The University of Houston, specifically the Air Quality Modeling and Forecasting Laboratory, has developed an Artificial Intelligence that could accurately forecast ozone levels two weeks in advance. This is a very notable improvement since with current systems they can only be predicted three days in advance. This AI could improve high ozone problems and also help solve climate change problems.
Ozone is a colorless gas that protects us by filtering ultraviolet radiation from the Sun and is useful in the correct amount and place. However, when there are high concentrations of it, it is toxic to the heart and lungs. It can also cause a sore throat, shortness of breath, asthma, or respiratory damage.
A breakthrough
This was very challenging. No one had ever done this before. I think we are the first to try to forecast ozone levels at the surface two weeks in advance, ”said Yunsoo Choi, professor of atmospheric chemistry and artificial intelligence deep learning. After three days, the precision in the system begins to decrease.
Team members added historical ozone data to the tests as they gradually adjusted the program’s reactions. This enables the Artificial Intelligence algorithm to accurately predict the results of ozone conditions in real life by recognizing what happened before in similar situations.
An Artificial Intelligence is given information and after many corrections and repetitions, the process refines over time and the program comes to know how to react to conditions that it already knows before.
Years of study and work
The team used data from four to five years of ozone. They described it as an evolutionary process of teaching the Artificial Intelligence system to recognize ozone conditions and to make ozone predictions and forecast, improving with time.
During the tests in the laboratory the team had a hard time for a few years, although they have finally managed to find the key, applying deep learning to air quality and weather forecasting. Finally, the system has already been adjusted enough after several years, since Artificial Intelligence already understands how to forecast.
Before success in the laboratory can lead to real-world service, many commercial steps are taken before the world can benefit from the discovery.
In fact, it is believed that this technology could help in the future not only with ozone, but could also help fight climate change and make our planet a safer place since knowing the quality of the air can make them understand each other many things related to the weather.
