AI shows 'potential for revolution' in weather forecasting

 

AI shows 'potential for revolution' in weather forecasting


A groundbreaking new artificial intelligence system has the potential to revolutionize weather forecasting, say its developers.


The inventor of this new technology is Google-owned AI company 'DeepMind'. The company says it can provide "highly accurate" 10-day weather forecasts in just one minute.


According to the report of the British daily Independent, the forecast made by the system called 'Graphcast' is not only accurate but also effective. This means it is possible to produce results faster using less data than ever before.


  This technology can also help in early warning by detecting adverse weather conditions like cyclones, floods and wildfires. Google says, through this, people will get a chance to prepare better and it will also play a role in saving people's lives.


Currently, weather forecasting relies on a system called 'Numerical Weather Prediction', where physics equations and computer algorithms are combined in supercomputers. This requires massive computing power as well as the expertise of weather forecasters.


The new system uses deep learning technology, which not only relies on physics equations, but also collects data from how Earth's weather changes over time.


Creating such a model is somewhat complicated. Because, in between, decades of weather data have to be trained. However, now that it has become a reality, the amount of data needed to predict the weather will be greatly reduced, the Independent reports.


A 10-day forecast now takes a minute on a machine, a process that once required hours on a supercomputer.


In a recent paper published in the scientific journal Science, its creators wrote, the new system was more accurate than conventional standards in 90 percent of tests.


In addition, the system detected adverse weather conditions without any training. For example, it predicted Hurricane Lee, which struck the northern United States in September, nine days in advance, while conventional systems detected it six days earlier.


DeepMind said Graphcast's flammability projections could be particularly useful during climate crises. And it can predict temperatures in places that are at risk of exceeding record highs. As a result, there will be an opportunity to make possible preparations before the fire strikes.


In addition, DeepMind may open the technology so others can use it. The company says it will also help with new tools and research to tackle environmental challenges.

Post a Comment

Previous Post Next Post

POST ADS1

POST ADS 2