Institute
Penn State University
Time & Place
Thu, 20 Jul 2023 16:00:00 NZST in ER 263
Abstract
Zoom Recording: Link
In this talk I will showcase research related to the generation of weather Analogs using the AnEn technique. Applications will include selected examples of how AnEn was used to forecast renewable energy generation. Some of the main challenges include the definition of a multi-dimensional similarity matrix based on deep learning for the analysis of full numerical model output realizations, and the generation of analogs for which no past similar conditions exist. This is a particularly important problem because of new weather and climate patters that are significantly different than what was observed in the past. Finally, I will talk about the importance of sustainable computing practices, especially when power-hungry methods like deep learning are employed.
Biography
Guido Cervone is E. Willard and Ruby S. Miller Professor, Geography, Meteorology & Atmospheric Science and Associate Director of the Institute for Computational and Data Sciences at Penn State. He has been at Penn State for nine winters, after living for about 20 years in Washington D.C. He is the elected President of the AGU Natural Hazards section.He also holds the appointments of Erskine Fellow, University of Canterbury, New Zealand, Affiliate Scientist, National Center for Atmospheric Research, Boulder, CO, and Affiliate Professor, Instituto Sant’ Anna, Italy.
He received the Ph.D. in Computational Science and Informatics in 2005, and M.S. and B.S. in Computer Science in 2000 and 1998. His expertise is in geoinformatics, machine learning and remote sensing, and his research focuses on the anomaly detection and prediction of rare and extreme events, especially associated with natural hazards and renewable energy. He sailed 5000 offshore miles, and uses an electric unicycle for his daily commute.