Atomospheric Deep Learning

cloud segmentation

Input -> Bootstraping samples -> training -> Post processing -> Output

In the post processing step, the author utilized:

  1. local boundary effect
  2. test time augumentation
  3. Gaussian smoothing

which derived from Kaggle competition (3rd place)

Deep Learning in general

What Do We Understand About Convolutional Networks?

94 pages overview of Convolutional Nerual Network: from technologies to research scope.

Deep learning in SPPs

Sadeghi, M., A.A. Asanjan, M. Faridzad, P. Nguyen, K. Hsu, S. Sorooshian, and D. Braithwaite, 2019: PERSIANN-CNN: Precipitation Estimation from Remotely Sensed Information Using Artificial Neural Networks–Convolutional Neural Networks. J. Hydrometeor., 20, 2273–2289, https://doi.org/10.1175/JHM-D-19-0110.1

Author used CNN to train GOES satellite to map the precipitation from IR band and water vapor channel.

Interepretable Machine Learning/Deep Learning

McGovern, A., R. Lagerquist, D. John Gagne, G. E. Jergensen, K. L. Elmore, C. R. Homeyer, and T. Smith, 2019: Making the Black Box More Transparent: Understanding the Physical Implications of Machine Learning. Bull. Amer. Meteor. Soc., 100, 2175–2199, https://doi.org/10.1175/BAMS-D-18-0195.1.

In this article, the authors emphasized the ways of interpretating “black box” model in atmospheric predictions to strengthen descision making.

Deep Learning in Hydrology

A Transdisciplinary Review of Deep Learning Research and Its Relevance for Water Resources Scientists

A nice review paper, introducing concepts of deep learning and machine learning and also their applications in hydrologic fields.