Super-resolution and uncertainty estimation from sparse sensors of dynamical physical systems

The goal of this study is to leverage emerging machine learning (ML) techniques to develop a framework for the global reconstruction of system variables from potentially scarce and noisy observations and to explore the epistemic uncertainty of these models. This work demonstrates the utility of exploiting the stochasticity of dropout and batch normalization schemes to infer uncertainty estimates of super-resolved field reconstruction from sparse sensor measurements. A Voronoi tessellation strategy is used to obtain a structured-grid representation from sensor observations, thus enabling the use of fully convolutional neural networks (FCNN) for global field estimation. An ensemble-based approach is developed using Monte-Carlo batch normalization (MCBN) and Monte-Carlo dropout (MCD) methods in order to perform approximate Bayesian inference over the neural network parameters, which facilitates the estimation of the epistemic uncertainty of predicted field values. We demonstrate these capabilities through numerical experiments that include sea-surface temperature, soil moisture, and incompressible near-surface flows over a wide range of parameterized flow configurations.

• Publication year

  2023

• Venue

  Frontiers in Water

• Publication date

  2023-03-14

• Fields of study

  Not labeled

• Identifiers
  DOI 10.3389/frwa.2023.1137110
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Super-resolution generative adversarial networks of randomly-seeded fields

  2022cited by this paper
• A robust super-resolution reconstruction model of turbulent flow data based on deep learning

  2022cited by this paper
• Super Resolution for Turbulent Flows in 2D: Stabilized Physics Informed Neural Networks

  2022influential reference
• Reduced Order Modeling Using Advection-Aware Autoencoders

  2022cited by this paper
• Capturing Model Uncertainty with Data Augmentation in Deep Learning

  2022cited by this paper
• A curated dataset for data-driven turbulence modelling

  2021influential reference
• Global field reconstruction from sparse sensors with Voronoi tessellation-assisted deep learning

  2021cited by this paper
• Using Physics-Informed Generative Adversarial Networks to Perform Super-Resolution for Multiphase Fluid Simulations

  2021influential reference
• A novel framework for cost-effectively reconstructing the global flow field by super-resolution

  2021cited by this paper
• Physics-informed machine learning

  2021cited by this paper
• Global soil moisture data derived through machine learning trained with in-situ measurements

  2021influential reference
• Understanding and Mitigating Gradient Flow Pathologies in Physics-Informed Neural Networks

  2021cited by this paper
• Solution of Physics-based Bayesian Inverse Problems with Deep Generative Priors

  2021cited by this paper
• Data-driven reduced order modeling of environmental hydrodynamics using deep autoencoders and neural ODEs

  2021cited by this paper
• A framework for data-driven solution and parameter estimation of PDEs using conditional generative adversarial networks

  2021cited by this paper
• Using physics-informed enhanced super-resolution generative adversarial networks for subfilter modeling in turbulent reactive flows

  2021cited by this paper
• Neural Ordinary Differential Equations for Data-Driven Reduced Order Modeling of Environmental Hydrodynamics

  2021cited by this paper
• SalsaNext: Fast, Uncertainty-Aware Semantic Segmentation of LiDAR Point Clouds

  2020cited by this paper
• Deep learning methods for super-resolution reconstruction of turbulent flows

  2020influential reference
• Unsupervised deep learning for super-resolution reconstruction of turbulence

  2020cited by this paper
• A perspective on machine learning methods in turbulence modeling

  2020cited by this paper
• Bathymetric Inversion and Uncertainty Estimation from Synthetic Surf-Zone Imagery with Machine Learning

  2020cited by this paper
• Super-resolution and denoising of fluid flow using physics-informed convolutional neural networks without high-resolution labels

  2020influential reference
• B-PINNs: Bayesian Physics-Informed Neural Networks for Forward and Inverse PDE Problems with Noisy Data

  2020cited by this paper
• A Review of Uncertainty Quantification in Deep Learning: Techniques, Applications and Challenges

  2020influential reference
• Uncertainty Quantification for Sparse Deep Learning

  2020influential reference
• Application of Deep Learning-based Interpolation Methods to Nearshore Bathymetry

  2020cited by this paper
• Super-resolution reconstruction of turbulent velocity fields using a generative adversarial network-based artificial intelligence framework

  2019cited by this paper
• Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations

  2019cited by this paper
• Deep Learning for Image Super-Resolution: A Survey

  2019cited by this paper
• Fast Bayesian Uncertainty Estimation and Reduction of Batch Normalized Single Image Super-Resolution Network

  2019cited by this paper
• Ensemble Distribution Distillation

  2019cited by this paper
• Machine Learning Interpretability: A Survey on Methods and Metrics

  2019cited by this paper
• InterpretML: A Unified Framework for Machine Learning Interpretability

  2019cited by this paper
• Using Physics-Informed Super-Resolution Generative Adversarial Networks for Subgrid Modeling in Turbulent Reactive Flows

  2019cited by this paper
• Uncertainty in Neural Networks: Approximately Bayesian Ensembling

  2018cited by this paper
• Uncertainty Estimation via Stochastic Batch Normalization

  2018cited by this paper
• Turbulence Modeling in the Age of Data

  2018cited by this paper
• Neural Ordinary Differential Equations

  2018cited by this paper
• Bayesian Uncertainty Estimation for Batch Normalized Deep Networks

  2018cited by this paper
• Sub-grid scale model classification and blending through deep learning

  2018influential reference
• Super-resolution reconstruction of turbulent flows with machine learning

  2018cited by this paper
• Concrete Dropout

  2017cited by this paper
• Incorporating Nesterov Momentum into Adam

  2016cited by this paper
• Image-to-Image Translation with Conditional Adversarial Networks

  2016cited by this paper
• Risk versus Uncertainty in Deep Learning: Bayes, Bootstrap and the Dangers of Dropout

  2016cited by this paper
• Visualizing the effects of predictor variables in black box supervised learning models

  2016cited by this paper
• Weight Uncertainty in Neural Networks

  2015cited by this paper
• Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift

  2015cited by this paper
• U-Net: Convolutional Networks for Biomedical Image Segmentation

  2015cited by this paper
• Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning

  2015influential reference
• Dropout: a simple way to prevent neural networks from overfitting

  2014cited by this paper
• Peeking Inside the Black Box: Visualizing Statistical Learning With Plots of Individual Conditional Expectation

  2013cited by this paper
• Dropout Training as Adaptive Regularization

  2013cited by this paper
• ImageNet classification with deep convolutional neural networks

  2012cited by this paper
• Daily High-Resolution-Blended Analyses for Sea Surface Temperature

  2007cited by this paper
• An Improved In Situ and Satellite SST Analysis for Climate

  2002cited by this paper
• Thesis Supervisor Accepted by.......................................................................

  2001cited by this paper
• Flash Floods in Arid and Semi-Arid Zones

  1999cited by this paper
• Biomedical image segmentation

  1998cited by this paper

• High-resolution wide-field magnetic imaging with sparse sampling using nitrogen-vacancy centers

  2026cites this paper
• Enhancing indoor temperature mapping: High-resolution insights through deep learning and computational fluid dynamics

  2025cites this paper
• Super-resolution reconstruction of sequential images based on an active shift via a hybrid attention calibration mechanism

  2025cites this paper
• Stochastic near-surface wind field estimation from sparse aerial swarm measurements to support wildfire behaviour predictions

  2025cites this paper
• A deep learning super-resolution model for turbulent image upscaling and its application to shock wave–boundary layer interaction

  2024cites this paper
• Convolutional neural networks for compressible turbulent flow reconstruction

  2023cites this paper