{"id":13971,"date":"2023-11-14T15:00:57","date_gmt":"2023-11-14T15:00:57","guid":{"rendered":"http:\/\/TheNextWeb=1401504"},"modified":"2023-11-14T15:00:57","modified_gmt":"2023-11-14T15:00:57","slug":"deepmind-says-its-new-ai-system-is-the-worlds-most-accurate-10-day-weather-forecaster","status":"publish","type":"post","link":"https:\/\/www.londonchiropracter.com\/?p=13971","title":{"rendered":"DeepMind says its new AI system is the world\u2019s most accurate 10-day weather forecaster"},"content":{"rendered":"\n

A new AI model from Google DeepMind is the world\u2019s most accurate 10-day global weather forecasting system, according to the London-based lab.<\/p>\n

Named GraphCast, the model promises medium-range weather forecasts of \u201cunprecedented accuracy.\u201d In research published today<\/a>, GraphCast was found to be more precise and faster than the industry gold standard for weather simulation, the High-Resolution Forecast (HRES).<\/p>\n

The system also predicted extreme weather further into the future than was previously possible. These insights were analysed by the European Centre for Medium-Range Weather Forecasts (ECMWF), an intergovernmental organisation that produces the HRES.<\/p>\n

A live version<\/a> of Graphcast was deployed on the ECMWF website. <\/span>In September, the system accurately predicted that Hurricane Lee would make landfall in Nova Scotia around nine days before it happened.
<\/span><\/p>\n

\n
<\/p>\n

Get your ticket NOW for TNW Conference – Super Earlybird is almost sold out!<\/p>\n

Unleash innovation, connect with thousands of tech lovers and shape the future on June 20-21, 2024.<\/p>\n

<\/a><\/div>\n<\/div>\n

In contrast, <\/span>traditional forecasting methods only spotlighted Nova Scotia around six days in advance. They also provided less consistent predictions of the time and location of landfall.<\/span><\/p>\n

<\/i><\/a>GraphCast mapped both the trajectory and speeds of Cyclone Lee. Credit: Google DeepMind.<\/figcaption><\/figure>\n

Intriguingly, GraphCast can identify dangerous weather events without being trained to find them. After integrating a simple cyclone tracker, the model predicted cyclone movements more accurately than the HRES method.<\/span><\/p>\n

Such data could save lives and livelihoods. As the climate becomes more extreme and unpredictable, fast and accurate forecasts will provide increasingly vital insights for disaster planning.<\/span><\/p>\n

Matthew Chantry, a machine learning<\/a> coordinator at the ECMWF, believes his industry has reached an inflection point.<\/p>\n

\u201cThere\u2019s probably more work to be done to create reliable operational products, but this is likely the beginning of a revolution,\u201d Chantry said at a press briefing.<\/span><\/p>\n

Meteorological organisations, he added, had previously expected AI<\/a> to be most useful when merged with physics. But recent breakthroughs show that machine learning can also directly forecast the weather.<\/span><\/p>\n

How GraphCast works<\/h2>\n

Conventional weather forecasts are based on intricate physics equations. <\/span>These are then adapted into algorithms that run on supercomputers.<\/span><\/p>\n

The process can be painstaking. It also requires specialist knowledge and vast computing resources.<\/span><\/p>\n

GraphCast harnesses a different technique. The model combines machine learning with Graph Neural Networks (GNNs), an architecture that\u2019s adept at processing spatially structured data.<\/span><\/p>\n

To learn the causes and effects that determine weather changes, the system was trained on de<\/span>cades of weather information. <\/span><\/p>\n

Traditional approaches are also incorporated. <\/span>The ECMWF supplied Graphcare with training data from around 40 years of weather reanalysis, which encompassed monitoring from satellites, radars and weather stations. <\/span><\/p>\n

When there are gaps in these observations, traditional physics-based prediction methods fill them in. The result is a detailed history of global weather. <\/span>GraphCast uses these lessons from the past to predict the future. <\/span><\/p>\n

\"Graphs
<\/i><\/a>For predicting cyclone movements (left), GraphCast maintains greater accuracy than HRES as the lead time grows. For atmospheric river prediction. the model\u2019s prediction errors were lower than HRES\u2019s across the 10 days. Credit: Google DeepMind<\/figcaption>