Generative Moment Matching Autoencoder with Perceptual Loss

Ning

Earth and Space Science

et al. 2023

Accurately inverting global and regional subsurface temperature (ST) by multisource satellite observations is a challenging but hot topic. This study proposes a new method to invert daily ST from the sea surface information in China's marginal seas based on generative adversarial network (GAN) model. The proposed GAN‐based model can project the STs from sea surface information (SLA, SSTA, SST) with a high resolution of 1/12°. A traditional regression‐based model, Modular Ocean Data Assimilation System (MODAS), is set up same experiments for comparison. The results show that the averaged root mean square error results are less than 1.45°C in the upper 200 m and the highest averaged R2 of 0.97 at the 70 m level, which is better than that of MODAS. Errors analysis and typical oceanographic phenomena analysis results show the superiority of the proposed GAN‐based model in this study. This study can provide high‐precision daily ST data from sea surface information, which can be expanded to further studies on the interior ocean variation characteristics.

Section: Methodsmentioning

confidence: 99%

Deriving Sea Subsurface Temperature Fields From Satellite Remote Sensing Data Using a Generative Adversarial Network Model

Ning

Earth and Space Science

et al. 2023

“…Using the analysis presented in Li et al [2015], Kiasari et al [2017], it can be shown that, for a Gaussian kernel, the moments (of all order) of the PCE response will be close to the moments of the output data y. The proposed model is depicted in Figure 1.…”

Section: Pce-netmentioning

confidence: 95%

“…For example, when the kernel K(•, •) is linear, E η [K] = µ η is simply the mean of η. Thus, choosing a Gaussian kernel allows us to capture high-order moments, and MMD acts as a moment matching approach, see Li et al [2015], Kiasari et al [2017] for details.…”

Section: Maximum Mean Discrepancymentioning

confidence: 99%

PCENet: High Dimensional Surrogate Modeling for Learning Uncertainty

Shustin¹,

Ubaru²,

Kalantzis³

et al. 2022

Preprint

Learning data representations under uncertainty is an important task that emerges in numerous machine learning applications. However, uncertainty quantification (UQ) techniques are computationally intensive and become prohibitively expensive for high-dimensional data. In this paper, we present a novel surrogate model for representation learning and uncertainty quantification, which aims to deal with data of moderate to high dimensions. The proposed model combines a neural network approach for dimensionality reduction of the (potentially high-dimensional) data, with a surrogate model method for learning the data distribution. We first employ a variational autoencoder (VAE) to learn a low-dimensional representation of the data distribution. We then propose to harness polynomial chaos expansion (PCE) formulation to map this distribution to the output target. The coefficients of PCE are learned from the distribution representation of the training data using a maximum mean discrepancy (MMD) approach. Our model enables us to (a) learn a representation of the data, (b) estimate uncertainty in the high-dimensional data system, and (c) match high order moments of the output distribution; without any prior statistical assumptions on the data. Numerical experimental results are presented to illustrate the performance of the proposed method.

“…In deep generative networks, some data generation losses are optimized in a grid-based manner (grid-to-grid), so generated data typically lack high-frequency details; thus, the perceptual difference between the real data and the generated data is not understood (Kiasari et al, 2017;Wang et al, 2018). Our study area covers a large area of China's marginal seas, including the BHS, YS and ECS, with a high spatial resolution.…”

Section: St Inversion Via Sea Surface Information-guided Gan (Ssig-g)mentioning

confidence: 99%

Deriving Sea Subsurface Temperature Fields from Satellite Remote Sensing Data Using a Generative Adversarial Network Model

Ning

et al. 2022

Preprint

Ingenious use of multisource satellite observations to accurately invert global and regional subsurface thermohaline structure is essential for understanding ocean interior processes, but extremely challenging. This study proposes a new method from the sea surface information inverting daily subsurface temperature (ST) based on generative adversarial network (GAN) model in China's marginal seas. The proposed GAN-based model can project the STs from sea surface information (SLA, SSTA, SST) with a high resolution of 1/12°. A traditional regression-based model, Modular Ocean Data Assimilation System (MODAS), is set up the same experiments for the sake of comparison. The results show that the averaged RMSE results are less than 1.45°C in upper 200m and the highest averaged R2 of 0.97 at the 70m level, which are better than that of MODAS. Errors analysis and typical oceanographic phenomena analysis results show the superiority of the proposed GAN-based model in this study. This study can provide high-precision daily ST data from sea surface information, which can be expanded to further studies on the ocean interior variation characteristics.