Research map / Probabilistic and scientific modeling / Differential equations and PDEs
NOMAD: Nonlinear Manifold Decoders for Operator Learning
Jacob H Seidman, Georgios Kissas, Paris Perdikaris, George J. Pappas
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 |
|---|---|---|---|
| replace a linear decoder with a nonlinear manifold decoder | operator learning models · Antiderivative operator | relative L2 testing error: by one order of magnitude | Improved |
| use a nonlinear manifold decoder with latent dimension n=1 | NOMAD · Antiderivative operator | relative prediction error: 10\% | Improved |
| replace a linear decoder with a nonlinear manifold decoder | operator learning models · Parametric advection PDE | relative L2 testing error | Improved |
| replace the linear decoder with NOMAD | operator learning model · Shallow water equations | relative L2 error, rho (%): 0.040 \pm 0.015 -> 0.048 \pm 0.017 | Hurt |
| replace LOCA with NOMAD | operator learning model · Shallow water equations | relative L2 error, v1 (%): 2.7 \pm 0.3 -> 2.0 \pm 0.4 | Improved |
| replace LOCA with NOMAD | operator learning model · Shallow water equations | relative L2 error, v2 (%): 2.9 \pm 0.4 -> 2.6 \pm 0.3 | Improved |
| use NOMAD instead of LOCA | operator learning model · Shallow water equations | training cost (minutes): 12.1 -> 5.5 | Improved |
| use NOMAD instead of DeepONet | operator learning model · Shallow water equations | relative L2 error, rho (%): 0.100 \pm 0.030 -> 0.048 \pm 0.017 | Improved |
| use NOMAD instead of DeepONet | operator learning model · Shallow water equations | relative L2 error, v1 (%): 5.5 \pm 1.2 -> 2.0 \pm 0.4 | Improved |
| use NOMAD instead of DeepONet | operator learning model · Shallow water equations | relative L2 error, v2 (%): 5.9 \pm 1.4 -> 2.6 \pm 0.3 | Improved |
| use a nonlinear decoder with latent dimension 20 rather than a linear decoder with latent dimension 480 | operator learning model · Shallow water equations | latent dimension: 480 -> 20 | Improved |
Later papers that compared with this method
1 later accepted paper reports NoMAD as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Flow Matching for Few-Trial Neural Adaptation with Stable Latent Dynamics (ICML 2025 · better on 2 of 2 shared results)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Optimal deep learning of holomorphic operators between Banach spaces (NeurIPS 2024 · similarity 0.87)
- Learning semilinear neural operators: A unified recursive framework for prediction and data assimilation. (ICLR 2024 · similarity 0.87)
- Data-Efficient Operator Learning via Unsupervised Pretraining and In-Context Learning (NeurIPS 2024 · similarity 0.86)
- RIGNO: A Graph-based Framework For Robust And Accurate Operator Learning For PDEs On Arbitrary Domains (NeurIPS 2025 · similarity 0.86)
- Neural Operators with Localized Integral and Differential Kernels (ICML 2024 · similarity 0.86)
- Geometry Aware Operator Transformer as an efficient and accurate neural surrogate for PDEs on arbitrary domains (NeurIPS 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.
