Researchers at Kobe University and Osaka University have developed AI that can extract hidden equations from ordinary data and create a model that follows the laws of physics.
The new technology will help find hidden equations and laws that were considered inexplicable. For example, it is possible to study ecosystem resilience through knowledge and modeling.
The ability to formulate physical phenomena using AI will help to model the surrounding phenomena extremely quickly and accurately. In modern research, AI is used to make transformed data that fit the equation of motion. Therefore, it is difficult to apply the algorithm to actual data for which there is no formula in advance.
According to the authors, in the future, it may be possible to discover the hidden physical laws behind the phenomena that were previously considered incompatible with Newton’s laws.
Usually, predictions of physical phenomena come from simulations using supercomputers. These simulations use mathematical models based on the laws of physics, but if the model is not highly reliable, this will affect the results. Therefore, it is necessary to develop a method for creating highly reliable models based on observations of phenomena.
The authors of the new work have created a method for finding new equations of motion based on observations of phenomena to which Newton’s laws can be applied. The research team succeeded in developing an AI that can find geometric properties in data. If the equations of motion can be extracted from the data, then they can be used to create models and simulations that follow the laws of physics.
Physical modeling is conducted in a wide variety of fields, such as weather forecasting, drug detection, building analysis, and vehicle design. You can also find the hidden equation of motion in animal population data that shows changes in population numbers.