Databases of extreme weather events identified in climate simulation models are of key importance for quantifying and projecting into the future the effects of climate change, including the risk it imposes on human livelihoods and activity. A major challenge in creating these databases is accurately and rapidly identifying such events, including atmospheric rivers, hurricanes, and tropical storms, in large-scale climate simulation datasets. This project aims to create the first of its kind, large-scale dataset of human-labeled extreme weather events to aid training of machine learning models for tasks such as detection and segmentation. Part of the research will be in designing the data labeling strategy (e.g., using concepts from active learning) as to most effectively leverage the time of both domain experts as well as non-experts that will contribute to the labeling effort. Another part of the research will require developing intuitive interfaces and annotation software specific to the data types found in climate modeling.
– Comfortable with Python programming (object-oriented programming, numpy, matplotlib, pandas etc.)
– Familiarity with modern machine learning frameworks such as PyTorch or TensorFlow
– Knowledge and familiarity with modern machine learning, including deep learning. At a minimum, class work and projects, but previous research work using/developing deep learning methods preferred (ideally having worked with GANs)
-Interest in scientific applications of AI including one or more of climate, energy, materials