Learning to Assimilate in Chaotic Dynamical Systems

Abstract

The accuracy of simulation-based forecasting in chaotic systems is heavily dependent on high-quality estimates of the system state at the time the forecast is initialized. Data assimilation methods are used to infer these initial conditions by systematically combining noisy, incomplete observations and numerical models of system dynamics to produce effective estimation schemes. We introduce amortized assimilation, a framework for learning to assimilate in dynamical systems from sequences of noisy observations with no need for ground truth data. We motivate the framework by extending powerful results from self-supervised denoising to the dynamical systems setting through the use of differentiable simulation. Experimental results across several benchmark systems highlight the improved effectiveness of our approach over widely-used data assimilation methods.

References in corpus (10)

• Auto-Encoding Variational Bayes
• Prevalence of Neural Collapse during the terminal phase of deep learning training
• Deep Learning for Post-Processing Ensemble Weather Forecasts
• Solver-in-the-Loop: Learning from Differentiable Physics to Interact with Iterative PDE-Solvers
• Noise2Self: Blind Denoising by Self-Supervision
• Auto-Encoding Sequential Monte Carlo
• Recurrent Kalman Networks: Factorized Inference in High-Dimensional Deep Feature Spaces
• A case for new neural network smoothness constraints
• Mechanics and Thermodynamics of a New Minimal Model of the Atmosphere
• Imputing Missing Events in Continuous-Time Event Streams