Research map / Probabilistic and scientific modeling / Differential equations and PDEs
$\boldsymbol{\partial^\infty}$-Grid: A Neural Differential Equation Solver with Differentiable Feature Grids
Navami Kairanda, Shanthika Naik, Marc Habermann, Avinash Sharma, Christian Theobalt, Vladislav Golyanik
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use differentiable RBF feature grid instead of linear-interpolated K-Planes | feature grid · gradient-supervised image reconstruction | PSNR: 17.96 -> 32.24 | Improved |
| use feature grid instead of Siren MLP | ∂∞-Grid · gradient-supervised image reconstruction | training time: 10mins -> 25s (PSNR 32.12 -> 32.24) | Improved |
| use feature grid instead of Siren MLP | ∂∞-Grid · Laplacian-supervised image reconstruction | training time: 1hr56mins -> 15mins (PSNR 11.82 -> 12.19) | Improved |
| use feature grid instead of Siren MLP | ∂∞-Grid · Helmholtz omega=20 | training time: 4x speed-up | Improved |
| use eps=2 with 1-ring instead of eps=1 with larger rings (ablation) | RBF interpolation · gradient image reconstruction 768x768 | PSNR: 28.74 -> 17.72 | Hurt |
| use 1-ring instead of 2-ring neighbourhood at eps=1 (ablation) | RBF interpolation · gradient image reconstruction 768x768 | PSNR: 28.82 -> 27.50 | Hurt |
| increase ring size from 2 to 4 at eps=1 (ablation) | RBF interpolation · gradient image reconstruction 768x768 | PSNR: 28.82 -> 28.74 | No clear effect |
| add multi-resolution feature grids (ablation) | ∂∞-Grid · Eikonal SDF and image recovery | convergence | Improved |
| use NeuRBF with differential supervision | NeuRBF · image fitting with Poisson loss | PSNR: 10.85 (failed to converge) | Hurt |
Closest papers by meaning
- Accurate Differential Operators for Hybrid Neural Fields (CVPR 2025 · similarity 0.87)
- UGrid: An Efficient-And-Rigorous Neural Multigrid Solver for Linear PDEs (ICML 2024 · similarity 0.87)
- Vectorized Conditional Neural Fields: A Framework for Solving Time-dependent Parametric Partial Differential Equations (ICML 2024 · similarity 0.86)
- Solving Differential Equations with Constrained Learning (ICLR 2025 · similarity 0.86)
- CALM-PDE: Continuous and Adaptive Convolutions for Latent Space Modeling of Time-dependent PDEs (NeurIPS 2025 · similarity 0.85)
- Toward Efficient Kernel-Based Solvers for Nonlinear PDEs (ICML 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
