What if all the data we need to prepare for disasters, better plan our urban environments, and protect our food supply is already collected? What if we just lack the tools to effectively analyze all the data that's already been collected?
An Arizona State University doctoral student is at the forefront of artificial intelligence (AI) research designed to help us act on what we know.
This June, leading data science researchers and graduate students gathered at the Dedan Kimathi University of Technology campus in Nyeri, Kenya, to tackle some of the planet's most pressing issues, including food insecurity, environmental protection and climate change.
Data Science Africa (DSA) hosted its annual Summer School, bringing together hundreds of students, industry professionals and academics from around the world to collaborate on the event's theme, “Data Science for Social Impact in the Era of Generative AI.”
Gedeon Muhawenayo, a doctoral student in the Department of Computing and Augmented Intelligence, part of ASU's Ira A. Fulton College of Engineering, attended the event representing the Kerner lab and described his team's work as part of the NASA Harvest program, a consortium of leading scientists and agricultural practitioners bringing together their expertise to ensure food security.
Hannah Kerner, an assistant professor of computer science and engineering at the Fulton School, develops tools to analyze vast amounts of Earth observation data and develop actionable plans. In her lab, her team uses a type of AI called machine learning to process huge sets of information.
One of the group's tasks is to feed satellite imagery to a fleet of computers and teach the AI system to recognize the types of crops being grown in the images. Once the AI can identify the crops, it will be able to create meaningful maps when it receives new images. This process allows experts to know how much food is produced in an area and whether it is enough to feed the population.
Satellite imagery showing crop fields in Nyagatare, Rwanda, taken on June 15, 2024. Muhawenayo is working on a form of AI called machine learning, creating a system to train computers to identify crop types from satellite observations. Photo by Hannah Kerner, adapted from European Space Agency's Sentinel-2 satellite image.
From satellites to social impacts
Kerner brought Muhawenayyo onto the team after working at the Rwanda Space Agency, where he developed machine learning systems for use in sustainable agriculture programs.
“Gedeon is an invaluable addition to the lab,” Kerner said. “He has the drive, dedication, and technical ability to advance AI methods that deliver meaningful results for real people, not just papers. I am confident that he will continue his innovative, societally impactful work in machine learning and beyond.”
Muhawenayyo is a rising expert in the field of machine learning for remote sensing, where he develops AI tools to study and process data collected by satellites and drones, including electromagnetic radiation, visible light, infrared and microwave signals.
“We started looking at ways to extract useful features from satellite imagery and quickly realized that machine learning algorithms were the key to making this happen,” Muhawenayyo says.
For Muhawenayoh, who is originally from Kigali, Rwanda, the mission to ensure food security holds special meaning.
“Many African countries are prone to a variety of natural disasters, putting a strain on the most vulnerable communities, especially smallholder farmers,” he says. “Furthermore, agriculture is a major economic sector, and advances in machine learning can support ways to increase yields for these communities, especially as they face challenges posed by climate change.”
Muhawenayyo joined the Kerner Institute after earning a master's degree in machine intelligence from the African Institute of Mathematics in Accra, Ghana.
“I was particularly impressed by the focus on user-driven research and the practical application of cutting-edge research,” he adds. “I am excited to be part of a research program that leverages technology for the benefit of society.”
Muhawenayyo arrives at the DSA 2024 Nyeri Summer School event, where he showed fellow graduate students how to process large amounts of Earth observation data during the Machine Learning for Remote Sensing workshop. Photo by Gedeon Muhawenayyo
Globalization and engagement
At the DSA Summer School event, Muhawenayo led a machine learning in remote sensing workshop, training graduate students on how to handle large amounts of Earth observation data. During his talk, he introduced students to sources of satellite data, explained how to budget for a research project, and introduced them to programming tools to process imagery. He also conducted a hands-on session that allowed participants to create their own cropland maps.
Muhawenayo’s work is part of the Fulton School’s ongoing effort to create world-class graduate programs and attract talented students from around the world.
Ross Maciejewski, dean of the School of Computing and Augmented Intelligence, said the outreach efforts are a key area of focus for faculty in the department.
“The collaboration between Hannah Kerner and Gedeon is a fantastic blend of technical insight and specialized and local knowledge that will produce real, practical results,” says Maciejewski. “Creating a home for excellent doctoral students is an important part of our research and academic mission.”
Meanwhile, Muhawenayo said he was inspired by the DSA 2024 Nyeri Summer School events and believes students today are serious about solving the world’s toughest problems.
“The participation of younger generations in the global community gives me hope that we can make a meaningful difference,” said Muhawenayoh. “I believe we can create a sustainable future and effectively address the adverse effects of climate change.”
