Russian scientists and their foreign partners have developed artificial intelligence that can predict the course of chaotic economic processes. This was reported by the press service of the RNF.
Modern science has learned to accurately model many processes, from the movement of planets in orbit to the destruction of a brick under load. There is an element of randomness in some processes, such as airflow around an aircraft, but these can be taken into account and form a reliable mathematical description of an aircraft’s flight. However, there are too many chaotic elements in the social sciences, including economics, and the market is affected not only by factors of production, but also by epidemics and subjective cultural preferences (demand for a particular product can be stimulated by a pop star). ) as well as the emergence of financial bubbles. Some major economic events, such as the 2008 crisis, were triggered by unknown innovative mechanisms (securitization of mortgage loans). In this regard, all modern economic models are far from astronomical and aerodynamic in reliability, and therefore scientists are looking for a way to calculate the effect of chaos.
Petersburg State University and their Russian and foreign colleagues developed a way to detect erratic fluctuations in the economy. The method is based on the use of a model that reproduces the development process of a society consisting of two generations, young and old, in a given economy throughout the lifetime of these generations. A pricing model was also considered, taking into account the regional location of markets with a network of connections. It has long been used to describe the complex chaotic dynamics of price formation in food markets, typical for fish markets, for example.
For calculations, evolutionary algorithms and reinforcement machine learning were used – areas of artificial intelligence that involve self-learning of a computer program in the process of solving a problem. The algorithm not only offers one or the other choice, but also takes into account past experiences of successful or incorrect responses. This task was given to a machine from the Czech National Supercomputer Center in Ostrava. He figured it out in 48 hours, while an ordinary computer would take more than 17 years.
At the same time, the authors warn that their method does not provide a comprehensive and accurate answer. Businessmen and government representatives make decisions based on intuition, which is the product of many experiences. Data obtained using the new algorithm could become just one of the tools at the disposal of a human analyst: Roughly speaking, artificial intelligence generalizes large amounts of data and makes hidden patterns open.
ancient biologists approvedthat the moths’ “tails” act as decoys for bats.