision-making in many areas, but the possibilities are endless” – says the Professor . Szczerbicki.

The concept of DNA decision-making has been implemented in more than 20 research projects. Thanks to it, the optimal places for the construction of geothermal installations in Australia were determined, the optimal time for servicing cars and machinery in Spain, or the decision was made on granting installment and investment loans to the banking sector in Poland. Applications in medicine and career planning are promising.

An “artificial” doctor with specialized knowledge

Many people with Alzheimer’s disease can be effectively helped if diagnosed early. In Spain, we conducted a study in a group of 350 patients. It turns out that an algorithm that collects data such as age, lifestyle and medical test results and analyzes them according to the concept of DNA decision-making is able to detect disease earlier than a doctor. There are many areas of medicine in which artificial intelligence is more effective than humans ”- says Prof. Szczerbicki, who collaborated in this field with Dr. Carlos Toro associated with the University of the Basque Country and the Spanish research center Vicomtech Research Institute.

Another example is sensors to recognize human activity. They assist, among other things, in the daily monitoring of elderly people in hospitals and nursing homes. Professor’s team. Szczerbicki proposed a new tool based on the use of expertise embedded in convolutional neural networks. Its effectiveness is much greater than that of existing solutions. The results are described in the scientific article “A new method for perceptual human activity recognition for the Internet of Things (HAR) using multi-headed convolutional attention” in the IEEE Internet of Things.

A smart program to make your dreams come true

Currently mr. Szczerbicki and his team are working on an app that helps achieve goals and dreams on the platform Dream . Technology. For example, a person who wants to study biochemical engineering in Australia or the United States in three years, while working while studying and staying there with his family, will receive a personal proposed business plan through DNA decision-making thanks to the application.

“The application is based on a questionnaire completed by the user, the life experiences of his community and the changing conditions of life. In a way, it is an interactive smart trainer that will motivate, motivate and help you get closer to your goal, all in real time” – summarizes Professor K. Szczerbicki.

Matrix, that is, after 30 years we will be hybrids

As the expert emphasizes, the DNA of decision-making is to meet the challenges posed by the development of the Internet of Things and the next steps in improving artificial intelligence. The collection of information stored by objects and applications will not only become a common phenomenon, but also reach an unprecedented scale. This is the challenge that DNA decision-making must face. It’s all about concentrating this knowledge into the most efficient tool possible. This can be compared to the Matrix trilogy, in which the protagonist learns martial arts by connecting to a special machine that installs new skills in his brain like programs on a hard drive.

“We are able to collect and formalize knowledge into collections and databases, which may consist of what is called SOEKS – a set of experience knowledge structure, i.e. a set of experiences from a given situation. We proposed a new tool for solving the traveling salesman problem, i.e. a typical logistical problem in the field of Operations Research. SOEKS absorbs past experiences from all experiences, regardless of whether it was a good or bad experience,” explains Prof. Edward Scherbecky.

In his opinion, AI-based decision support can be used in almost every area of ​​life. The aforementioned roving vendor problem is described in the scientific article “Evolutionary Algorithm and Critical DNA of the Multi-Vendor Problem” in the journal Neurocomputing.

***

a. Szczerbicki specializes in knowledge engineering and artificial intelligence. To date, the DNA decision-making concept developed by his team and its practical applications has been presented in more than 300 scientific articles. In its development involved, inter alia, Prof. Cesar Sanin of Newcastle University, A. Haoxi Zhang of the University of Information Technology in Chengdu (China), Professor. Syed Imran Shafiq from Aligarh Islamic University (India) and Dr. Carlos Toro from Vicomtec Research Institute (Spain).

PAP – Science in Poland, Karolina Duszczyk

cool / agt /