First Protein folding, Now Weather Forecast: London-based AI firm Deep Mind continues to apply deep learning for difficult science problems. The UK National Weather Service, working with the Met Office, has developed a deep learning tool called Deep Mind DGMR that can accurately predict the possibility of rain in the next 90 minutes অন্যতম one of the most difficult challenges in weather forecasting.
By making blind comparisons with existing equipment, dozens of experts have judged DGMR forecasting to be the best among a variety of factors – including location, range, movement, and rainfall intensity – -% of the time. The results were Published On a nature paper Today.
Deep Mind’s new tool No Alphafold, which has opened up a major problem in biology Scientists have been fighting for decades. Yet little improvement in forecasting.
Rain forecasts, especially heavy rains, are important for many industries, ranging from outdoor events to emergency services. But it’s hard to do well. Finding out how much water is in the sky and when and where it will fall depends on weather processes, such as temperature changes, cloud formation and wind. All of these factors are complex enough in themselves, but they are even more complex when taken together.
Existing prediction techniques use extensive computer simulations of atmospheric physics. These work well for long-term forecasting but are less good for predicting what is going to happen in the next hour or so, known as nookasting. Previous deep-learning strategies have been improved, but they usually do better in one thing, such as predicting location, at another expense, such as predicting intensity.
Greg Carbin, head of forecasting activities at the NOAA Weather Forecast Center in the United States, said: “Rainfall remains a major challenge for meteorologists who were not involved.
The Deep Mind team trained their AI on radar data. Many countries publish frequent snapshots throughout the day of radar measurements that track cloud formation and movement. In the UK, for example, a new lesson is published every five minutes. Putting these snapshots together provides an up-to-date stop-motion video that shows how the types of rain are moving across a country, similar to the forecast visuals you see on TV.