AOSC Seminar by Dr. Maria Molina, 9/9/2021

AOSC Seminar

Dr. Maria Molina

NCAR

Title: Deep Learning across Weather and Climate Scales: Opportunities, Challenges, and a Future Outlook

Abstract:

Deep learning is a subfield of machine learning that includes multi-layered neural networks, which can enable computers to learn complex patterns and relationships in data. Recent computational and deep learning advances present an opportunity to answer pressing questions in the Earth sciences and improve societally relevant forecasts across time scales. Research has shown promising results, including skillful classification of severe convective storms and the ability to learn precipitation forecast biases at subseasonal lead times (weeks 3-4). However, deep learning is susceptible to errors when encountering outlier events or when the training data exhibits nonstationarity, which are limitations that can be manifested by extreme weather events or climate trends. This seminar will highlight applications of deep learning for weather and climate, propose solutions to identified challenges, and share a future outlook for the use of deep learning for Earth science.

Contact: Jonathan Poterjoy

AOSC Seminar

Pre-seminar refreshment: N/A
Seminar: 3:30-4:30pm, Room: ATL 2400
Meet-the-Speaker: 4:30-5:00pm, Room: ATL 3400 [For AOSC Students only]

[ Subscribe to receive seminar announcements | Email aosc-deptseminar@umd.edu to give a seminar ]