Deep Learning for Climate Science
Climate change poses a major challenge for humanity in the 21st century. Characterizing the impact of climate change requires sophisticated analysis of complex datasets produced by climate simulations. In recent years, we have shown that Deep Learning based techniques are capable of classifying, detecting and segmenting extreme weather patterns in O(10) TB sized datasets. We have successfully scaled Deep Learning architectures to the largest CPU- and GPU-based HPC systems in the world. This talk will review our latest results, and conclude with efforts to create ClimateNet - an open labeled dataset and reference architecture for collaboration with the broader community.
- Deep Learning can be applied to climate science problems
- Analyzing large datasets requires HPC systems
- Access to labeled data is a key challenge
Prabhat leads the Data and Analytics Services team at NERSC; his group is responsible for supporting over 7000 scientific users on NERSC’s HPC systems. His current research interests include Deep Learning, Machine Learning, Applied Statistics and High Performance Computing. Prabhat has co-authored over 150 papers spanning several domain sciences and topics in computer science. He has won 5 Best Paper Awards, 3 Industry Innovation Awards, and he was a part of the team that won the 2018 Gordon Bell Prize for their work on ‘Exascale Deep Learning’.