CS&E’s New AI for Earth Program Addresses Environmental Challenges
The Department of Computer Science & Engineering (CS&E) hosted a new program this summer led by professor Vipin Kumar and administered by lecturer and research associate Eman Ramadan. AI for Earth Summer School is an eight-week program that introduces the potential of artificial intelligence (AI) and machine learning (ML) research to students from across the country. This program aims to apply AI/ML research to help address environmental challenges, providing hands-on machine learning experience in the process.
The summer school was made possible by funding from the National Science Foundation (NSF), which was a supplement to an existing award “Advancing Deep Learning for Inverse Modeling Project” through Kumar’s research group. The first seven weeks of the program were hosted remotely, with the final week hosted on campus. Collaborating with different services across the University of Minnesota, AI for Earth provided housing for students coming from out of state and coordinated flights and meal plans for incoming students.
“Originally, our vision was to organize a summer school that would open the door for students who may not yet have experience in machine learning, AI, or environmental science,” Eman explained. “A key goal is to create a space where computer science students and non-computer science students can learn side by side. By involving students from outside computer science, we hope to give them meaningful, hands-on experience at the intersection of two scientific disciplines and demonstrate how machine learning can be applied to pressing environmental challenges.”
“The summer school curriculum featured two weekly virtual lectures, designed and delivered by University of Minnesota faculty members from diverse disciplines alongside PhD students. The program culminated in a week-long, in-person session with guest lectures and career-building activities, including industry and graduate panels. To deepen learning and foster collaboration, each participant joined a multidisciplinary team tasked with developing a project that addressed a real-world environmental challenge.”
“My project was to predict precipitation, but our smaller section was to predict two meter air temperature,” Emilia Szynwald, a computer science student from Hofstra University who participated in this program, said. “That is the best indicator for how people feel on ground level without having to take into account the effects of the heating and cooling of the earth’s surface. The benefit is that it can predict any extreme weather that is going to happen. It’s applicable to see when the climate is changing and any indication that might help us prevent future climate issues.”
“My project involved data from this thing called FlexNet, which essentially consists of a bunch of data gathered from 200 different locations around the world,” Sergey Barabanoff, a University of Minnesota computer science and biomedical student who participated in this program, said. “They use these towers to measure precipitation and how much carbon dioxide is present in that environment. Using a model like this, we are able to see if certain areas are making more carbon dioxide than they should be, or if they are contributing to lowering the amount of carbon dioxide in the atmosphere. It helps us see if a certain area is being detrimental or beneficial to the climate.”
“The University of Minnesota’s program combined both online and in-person components to maximize flexibility and engagement. The online portion featured recorded lectures and accompanying exercises, allowing students to learn at their own pace and build foundational knowledge ahead of the final week. The in-person component emphasized hands-on experience, enabling participants to apply machine learning techniques to real-world environmental problems while collaborating closely with peers on team projects.”
“The professors I was interacting with during this program were very interested in environmental impacts going on and trying to see how you can use computer science and machine learning to try and understand any climate issues going on,” Barabanoff said. “It was interesting to see different perspectives of how computer science could be applicable in other areas, and also coming from a computer science background, how they can understand the problems we’re trying to solve on a different level.”
“Preparing the materials and projects was an exciting and rewarding experience,” Eman said. “We created opportunities for students to learn independently and then come together to collaborate in teams. Through this process, they developed a genuine passion for tackling real-world challenges—motivated to apply their computer science skills, or even acquire new ones if they lacked prior experience—to make a meaningful impact on the environment and the communities where we live.”
The goal is to open this summer school to participants worldwide, giving them the opportunity to apply their existing knowledge to solving problems they care about most. For many students, the program also provided a glimpse into what graduate school could entail, offering valuable insights and experience for those interested in pursuing their studies beyond the undergraduate level. CS&E will be hosting this summer school again next year—stay tuned for more details and opportunities to get involved!