In December 2017, two years after the Paris climate accord was adopted, French President Emmanuel Macron led government, business and civic leaders in a conference called The One Planet Summit. President Trump, who earlier in the year announced his commitment to withdraw the U.S. from the historic climate accord, was not invited. At this event, Microsoft's President and Chief Legal Officer Brad Smith announced the company would be committing $50 million over the following five years as part of a new strategy to provide access to artificial intelligence (AI) for groups and people who want to use it for the good for the planet. Microsoft's AI for Earth, a program with the goal of using AI to address environmental challenges, launched six months before this announcement.
Announcing its plan to broaden the AI for Earth programme, Microsoft has pledged $50 million over the next five years to put artificial intelligence technology in the hands of those who are working to mitigate climate change. Microsoft rolled out the AI for Earth programme six months ago with an aim to put the power of artificial intelligence towards tackling environmental challenges. "At Microsoft, we believe artificial intelligence is a game changer. Our approach as a company is focused on democratising AI so its features and capabilities can be put to use by individuals and organisations around the world to improve real-world outcomes," Microsoft President and Chief Legal Officer Brad Smith wrote in a blog post on Monday. The announcement came on the eve of the second anniversary of the Paris Agreement.
Then detection and identification of extreme weather events in large-scale climate simulations is an important problem for risk management, informing governmental policy decisions and advancing our basic understanding of the climate system. Recent work has shown that fully supervised convolutional neural networks (CNNs) can yield acceptable accuracy for classifying well-known types of extreme weather events when large amounts of labeled data are available. However, many different types of spatially localized climate patterns are of interest including hurricanes, extra-tropical cyclones, weather fronts, and blocking events among others. Existing labeled data for these patterns can be incomplete in various ways, such as covering only certain years or geographic areas and having false negatives. This type of climate data therefore poses a number of interesting machine learning challenges. We present a multichannel spatiotemporal CNN architecture for semi-supervised bounding box prediction and exploratory data analysis. We demonstrate that our approach is able to leverage temporal information and unlabeled data to improve the localization of extreme weather events. Further, we explore the representations learned by our model in order to better understand this important data. We present a dataset, ExtremeWeather, to encourage machine learning research in this area and to help facilitate further work in understanding and mitigating the effects of climate change. The dataset is available at extremeweatherdataset.github.io and the code is available at https://github.com/eracah/hur-detect.