Scientists believe that machine learning systems based on machine learning are unable to understand the physical laws governing the system under study, which increases the likelihood of error. By examining the weather forecast data with the method «Random forestThey found that artificial intelligence has many errors in predicting local weather. According to them, artificial intelligence can play the role of an assistant in performing complex high-speed computing for humans.
A century ago, the English mathematician «Lewis Fry RichardsonCame up with an amazing idea: to build a systematic mathematical process that could predict the weather. Richardson in his book entitled «Weather forecasting with numerical processHe tried to write an equation that he could use to analyze the dynamics of the atmosphere.
Unfortunately, this equation was not successful because there was not enough knowledge about the atmosphere at that time. A century later, weather forecasts were based on complex calculations that Richardson had made in his day.
However, there are still fundamental challenges. Predicting the thunderstorms that caused Tornado، تگرگ And heavy rain It is difficult.
Like many other disciplines, the development of tools such as Artificial intelligence And Machine learning Promises good news for weather forecasting. But as we believe these tools offer new opportunities, many areas of work require the skills of experienced people.
In a recently published article, scientists focus on forecasting Rainfall Have focused. Using the random forest method in machine learning, they have tried to categorize the data and use it to predict different outcomes. They found that the smaller the phenomenon in question – for example, lightning – the greater the artificial intelligence error in prediction.
Various other groups, on the other hand, have focused on forecasting tornadoes, hail, and lightning. Although machine learning is capable of doing Advanced calculations in a very short time Help scientists, but in the end, artificial intelligence can not replace experienced people in this field.
These tools can have general predictions, meaning that they only predict large-scale events with high accuracy and do not have the ability to analyze local weather.
Artificial intelligence and humans
Weather forecasting requires an accurate understanding of the physics of the problem. Unfortunately, artificial intelligence is not able to understand the laws of physics in such complex situations. This makes the results seem unrealistic – for example, artificial intelligence makes extreme predictions of Temperature In the abnormal range.
This is where the role of man comes into play. Scientists have finally found aYen is the only job in which artificial intelligence can not replace humans. In fact, artificial intelligence in this area is the only assistant and auxiliary force for humans to help humans perform high-speed calculations.