A Tailored Convolutional Neural Network for Nonlinear Manifold Learning of Computational Physics Data using Unstructured Spatial Discretizations
arXiv:2006.06154 · doi:10.1137/20M1344263
Abstract
We propose a nonlinear manifold learning technique based on deep convolutional autoencoders that is appropriate for model order reduction of physical systems in complex geometries. Convolutional neural networks have proven to be highly advantageous for compressing data arising from systems demonstrating a slow-decaying Kolmogorov n-width. However, these networks are restricted to data on structured meshes. Unstructured meshes are often required for performing analyses of real systems with complex geometry. Our custom graph convolution operators based on the available differential operators for a given spatial discretization effectively extend the application space of deep convolutional autoencoders to systems with arbitrarily complex geometry that are typically discretized using unstructured meshes. We propose sets of convolution operators based on the spatial derivative operators for the underlying spatial discretization, making the method particularly well suited to data arising from the solution of partial differential equations. We demonstrate the method using examples from heat transfer and fluid mechanics and show better than an order of magnitude improvement in accuracy over linear methods.
arXiv commentPreprint
References in corpus (41)
- Deep Learning in Neural Networks: An Overview
- The Emerging Field of Signal Processing on Graphs: Extending High-Dimensional Data Analysis to Networks and Other Irregular Domains
- Neural Message Passing for Quantum Chemistry
- Crystal Graph Convolutional Neural Networks for an Accurate and Interpretable Prediction of Material Properties
- Spectral Networks and Locally Connected Networks on Graphs
- A Data-Driven Approximation of the Koopman Operator: Extending Dynamic Mode Decomposition
- Molecular Graph Convolutions: Moving Beyond Fingerprints
- Graph Networks as a Universal Machine Learning Framework for Molecules and Crystals
- Modeling polypharmacy side effects with graph convolutional networks
- Deep learning for universal linear embeddings of nonlinear dynamics
- Graph Neural Networks: A Review of Methods and Applications
- Deep Convolutional Networks on Graph-Structured Data
- Physics-Constrained Deep Learning for High-dimensional Surrogate Modeling and Uncertainty Quantification without Labeled Data
- Convolutional Neural Network Architectures for Matching Natural Language Sentences
- The GNAT method for nonlinear model reduction: effective implementation and application to computational fluid dynamics and turbulent flows
- Physics-Informed Multi-LSTM Networks for Metamodeling of Nonlinear Structures
- ParticleNet: Jet Tagging via Particle Clouds
- Nonlinear mode decomposition with convolutional neural networks for fluid dynamics
- Reduced-order modeling of advection-dominated systems with recurrent neural networks and convolutional autoencoders
- Data-assisted reduced-order modeling of extreme events in complex dynamical systems
- Convolutional neural network and long short-term memory based reduced order surrogate for minimal turbulent channel flow
- Machine-learning-based reduced order modeling for unsteady flows around bluff bodies of various shapes
- A deep learning enabler for non-intrusive reduced order modeling of fluid flows
- Orbital Graph Convolutional Neural Network for Material Property Prediction
- Learning representations of irregular particle-detector geometry with distance-weighted graph networks
- A Deep Learning based Approach to Reduced Order Modeling for Turbulent Flow Control using LSTM Neural Networks
- Time-series learning of latent-space dynamics for reduced-order model closure
- Deep Learning the Physics of Transport Phenomena
- SocialGCN: An Efficient Graph Convolutional Network based Model for Social Recommendation
- Knowledge Transfer for Out-of-Knowledge-Base Entities: A Graph Neural Network Approach
- Deep convolutional recurrent autoencoders for learning low-dimensional feature dynamics of fluid systems
- Recovering missing CFD data for high-order discretizations using deep neural networks and dynamics learning
- Partial Convolution based Padding
- Data-Driven, Physics-Based Feature Extraction from Fluid Flow Fields
- Hierarchical Visualization of Materials Space with Graph Convolutional Neural Networks
- Memory embedded non-intrusive reduced order modeling of non-ergodic flows
- Embedding Hard Physical Constraints in Neural Network Coarse-Graining of 3D Turbulence
- DCFNet: Deep Neural Network with Decomposed Convolutional Filters
- Graph Convolutional Neural Networks for Polymers Property Prediction
- Learning on the Edge: Explicit Boundary Handling in CNNs
- Empirical Study of the Benefits of Overparameterization in Learning Latent Variable Models