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>

Distributed Neural Representation for Reactive in situ Visualization

Interactive Volume Visualization via Multi-Resolution Hash Encoding based Neural Representation

HyperINR: A Fast and Predictive Hypernetwork for Implicit Neural Representations via Knowledge Distillation

UGINR: large-scale unstructured grid reduction via implicit neural representation

A Novel Implicit Neural Representation for Volume Data

Scalable Neural Indoor Scene Rendering.

Going Off-Grid: Continuous Implicit Neural Representations for 3D Vascular Modeling

Adaptively Placed Multi-Grid Scene Representation Networks for Large-Scale Data Visualization

A Compact Implicit Neural Representation for Efficient Storage of Massive 4D Functional Magnetic Resonance Imaging

Towards Scalable Neural Representation for Diverse Videos

Efficient Neural Representation of Volumetric Data using Coordinate-Based Networks

ECNR: Efficient Compressive Neural Representation of Time-Varying Volumetric Datasets

Fast Neural Representations for Direct Volume Rendering

Distributed Stochastic Optimization of a Neural Representation Network for Time-Space Tomography Reconstruction

STSR-INR: Spatiotemporal super-resolution for multivariate time-varying volumetric data via implicit neural representation

Implicit Neural Spatial Representations for Time-dependent PDEs

INV: Towards Streaming Incremental Neural Videos

Towards Full-scene Volumetric Video Streaming Via Spatially Layered Representation and NeRF Generation

NeuralVDB: High-resolution Sparse Volume Representation using Hierarchical Neural Networks

An In-Situ Visual Analytics Framework for Deep Neural Networks

Parallel Compositing of Volumetric Depth Images for Interactive Visualization of Distributed Volumes at High Frame Rates