Google's DeepMind can predict the changing winds one day in advance

Google and DeepMind have started testing the machine learning of Google's own wind turbines, which are part of the company's renewable energy projects. Since last year, they have provided weather forecasts and data from existing wind turbines to DeepMind's machine learning platform, which was producing wind forecasts with 36 hours ahead of production. real energy. Google could then make supply commitments to power grids one full day before delivery. This predictability allows power systems to rely more easily on wind power, which increases Google's wind energy value by 20%.

Not only does this explanation about how machine learning can promote the adoption of wind energy, it is also an example of how to effectively use machine learning. : solve critical problems and not just jump in your text thread to recommend a restaurant when you talk about tapas. For DeepMind, this is a leading use of its technology and evidence that it is not only useful for beating professionals. StarCraft II players.