Google DeepMind’s Post

Wow, this is cool. And it raises many other questions, some of which have deep implications for society. As AI models increase accuracy in complex systems prediction it's not just the weather it will be able to forecast. Decoding human behaviour, attitudinal responses and actions will come next. Forecasting reactions to economic, political, legal, social and cultural drivers will all be on the table - as will the data to know how to optimise stimuli to influence actions. Indeed these kind of technologies are already existent with groups like Palantir for use in policing and military. The question will be who's going to lead the regulation of these powerful society shaping capabilities and what ethics should the public have say in? If accepting cookies were the internets opt in governance, what comes next? #ai #governance #socialethics #futures

The model utilizes Graph Neural Networks to analyze vast amounts of data, allowing it to predict weather conditions up to 10 days in advance with remarkable accuracy. Operating efficiently on Google's TPU v4 machines, producing forecasts in under one minute. This will enhance forecast accuracy and improve early warnings for severe weather events, potentially saving lives and minimizing economic impacts. Back in 2019, here at Velebit AI, we tried to enter the field of applying neural networks to short-term weather prognoses. We were part of an Algebra Lab Startup Incubator with a project for Very short-term precipitation prediction. https://lnkd.in/dPcJpTwB We didn't move past the prototype because we didn't manage to find investment, and we realized that the problem was potentially too big for our small team. I'm glad to see that using AI methods in weather prognosis and modeling has become mainstream. It was the obvious way to go, even for our small team 5 years ago. 😉 🙂 #AIforcast #AIweather #WeatherForcast

A great achievement, no doubt about that. What is unfortunately missing in Google ´ announcement is a recognition that such an achievement would be impossible without high quality re-analysis ( e.g. ECMWF ERA5 as pointed out already by our colleague Dick Dee). All of this would be impossible without WMO basic infrastructure operated by 193 National Meteorological and Hydrological Services, National and international oceanographic organizations, polar research institutions, National and international space agencies and academic institutions. AI and ML approached are only so good as the availability of historical data for training , the real time availability of data for the initial conditions and the on-going availability of data for continuous improvements. Thanks to WMO and its 193 country Members and supporting organizations.

Google DeepMind breaking new ground once again. Accurate 10 day weather forecasting has always been a bit iffy, so using #AI to make it more accurate and deliver forecasts in 45 seconds is 🤯. Next challenge for GDM is to fix the British Summer 🌧 "Instead of the traditional numerical weather prediction methodologies, GraphCast uses machine learning to provide highly accurate and timely weather predictions up to ten days in advance. It harnesses the power of graph neural networks to model complex weather patterns and offers unprecedented precision in forecasts and an enhanced ability to predict extreme weather conditions."

Congratulations to the Google DeepMind team for this recognition from the Royal Academy of Engineering for their work on GraphCast. Our open-sourced GraphCast AI model can deliver 10-day weather forecasts with unprecedented accuracy — especially critical as climate change continues to accelerate, as proven by recent heat waves and natural disasters around Asia-Pacific. Accessible tools like GraphCast can help people prepare early for catastrophic events, and help billions live safer day-to-day.

In the next episode of GT Tech Tips, we dive deep into the capabilities of Gamma Technologies' GT-SUITE and creating dynamic metamodels for system simulation! Watch the episode here: https://lnkd.in/gS3KCYnV Join us as we explore how GT-SUITE's machine learning assistant leverages a recurrent neural network to accurately capture the complex transient behavior of a system. Subscribe to the Gamma Technologies YouTube channel and turn on the notification bell for more insightful episodes! #GTTechTip #GTSUITE #machinelearning #technology #engineering #simulation #neuralnetwork #metamodel

AI enhances weather forecasting by analyzing complex data patterns globally, improving accuracy and efficiency. Utilizing machine learning and data analytics, AI predicts weather conditions, aiding governments in risk mitigation and public safety. Hybrid models blend AI-driven learned physics modules for short-term precision and long-term reliability. Additionally, random forests algorithm enhances predictions by employing multiple decision trees. . . Admission for our AI fundamentals course is open now. 👇 https://lnkd.in/eEBPW7Mw #GIT #ai #weatherAI #aiforecasting #dataanalytics #publicsafety

Along with predictive machine learning software for optimal blast planning and analysing blast outcomes using computer vision, New Gradient has also developed computer vision tools for analysing rock size during post-blast processing. This has been in collaboration with Tarmac at one of its flagship quarries. The implementation of on-demand computer vision equips crusher operators with insights for fine-tuning settings, leading to heightened processing efficiency and a reduction in CO2 emissions. Spearheaded as a component of Tarmac’s IoQ Emerald Challenge-winning initiative, this technological advancement underscores a commitment to fostering sustainability within the mineral products industry. New Gradient takes pride in contributing cutting-edge, tech-driven solutions to assist the industry in achieving its emissions reduction objectives. Curious about how New Gradient can employ artificial intelligence and machine learning to benefit your industry? Reach out to us through the link in our bio! #AI #MachineLearning #Technology #Analytics #Data

Dr. Paul Roebber, Distinguished Professor of Atmospheric Sciences at the University of Wisconsin-Milwaukee and #NMDSI affiliated faculty member, is regarded as one of the leading forecasting experts in the United States. In his recent NMDSI thought leadership article, “The Role of AI and Machine Learning in Weather Forecasting,” he explores how to implement machine learning tools to optimize weather prediction and improve forecast accuracy on a national scale. Read his full piece on our website: https://lnkd.in/ge6PNbXd #nmdsithoughtleader #machinelearning #weatherprediction

𝙂𝙤𝙤𝙜𝙡𝙚'𝙨 𝘼𝙄-𝙀𝙣𝙝𝙖𝙣𝙘𝙚𝙙 𝙒𝙚𝙖𝙩𝙝𝙚𝙧 𝙎𝙞𝙢𝙪𝙡𝙖𝙩𝙤𝙧: 𝘼 𝙇𝙚𝙖𝙥 𝙁𝙤𝙧𝙬𝙖𝙧𝙙 𝙞𝙣 𝘾𝙡𝙞𝙢𝙖𝙩𝙚 𝙈𝙤𝙙𝙚𝙡𝙞𝙣𝙜 Google has introduced an innovative AI-enhanced weather simulator, NeuralGCM. This new model combines conventional physics-driven models with machine learning to enhance the accuracy and efficiency of weather and climate forecasts. It matches the accuracy of traditional forecasts while requiring significantly less computing power. Developed by a Google-led research team and detailed in a recent Nature publication, NeuralGCM focuses on simulating atmospheric temperatures and could be a game-changer for long-term weather predictions and climate modeling. Notably, this model estimates smaller-scale phenomena like clouds and tornadoes that are challenging for general equations to capture. Experts like Texas A&M University’s Professor R. Saravanan recognize the model's potential, especially for predicting weather on sub-seasonal to seasonal scales. Despite its limitations, such as not varying greenhouse gas levels, the model's ability to run on minimal computing resources like a laptop could democratize weather forecasting technology. The development signals a promising direction for using AI in climate science, aligning with ongoing efforts to integrate AI across various scientific disciplines, from materials science to protein folding. #AI #ClimateChange #WeatherForecasting #MachineLearning #GoogleResearch #Innovation #Tech #Sustainability #ClimateModeling #EnvironmentalScience #GreenTech #FutureOfWeather https://lnkd.in/gfPi5r7c