Learning 4DVAR Inversion Directly From Observations

Arthur Filoche (1) , Julien Brajard (2, 3) , Anastase Charantonis (4, 5) , Dominique Béréziat (1)

1 PEQUAN - Performance et Qualité des Algorithmes Numériques
2 LOCEAN - Laboratoire d'Océanographie et du Climat : Expérimentations et Approches Numériques
3 NERSC - Nansen Environmental and Remote Sensing Center [Bergen]
4 LaMME - Laboratoire de Mathématiques et Modélisation d'Evry
5 ENSIIE - Ecole Nationale Supérieure d'Informatique pour l'Industrie et l'Entreprise

Variational data assimilation and deep learning share many algorithmic aspects in common. While the former focuses on system state estimation, the latter provides great inductive biases to learn complex relationships. We here design a hybrid architecture learning the assimilation task directly from partial and noisy observations, using the mechanistic constraint of the 4DVAR algorithm. Finally, we show in an experiment that the proposed method was able to learn the desired inversion with interesting regularizing properties and that it also has computational interests.

Keywords: Data Assimilation; Unsupervised Inversion; Differentiable Physics; Learning Simulation

Domains: Image Processing [eess.IV]

Arthur Filoche, Julien Brajard, Anastase Charantonis, Dominique Béréziat. Learning 4DVAR Inversion Directly From Observations. 2022. ⟨hal-03863390v1⟩