Research map / Probabilistic and scientific modeling / Differential equations and PDEs
Expanding the Chaos: Neural Operator for Stochastic (Partial) Differential Equations
Dai Shi, Lequan Lin, Andi Han, Luke Thompson, José Miguel Hernández-Lobato, Zhiyong Wang, Junbin Gao
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 FNO with F-SPDENO (add Wiener-chaos Wick noise features to FNO) | FNO · Stochastic Navier-Stokes 64x64 (W->X) | relative L2 error: 0.051 -> 0.037 | Improved |
| replace FNO with F-SPDENO (add Wiener-chaos Wick noise features to FNO) | FNO · Stochastic Navier-Stokes 64x64 ((X0,W)->X) | relative L2 error: 0.073 -> 0.031 | Improved |
| vary initial condition for deterministic FNO (W->X to (X0,W)->X) | FNO · Stochastic Navier-Stokes 64x64 | relative L2 error: 0.051 to 0.073 | Hurt |
| use F-SPDENO instead of NSPDE | NSPDE · Stochastic Navier-Stokes 64x64 (W->X) | relative L2 error: 0.040 to 0.037 | Improved |
| use F-SPDENO instead of NSPDE | NSPDE · Stochastic Navier-Stokes 64x64 ((X0,W)->X) | relative L2 error: 0.047 to 0.031 | Improved |
| increase training trajectories N from 1000 to 10000 | F-SPDENO · Dynamic Phi41 (W->X) | relative L2 error: 0.012 -> 0.011 | No clear effect |
| increase training trajectories N from 1000 to 10000 | NSPDE · Dynamic Phi41 (W->X) | relative L2 error: 0.021 -> 0.013 | Improved |
| increase training trajectories N from 1000 to 10000 | NCDE-FNO · Dynamic Phi41 (W->X) | relative L2 error: 0.077 -> 0.091 | Hurt |
| use U-SDENO one-step sampler instead of pretrained multi-step DDPM | DDPM CIFAR-10 · CIFAR-10 one-step sampling | FID-50K: 3.17 -> 4.89 | Hurt |
| use Manifold-SDENO instead of RSGM | RSGM · Flood dataset on S2 | NLL: 0.49 +- 0.20 -> 0.51 +- 0.12 | No clear effect |
| use G-SDENO instead of TSBM | TSBM · Brain signal interpolation | 1-Wasserstein distance: 9.51 -> 9.60 | No clear effect |
Closest papers by meaning
- Neural Stochastic Flows: Solver-Free Modelling and Inference for SDE Solutions (NeurIPS 2025 · similarity 0.88)
- Stable Neural Stochastic Differential Equations in Analyzing Irregular Time Series Data (ICLR 2024 · similarity 0.87)
- In-Context Learning of Stochastic Differential Equations with Foundation Inference Models (NeurIPS 2025 · similarity 0.87)
- An Efficient High-dimensional Gradient Estimator for Stochastic Differential Equations (NeurIPS 2024 · similarity 0.87)
- Efficient Training of Neural Stochastic Differential Equations by Matching Finite Dimensional Distributions (ICLR 2025 · similarity 0.87)
- SPDEBench: An Extensive Benchmark for Learning Stochastic PDEs (NeurIPS 2026 · similarity 0.87)
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.
