Data, Microsoft President Brad Smith wrote in an April 2019 blog post, “is a critical part of our work and a global transition to a low-carbon future.”
All of the signs point to Earth facing a climate crisis; fortunately, there are tools and technology that could help. Big data, along with the machine learning and artificial intelligence needed to transform it into actionable information, can give us a leg up on our environmental problems–provided we use it.
Microsoft wants to make accessing powerful AI and machine learning technology more practical for scientists and environmental researchers, which is why it started AI for Earth in 2017. In the mere two years since Microsoft founded AI for Earth, it has awarded over 230 grants to teams in 63 countries, covering four research areas: Climate change, agriculture, biodiversity, and waters.
Grants come in two forms: Data labeling services, which help get large datasets ready for AI processing, and Microsoft Azure computer credits. Grant applicants are able to apply for one or both forms of grants.
Microsoft has built two APIs specifically for AI for Earth, and has stated that it continues to work on more. Currently available is a land coverage mapping API and a species classification API. The APIs are available to anyone who wants to use them.
AI for Earth projects
With over 230 grants being awarded, there’s a huge number of projects going on around the globe. From apps average citizens can use to take pictures of wildlife for AI to classify to anti-poaching measures in Africa, the uses for AI in protecting natural resources is vast.
FarmBeats, a project that has been ongoing since 2015, is working on ways to improve agriculture using AI, edge computing, and IoT technology. FarmBeats wants to establish an end-to-end platform that takes the manual tech work and know-how out of the farmer’s hands. By using sensors and automated drones, FarmBeats wants to give farmers the tools they need to plant more efficiently, using less resources and getting higher yields.
Project Premonition aims to use drones and other automated robots to find, trap, and gene sequence mosquitoes to prevent the spread of diseases like malaria, Zika, and Ebola. It uses drones to map possible mosquito locations and deploys robots to trap the mosquitoes and deliver them to a lab, where scientists can sequence their genes. The data gathered by gene sequencing doesn’t focus on a particular pathogen, and instead gives scientists data like the types of animals they’ve bitten and any possible pathogen exposure.
SilviaTerra used its AI for Earth grant to develop a number of tools for forestry professionals looking to conserve woodlands more efficiently. Built on the back of satellite imaging that predicts tree populations, SilviaTerra’s software makes forestry mapping trips more efficient by improving routes and equipment inventories for scientists in the field.
Putting important data in one place
Science generates data, but collating and transforming that data into action can be difficult. When you add to that the sheer volume of data available, things get even more difficult, especially if all of that data isn’t available to everyone who may need it.
AI for Earth takes a step toward rectifying that by offering services to organizations that may not be able to access them otherwise, and by centralizing data from government research agencies from around the globe in Azure cloud.
SEE: The Internet of Wild Things: Technology and the battle against biodiversity loss and climate change (TechRepublic cover story) | Download the free PDF version (TechRepublic)
In addition, grant applicants that apply for data labeling have their data published on Microsoft Azure for the use of other researchers, further expanding the base of knowledge needed to turn AI into a tool for improving the planet.
Research takes time, and the impact of a project like AI for Earth isn’t immediately available. However, with all the data it’s hosting and projects it has launched, Microsoft Azure could be the place to go for scientists looking for the data they need to make an impact.