Data Assimilation Predictive GAN (DA-PredGAN) Applied to a Spatio-Temporal Compartmental Model in Epidemiology.

Autor: Silva VLS; Applied Modelling and Computation Group, Imperial College London, London, UK.; Department of Earth Science and Engineering, Imperial College London, London, UK., Heaney CE; Applied Modelling and Computation Group, Imperial College London, London, UK.; Department of Earth Science and Engineering, Imperial College London, London, UK., Li Y; Department of Earth Science and Engineering, Imperial College London, London, UK., Pain CC; Applied Modelling and Computation Group, Imperial College London, London, UK.; Department of Earth Science and Engineering, Imperial College London, London, UK.; Data Assimilation Laboratory, Data Science Institute, Imperial College London, London, UK.
Jazyk: angličtina
Zdroj: Journal of scientific computing [J Sci Comput] 2023; Vol. 94 (1), pp. 25. Date of Electronic Publication: 2022 Dec 28.
DOI: 10.1007/s10915-022-02078-1
Abstrakt: We propose a novel use of generative adversarial networks (GANs) (i) to make predictions in time (PredGAN) and (ii) to assimilate measurements (DA-PredGAN). In the latter case, we take advantage of the natural adjoint-like properties of generative models and the ability to simulate forwards and backwards in time. GANs have received much attention recently, after achieving excellent results for their generation of realistic-looking images. We wish to explore how this property translates to new applications in computational modelling and to exploit the adjoint-like properties for efficient data assimilation. We apply these methods to a compartmental model in epidemiology that is able to model space and time variations, and that mimics the spread of COVID-19 in an idealised town. To do this, the GAN is set within a reduced-order model, which uses a low-dimensional space for the spatial distribution of the simulation states. Then the GAN learns the evolution of the low-dimensional states over time. The results show that the proposed methods can accurately predict the evolution of the high-fidelity numerical simulation, and can efficiently assimilate observed data and determine the corresponding model parameters.
Competing Interests: Conflict of interestThe authors declare that they have no conflict of interest.
(© The Author(s) 2022.)
Databáze: MEDLINE
Externí odkaz: