Learning Variational Data Assimilation Models and Solvers
In one sentence. End-to-end neural variational data assimilation, with learned model and solver, outperforms classical gradient descent on Lorenz-63 and Lorenz-96 reconstructions.
Abstract
This paper addresses variational data assimilation from a learning point of view. Data assimilation aims to reconstruct the time evolution of some state given a series of observations, possibly noisy and irregularly-sampled. Using automatic differentiation tools embedded in deep learning frameworks, we introduce end-to-end neural network architectures for data assimilation. It comprises two key components: a variational model and a gradient-based solver both implemented as neural networks. A key feature of the proposed end-to-end learning architecture is that we may train the NN models using both supervised and unsupervised strategies. Our numerical experiments on Lorenz-63 and Lorenz-96 systems report significant gain w.r.t. a classic gradient-based minimization of the variational cost both in terms of reconstruction performance and optimization complexity. Intriguingly, we also show that the variational models issued from the true Lorenz-63 and Lorenz-96 ODE representations may not lead to the best reconstruction performance. We believe these results may open new research avenues for the specification of assimilation models in geoscience.
Labels
- Topic
- Data assimilation
- Type
- New model or method
- Method
- Differentiable physical model
- Data
- Other
- Status
- Research
- Identifiers
- arXiv:2007.12941 DOI:10.1029/2021ms002572
- Added
- 2026-10-22
Labels and summary are generated by Claude and reviewed weekly. Wrong? Edit this entry.