Wednesday, April 27, 2022 - 4:00pm to 4:30pm
Event Calendar Category
LIDS & Stats Tea
Building and Room Number
Would a global carbon tax reduce the flood risk at MIT? The answer to this question of local impact and risk is critical for policy making or climate-resilient infrastructure development. But, localized climate models are computationally too expensive for robust uncertainty quantification or exploring many policy scenarios. In this talk, I will introduce The Climate Pocket: a climate education simulation that leverages fast machine learning (ML)-based climate models to illustrate local science-based flood impacts of global climate policy decisions, as shown in http://trillium.tech/eie. While ML is promising copies of climate models, called surrogates/reduced-order/or pruned models, that are order of magnitudes faster, ML must verify physical constraints to be trusted. I will answer the question of trust by focusing on our novel model, called multiscale neural operator, that learns surrogates of multiscale partial differential equations while using known large-scale physics as a prior and focusing the learning on the hard-to-model small-scale physics.
Björn Lütjens is a PhD Candidate at the Human Systems Laboratory in the MIT Department of Aeronautics and Astronautics. His research is tackling climate change with machine learning, little-by-little, together with Prof. Dava Newman, Cait Crawford, and Chris Hill. His research has won grants by NSF, Climatechange.ai, ESA, Portugal Space, NASA, IBM, MIT MSCS, MIT Pkg, MIT Legatum and hardware grants by Microsoft and NVIDIA. He has advised two teams of senior researchers at the NASA/SETI Frontier Development Lab, co-founded the ForestBench Consortium, interned with IBM Future of Climate and BRT (John Deere), earned an M.Sc. from MIT in safe and robust deep reinforcement learning with Prof. How , and a B.Sc. from TUM in Engineering Science. He also windsurfs poorly, jams, jokes, and loves meeting new people.