Research map / Architectures and efficiency / RNNs and normalization
Beyond Uniformity: Regularizing Implicit Neural Representations through a Lipschitz Lens
Julian McGinnis, Suprosanna Shit, Florian A. Hölzl, Paul Friedrich, Paul Büschl, Vasiliki Sideri-Lampretsa, Mark Mühlau, Philippe C. Cattin, Bjoern Menze, Daniel Rueckert, Benedikt Wiestler
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 |
|---|---|---|---|
| add Lipschitz budget (K=2) constraint to SIREN | SIREN · Lung registration case 9 | TRE: 21.69 ± 15.52 -> 1.30 ± 1.01 | Improved |
| add Lipschitz budget (K=2) constraint to SIREN | SIREN · Lung registration case 7 | TRE: 9.39 ± 10.48 -> 1.41 ± 1.09 | Improved |
| add Lipschitz budget constraint to SIREN | SIREN · Lung registration case 8 | TRE: 1.22 ± 1.26 -> 1.52 ± 1.27 | Hurt |
| use Lipschitz SIREN instead of Jacobian-regularized SIREN | SIREN · Lung registration (10 cases) | folding ratio: better folding ratio in 8 out of 10 cases | Improved |
| use Lipschitz SIREN instead of NODEO | NODEO · Lung registration (10 cases) | TRE: 4.72 ± 5.14 -> 1.42 ± 1.33 (case 4) | Improved |
| use non-uniform (e.g. exponential) budget allocation instead of uniform (ablation) | Lipschitz INRs · Lung registration | TRE | Improved |
| use non-uniform budget allocation instead of uniform (ablation) | Lipschitz INRs · CelebA inpainting | PSNR | Improved |
| use non-uniform budget allocation instead of uniform 1-Lipschitz (ablation) | 1-Lipschitz SDFs · Stanford bunny SDFs | Chamfer Distance | No clear effect |
| use Bjorck or SLL normalization instead of spectral normalization (ablation) | 1-Lipschitz SDF · Stanford bunny | Chamfer Distance | Improved |
| use MaxMin/Householder gradient-preserving activations instead of ReLU (ablation) | 1-Lipschitz SDF · Stanford bunny | Chamfer Distance | Improved |
| move Lipschitz budget away from oracle estimate (ablation) | Lipschitz FFN · CelebA inpainting | PSNR | Hurt |
| scale SIREN first and hidden layer weights by alpha (ablation) | SIREN · inpainting (25% pixels) | Lipschitz bound | Improved |
Closest papers by meaning
- Some Fundamental Aspects about Lipschitz Continuity of Neural Networks (ICLR 2024 · similarity 0.88)
- ECLipsE: Efficient Compositional Lipschitz Constant Estimation for Deep Neural Networks (NeurIPS 2024 · similarity 0.87)
- Grids Often Outperform Implicit Neural Representation at Compressing Dense Signals (NeurIPS 2025 · similarity 0.86)
- A Rescaling-Invariant Lipschitz Bound Based on Path-Metrics for Modern ReLU Network Parameterizations (ICML 2025 · similarity 0.86)
- Approximation theory for 1-Lipschitz ResNets (NeurIPS 2025 · similarity 0.86)
- VI^3NR: Variance Informed Initialization for Implicit Neural Representations (CVPR 2025 · similarity 0.86)
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.
