Research map / Probabilistic and scientific modeling / Differential equations and PDEs
Meta-Auto-Decoder for Solving Parametric Partial Differential Equations
Xiang Huang, Zhanhong Ye, Hongsheng Liu, Shi Bei Ji, Zidong Wang, Kang Yang, Yang Li, Min Wang, Haotian CHU, Fan Yu, Bei Hua, Lei Chen and 1 more
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 |
|---|---|---|---|
| Use MAD-LM instead of training from scratch | Burgers' equation · Burgers' equation | convergence iterations: about 1200 iterations -> about 200 iterations | Improved |
| Use MAD-LM instead of Transfer-Learning | Burgers' equation · Burgers' equation | convergence iterations: about 1200 iterations -> about 200 iterations | Improved |
| Fine-tune model weights along with the latent vector rather than optimizing only the latent vector (ablation) | MAD · Burgers' equation | mean L2 error: 0.028 -> 0.013 | Improved |
| Fine-tune model weights along with the latent vector rather than optimizing only the latent vector (ablation) | MAD · Maxwell's equations | L2 error: 0.037 -> 0.030 | Improved |
| Fine-tune model weights along with the latent vector rather than optimizing only the latent vector (ablation) | MAD · Laplace's equation | L2 error: 0.0015 -> 0.0005 | Improved |
| Increase the number of pre-training samples from 10 to 50 (ablation) | MAD-L · Burgers' equation | mean L2 error: 0.365 -> 0.046 | Improved |
| Increase the number of pre-training samples from 10 to 50 (ablation) | MAD-LM · Burgers' equation | mean L2 error: 0.015 -> 0.013 | Improved |
| Increase the number of pre-training samples from 50 to 100 (ablation) | PI-DeepONet · Burgers' equation | mean L2 error: 0.247 -> 0.217 | Improved |
| Increase the number of pre-training samples from 200 to 300 (ablation) | PI-DeepONet · Burgers' equation | mean L2 error: 0.169 -> 0.181 | Hurt |
| Increase latent-vector size (ablation) | MAD-L · Burgers' equation | mean L2 error | Mixed |
| Increase latent-vector size (ablation) | MAD-LM · Burgers' equation | mean L2 error | No clear effect |
| Use MAD-LM instead of PI-DeepONet | Maxwell's equations · Maxwell's equations | mean L2 error: 0.672 -> 0.028 | Improved |
Closest papers by meaning
- Learning a Neural Solver for Parametric PDEs to Enhance Physics-Informed Methods (ICLR 2025 · similarity 0.88)
- Mechanistic PDE Networks for Discovery of Governing Equations (ICML 2025 · similarity 0.87)
- INC: An Indirect Neural Corrector for Auto-Regressive Hybrid PDE Solvers (NeurIPS 2025 · similarity 0.87)
- Interpolating Neural Network-Tensor Decomposition (INN-TD): a scalable and interpretable approach for large-scale physics-based problems (ICML 2025 · similarity 0.86)
- Derivative-enhanced Deep Operator Network (NeurIPS 2024 · similarity 0.86)
- Better Neural PDE Solvers Through Data-Free Mesh Movers (ICLR 2024 · 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.
