Research map / Probabilistic and scientific modeling / Differential equations and PDEs
Enhancing Stability of Physics-Informed Neural Network Training Through Saddle-Point Reformulation
Dmitry Bylinkin, Mikhail Aleksandrov, Savelii Chezhegov, Aleksandr Beznosikov
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 |
|---|---|---|---|
| use Adam for descent and RMSProp for ascent instead of Adam+Adam (ablation) | AdaptiveBGDA · Poisson 2d-C | L2RE: 4.45E-2 -> 8.16E-3 | Improved |
| use Adam for descent and RMSProp for ascent instead of RMSProp+RMSProp (ablation) | AdaptiveBGDA · Poisson 2d-C | L2RE: 6.02E-1 -> 8.16E-3 | Improved |
| replace best previous optimizer with saddle-point AdaptiveBGDA | vanilla PINN · PINNacle benchmark PDEs | L2RE: dominant in 72.7% cases; previous record 22.7% (LRA) | Improved |
| replace best previous optimizer with AdaptiveBGDA | vanilla PINN · PINNacle benchmark PDEs | L2RE: quality improved by more than double in 18.2% cases | Improved |
| apply AdaptiveBGDA to Burgers 2d-C | vanilla PINN · Burgers 2d-C | L2RE: 2.60E-1 -> 4.21E-1 | Hurt |
| apply AdaptiveBGDA to gPINN architecture | gPINN · PINNacle benchmark PDEs | L2RE: superiority in 81.8% of cases | Improved |
| apply AdaptiveBGDA to LAAF architecture | LAAF · PINNacle benchmark PDEs | L2RE: superiority in 72.7% of cases | Improved |
| apply AdaptiveBGDA to GAAF architecture | GAAF · PINNacle benchmark PDEs | L2RE: superiority in 90.1% of cases | Improved |
| apply AdaptiveBGDA to Burgers 2d-C with modern architectures | gPINN · Burgers 2d-C | L2RE: 3.27E-1 -> 5.11E-1 | Hurt |
Closest papers by meaning
- ConFIG: Towards Conflict-free Training of Physics Informed Neural Networks (ICLR 2025 · similarity 0.89)
- PINNsFormer: A Transformer-Based Framework For Physics-Informed Neural Networks (ICLR 2024 · similarity 0.89)
- Sub-Sequential Physics-Informed Learning with State Space Model (ICML 2025 · similarity 0.89)
- Challenges in Training PINNs: A Loss Landscape Perspective (ICML 2024 · similarity 0.88)
- Per-Loss Adapters for Gradient Conflict in Physics-Informed Neural Networks (NeurIPS 2026 · similarity 0.88)
- Efficient Error Certification for Physics-Informed Neural Networks (ICML 2024 · similarity 0.88)
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.
