Research map / Probabilistic and scientific modeling / Differential equations and PDEs
Improving physics-informed neural networks with meta-learned optimization
Alex Bihlo
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 Adam with meta-learned optimizer | PINN · 1D linear advection | training loss and point-wise error: more than 10 times smaller | Improved |
| replace Adam with meta-learned optimizer | PINN · 1D linear advection | epochs to reach loss: 500 epochs vs 3000 epochs | Improved |
| replace Adam with meta-learned optimizer | PINN · 2D Poisson equation | training loss and point-wise error | Improved |
| replace Adam with meta-learned optimizer | PINN · KdV equation | epochs to reach loss: 200 epochs vs 1000 epochs | Improved |
| meta-train optimizer on ensemble of initial conditions (transfer to unseen initial condition) | meta-learned optimizer on PINN · KdV equation, unseen initial condition k=2, phi=-pi/4 | training loss: 200 epochs vs 1000 epochs | Improved |
| replace Adam with meta-learned optimizer trained on Burgers equation | PINN · Burgers equation | training loss and error | Improved |
| use optimizer meta-trained on linear advection equation instead of Adam | PINN · Burgers equation (cross-equation transfer) | training loss | Mixed |
Closest papers by meaning
- ConFIG: Towards Conflict-free Training of Physics Informed Neural Networks (ICLR 2025 · similarity 0.86)
- Learning a Neural Solver for Parametric PDEs to Enhance Physics-Informed Methods (ICLR 2025 · similarity 0.86)
- Improving Energy Natural Gradient Descent through Woodbury, Momentum, and Randomization (NeurIPS 2025 · similarity 0.85)
- Challenges in Training PINNs: A Loss Landscape Perspective (ICML 2024 · similarity 0.84)
- Gradient Alignment in Physics-informed Neural Networks: A Second-Order Optimization Perspective (NeurIPS 2025 · similarity 0.84)
- Parameterized Physics-informed Neural Networks for Parameterized PDEs (ICML 2024 · similarity 0.84)
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.
