Abstract:Implicit Neural Spatial Representation (INSR) has emerged as an effective representation of spatially-dependent vector fields. This work explores solving time-dependent PDEs with INSR. Classical PDE solvers introduce both temporal and spatial discretizations. Common spatial discretizations include meshes and meshless point clouds, where each degree-of-freedom corresponds to a location in space. While these explicit spatial correspondences are intuitive to model and understand, these representations are not necessarily optimal for accuracy, memory usage, or adaptivity. Keeping the classical temporal discretization unchanged (e.g., explicit/implicit Euler), we explore INSR as an alternative spatial discretization, where spatial information is implicitly stored in the neural network weights. The network weights then evolve over time via time integration. Our approach does not require any training data generated by existing solvers because our approach is the solver itself. We validate our approach on various PDEs with examples involving large elastic deformations, turbulent fluids, and multi-scale phenomena. While slower to compute than traditional representations, our approach exhibits higher accuracy and lower memory consumption. Whereas classical solvers can dynamically adapt their spatial representation only by resorting to complex remeshing algorithms, our INSR approach is intrinsically adaptive. By tapping into the rich literature of classic time integrators, e.g., operator-splitting schemes, our method enables challenging simulations in contact mechanics and turbulent flows where previous neural-physics approaches struggle. Videos and codes are available on the project page: <a class="link-external link-http" href="http://www.cs.columbia.edu/cg/INSR-PDE/" rel="external noopener nofollow">this http URL</a>

Implicit Approximation of Neural Network and Applications

Implicit interpolation of time vector series.

Processing of well log data based on backpropagation neural network implicit approximation

Reservoir Static Property Estimation Using Nearest-Neighbor Neural Network

Artificial Intelligence Methods for Oil and Gas Reservoir Development: Current Progresses and Perspectives

3D reservoir modelling with the aid of artificial neural networks

A Process Neural Network for Continuous Process Approximation and Its Application

On the Approximation and Complexity of Deep Neural Networks to Invariant Functions

Implicit Neural Spatial Representations for Time-dependent PDEs

Implicit curves and surfaces based on BP neural network

Integral transform and its application to neural network approximation

Artificial neural network–based time-domain interwell tracer testing for ultralow-permeability fractured reservoirs

Neural Network Aided Approximation and Parameter Inference of Non-Markovian Models of Gene Expression

A Physics-Informed Spatial-Temporal Neural Network for Reservoir Simulation and Uncertainty Quantification

Parameterizing uncertainty by deep invertible networks, an application to reservoir characterization

Physics-Informed Neural Network for the Transient Diffusivity Equation in Reservoir Engineering

Neural Likelihood Approximation for Integer Valued Time Series Data

The Neural Network Approach to Inverse Problems in Differential Equations

Automatic Well Test Interpretation Method for Circular Reservoirs with Changing Wellbore Storage Using One-Dimensional Convolutional Neural Network

Neural Network Approximations of Compositional Functions With Applications to Dynamical Systems

Simulating the Behavior of Reservoirs with Convolutional and Recurrent Neural Networks