Causal inference for time series

Differentiable modelling to unify machine learning and physical models for geosciences

A review article by Chaopeng Shen at UPenn.

‘Differentiable’ refers to the ability to accurately and efficiently calculate gradients with respect to model variables or parameters, enabling the discovery of high-dimensional relationships.

Physics-aware Machine Learning Revolutionizes Scientific Paradigm for Machine Learning and Process-based Hydrology

In this paper, the authors clearly laid out four directions for involving AI in Earth science: 1. physical data-guided ML: extracting information from training data, 2. physics-informed ML: include domain-specific knowledge into training process such as customized loss function, 3. physics-embedded ML: adapt Neural Networks to physical laws such as conservation of mass and momentum, 4. physics-aware hybrid learning: directly combining pure physics-based models, such as numerical methods and hydrology models, with ML models.

image


Table of contents