Research map / Probabilistic and scientific modeling / Differential equations and PDEs
Unbiased and Second-Order-Free Training for High-Dimensional PDEs
Jaemin Seo, Su Rin Lee, Jae Yong Lee
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 independent-group product debiasing (Un-EM) to multi-shot EM-BSDE (ablation) | Multi-Shot EM-BSDE · HJB (soft constraint) | relative L2 error (x1e-2): 0.1617 -> 0.1348 | Improved |
| add independent-group product debiasing (Un-EM) to multi-shot EM-BSDE (ablation) | Multi-Shot EM-BSDE · BSB (soft constraint) | relative L2 error (x1e-2): 0.1046 -> 0.0814 | Improved |
| add independent-group product debiasing (Un-EM) to multi-shot EM-BSDE (ablation) | Multi-Shot EM-BSDE · AC (soft constraint) | relative L2 error (x1e-2): 0.0206 -> 0.0147 | Improved |
| add independent-group product debiasing (Un-EM) to multi-shot EM-BSDE (ablation) | Multi-Shot EM-BSDE · BZ (soft constraint) | relative L2 error (x1e-2): 5.3379 -> 5.1778 | Improved |
| add independent-group product debiasing (Un-EM) to multi-shot EM-BSDE (ablation) | Multi-Shot EM-BSDE · PIDE (soft constraint) | relative L2 error (x1e-2): 1.0771 -> 1.0878 | Hurt |
| add independent-group product debiasing (Un-EM) to multi-shot EM-BSDE (ablation) | Multi-Shot EM-BSDE · HJB (hard constraint) | relative L2 error (x1e-2): 0.1172 -> 0.1444 | Hurt |
| add independent-group product debiasing (Un-EM) to multi-shot EM-BSDE (ablation) | Multi-Shot EM-BSDE · BSB (hard constraint) | relative L2 error (x1e-2): 0.0739 -> 0.0120 | Improved |
| add independent-group product debiasing (Un-EM) to multi-shot EM-BSDE (ablation) | Multi-Shot EM-BSDE · AC (hard constraint) | relative L2 error (x1e-2): 0.0113 -> 0.0034 | Improved |
| add independent-group product debiasing (Un-EM) to multi-shot EM-BSDE (ablation) | Multi-Shot EM-BSDE · BZ (hard constraint) | relative L2 error (x1e-2): 1.1952 -> 0.2279 | Improved |
| add independent-group product debiasing (Un-EM) to multi-shot EM-BSDE (ablation) | Multi-Shot EM-BSDE · PIDE (hard constraint) | relative L2 error (x1e-2): 0.0245 -> 0.0226 | Improved |
| apply debiasing wrapper to Shotgun | Shotgun · BSB (hard constraint) | RL2 error: reduces RL2 error by approximately 2.67x, training time 1.78x | Improved |
| use Un-EM-BSDE instead of Heun-BSDE for unbiased training | Un-EM-BSDE · training time relative to EM-BSDE | training time: 42.91x -> 1.79x | Improved |
Closest papers by meaning
- Integration Matters for Learning PDEs with Backwards SDEs (NeurIPS 2025 · similarity 0.93)
- Efficient Training of Neural Fractional-Order Differential Equation via Adjoint Backpropagation (AAAI 2025 · similarity 0.86)
- Efficient Training of Neural Stochastic Differential Equations by Matching Finite Dimensional Distributions (ICLR 2025 · similarity 0.86)
- MC^2: Monte Carlo Correction for Fast Elliptic PDE Solving (NeurIPS 2026 · similarity 0.86)
- S$^{2}$-PINN: Stochastic Separable Physics-Informed Neural Networks (NeurIPS 2026 · similarity 0.85)
- Learning a Neural Solver for Parametric PDEs to Enhance Physics-Informed Methods (ICLR 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.
