Past Recording
Share
Star()
Deep Unsupervised Learning for Climate Informatics
Thursday Mar 4 2021 17:00 GMT
Please to join the live chat.
Deep Unsupervised Learning for Climate Informatics
Why This Is Interesting

Prediction of Global Climate Change is an important problem for adaptation, but Global Climate Models still have many errors/biases and the resolution is too low for impact modeling. At the same time, bias-correction and “downscaling” (upsampling) can be framed as classic ML problems.

Discussion Points
  • Bias-correction and “downscaling” (upsampling)
  • Supervised vs. unsupervised bias-correction/downscaling
  • Domain Alignment, Normalizing Flows and Variational Autoencoders
  • Interpretation of Latent Space
Time of Recording: Thursday Mar 4 2021 17:00 GMT