UNIVERSITY PARK, Pa. — The Tackling Climate Change with Machine Learning Workshop at the International Conference on Learning Representations (ICLR), which took place in Vienna, Austria, and virtually in May, presented a best paper award to a group researching and using machine learning to forecast weather.
Romit Maulik, Penn State Institute for Computational and Data Sciences (ICDS) co-hire and assistant professor in the College of Information Sciences and Technology, was a co-author on the winning paper, “Scaling transformer neural networks for skillful and reliable medium-range weather forecasting.”
“This has been a year-long collaboration between the University of California-Los Angeles, Carnegie Mellon University, Argonne National Laboratory and Penn State,” Maulik said.
The research investigated the use of modern artificial intelligence (AI) tools for forecasting as compared to classical methods currently in use for operational forecasts.
“It’s a paradigm shift from looking at the classical forecasts provided by several agencies,” Maulik said. “Those forecasts are typically obtained with very large computing resources, and it can be computationally costly. We thought, what if we took an alternative route?”
Maulik described that the AI model, based on computer vision techniques, takes data from historical information such as archival forecasts and satellite images to learn weather patterns.
“Then, a trained model can make forecasts in real time, without requiring access to very large computational resources,” Maulik said. “Once the neural networks are trained and released, the model deployment can be done effectively on a laptop and, eventually, on increasingly smaller resources such as cell phones.”
ICLR has several workshops on AI subtopics, Maulik said, where researchers can present their papers and get feedback.
“Getting accepted into a workshop, which is quite competitive, maximizes the paper’s visibility,” Maulik said. “It helps us get good feedback from both the AI and the domain sciences community and significantly improve our methods. The award itself is great; it validates our hard work. However, our long-term goal remains the same. We want to find ways to improve our current models and provide a viable competitor to classical weather forecasting approaches.”
One of the researchers' goals is to be able to more effectively forecast weather extremes, which current models may struggle to do, according to Maulik.
“Our eyes are set on grander challenges,” Maulik said. “That being said, as computational scientists, we want to solve the problem and we think of the tool after. We try to balance classical and machine learning methods and are not partial to either.”
The authors and collaborators on the article include Maulik, Tung Nguyen, Rohan Shah, Hritik Bansal, Troy Arcomano, Veerabhadra Rao Kotamarthi, Ian Foster, Sandeep Madireddy and Aditya Grover.