How artificial intelligence is helping two environmental scientists unravel the mysteries of the natural world > News > USC Dornsife

How to measure the cloud? How to write a bee swarm? Machine learning makes it possible to understand complex natural phenomena.

Bee data can be biased and geographically concentrated, with data clusters around cities and near roads, but not in more remote locations.  Machine learning can help fill in the gaps.  (Image source: iStock.)

Bee data can be biased and geographically concentrated, with data clusters around cities and near roads, but not in more remote locations. Machine learning can help fill in the gaps. (Image source: iStock.)

Machine learning is a very specific form of artificial intelligence. Using algorithms designed to learn from experience, machine learning—also known as ML—adapts and increases in effectiveness over time as new data is added. An ML-driven program “learns” from its mistakes and can therefore reduce the time it takes to analyze mountains of data from years to minutes.

Melissa Guzman and Sam Silva use machine learning to find insights into the patterns that underlie the natural world. (Photo: Courtesy of Melissa Guzman and Sam Silva.)

Two recently hired faculty, Melissa Guzman, Gabilan associate professor of biological sciences, and Sam Silva, associate professor of earth sciences, both in USC’s Dornsife College of Letters, Arts and Sciences, are already attracting attention for their use of machine learning to find understanding of what seemed b, unrecognizable – regularities underlying the natural world.

Guzman looks for trends in the migration patterns of bees among our most important pollinators, as well as their community composition.

Silva studies the chemical composition of clouds. Recently named recipients of the Faculty Innovation Award from the USC Dornsife Wrigley Institute for Environmental Studies, both are using their expertise to develop solutions to environmental challenges.

Climate change disrupts bee migration patterns and community formation

California is home to the largest and most diverse population of bees in all of North America. However, as their numbers have declined over the past decade, identifying and protecting safe and sustainable havens for bees has become increasingly important. But according to Guzman, finding a place where they are most likely to thrive is more difficult than you might think. Therefore, it uses machine learning tools to speed up the data cleaning process and isolate and correct incorrect data points from various sources.

“Bumblebees are a completely different type of bee—they’re big, fussy, hairy—and they usually like more temperate areas,” Guzman says. “We want to use life-history traits to understand which species benefit the most from things like climate change, and which are most hindered. One of the things we’ve found with bumblebees is that not every species is in decline.’

AI and Science: Towards More Accurate and Faster Climate Models

The air in Los Angeles is legendary, albeit for the wrong reasons. For Silva, this is an ideal option for his research: analyzing the chemical composition of the atmosphere.

Silva describes clouds as “one of the biggest uncertainties in our understanding of physical climate” because of their complex mix of physics (wind speed and direction) and chemistry (different molecules mixing in the atmosphere). Understanding their behavior is important because of the role they play in reflecting sunlight back into space and global hydrologic cycles. Accurately measuring their location, brightness, and duration is essential to properly understanding and predicting their behavior.

“The chemical composition of clouds and the Earth’s atmosphere is important in almost every aspect of air quality and climate change,” says Silva, who is holding a joint meeting at USC’s Viterbi School of Engineering. “We’re using machine learning to help us sift through the data we have—sometimes it’s a huge amount of partially relevant data—and figure out what’s going on in those areas.”

And what he learns in Los Angeles will unfortunately take on greater significance as conditions in other cities begin to mimic those in Southern California.

“Most cities have a large population, a lot of cars, and they’re not very walkable,” he says. “The chemistry we learn about in Los Angeles is transferable to many other places. What’s happening here has something to do with people’s health and air quality.”

See USC News >> for more on this story

Leave a Reply

Your email address will not be published. Required fields are marked *