AI Paper Atlas

Research map / Learning theory and algorithms / Combinatorial and search optimization

Neural Fields with Hard Constraints of Arbitrary Differential Order

Fangcheng Zhong, Kyle Thomas Fogarty, Param Hanji, Tianhao Walter Wu, Alejandro Sztrajman, Andrew Everett Spielberg, Andrea Tagliasacchi, Petra Bosilj, Cengiz Oztireli

NeurIPS 2023 · Learning theory and algorithms › Combinatorial and search optimization · approach: Differentiable and symbolic reasoning · 1 citations (OpenAlex)

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 changedWhereMeasuredEffect
use exact normal constraints instead of pseudo-normal constraints (ablation)CNF hypernet kernel · 2D circle normal interpolationmean normal error: 0.499 -> 1.083\times 10^{-6}Improved
use exact normal constraints instead of pseudo-normal constraints (ablation)CNF hypernet kernel · 2D triangle normal interpolationmean normal error: 0.238 -> 6.292\times 10^{-6}Improved
use exact normal constraints instead of SIRENCNF · 2D diamond normal interpolationmean normal error: 1.05 \times 10^{-1} -> 3.368\times 10^{-6}Improved
train after initialization with Eikonal lossCNF exact normal · 2D line normal interpolationmean normal error: 2.737\times 10^{-6} -> 6.963\times 10^{-6}Hurt
pre-train skewed RBF hyperparameters on mu=0 instead of random init (ablation)Skewed RBF · 1D advection, shift mu=10000RMSE: 635.78 -> 23.844Improved
pre-train skewed RBF hyperparameters instead of random init (ablation)Skewed RBF · 1D advection, no shiftRMSE: 0.0127 -> 0.0070Improved
pre-train skewed RBF hyperparameters instead of random init (ablation)Skewed RBF · 1D advection, shift mu=10nRMSE: 0.0594 -> 0.0018Improved
replace Gaussian kernel basis with dot-product kernel (ablation)CNF with SIREN encoder · BRDF fitting (MERL)reconstruction qualityHurt
replace Gaussian kernel basis with hypernet basis or constraint layer (ablation)CNF · BRDF fitting (MERL)reconstruction qualityHurt
use Gaussian kernel basis instead of other neural basis functions (ablation)CNF · 2D function with 4096 constraintscondition numberImproved
use skewed RBF with learnable per-basis covariance instead of fixed RBF (ablation)Kansa-style CNF · 1D advection on irregular griderrorImproved
pretrain encoder weights as circle SDFCNF hypernet kernel · 2D shape reconstructionplausibility of surfaceImproved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy