Research map / Probabilistic and scientific modeling / Differential equations and PDEs
Hypernetwork-based Meta-Learning for Low-Rank Physics-Informed Neural Networks
Woojin Cho, Kookjin Lee, Donsub Rim, Noseong Park
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 hypernetwork-based meta-learning instead of MAML | Naive-LR-PINN · Convection beta=30 | relative error: 0.6294 -> 0.0379 | Improved |
| use hypernetwork-based meta-learning instead of Reptile | Naive-LR-PINN · Convection beta=40 | relative error: 0.5632 -> 0.0722 | Improved |
| add curriculum learning pre-training | Naive-LR-PINN · Convection beta=35 | relative error: 0.5357 -> 0.5465 | No clear effect |
| use adaptive rank instead of full rank | Hyper-LR-PINN · Convection beta=30 | relative error: 0.0379 -> 0.0389 | No clear effect |
| use adaptive rank instead of full rank | Hyper-LR-PINN · Convection beta=40 | relative error: 0.0655 -> 0.0722 | Hurt |
| replace vanilla PINN with Hyper-LR-PINN | PINN · Convection, general CDR | relative error: 0.0327 -> 0.0085 | Improved |
| replace MAML with Hyper-LR-PINN | MAML · Conv-Diff-Reac | relative error: 0.0144 -> 0.0329 | Hurt |
| replace MAML with Hyper-LR-PINN | MAML · Reaction | relative error: 0.0057 -> 0.0045 | Improved |
| replace PINN with low-rank Hyper-LR-PINN (parameter count) | PINN · model size | # trainable parameters: 10,401 -> 351 | Improved |
| low-rank format with hypernetwork vs vanilla PINN | PINN · training time | training time: more than 4x speed up | Improved |
| use fixed instead of trainable basis vectors (ablation) | Hyper-LR-PINN · Convection | accuracy | Improved |
| add orthogonality constraint on basis vectors (ablation) | Hyper-LR-PINN · Convection | accuracy | Improved |
Closest papers by meaning
- PINNsFormer: A Transformer-Based Framework For Physics-Informed Neural Networks (ICLR 2024 · similarity 0.91)
- Sub-Sequential Physics-Informed Learning with State Space Model (ICML 2025 · similarity 0.91)
- RoPINN: Region Optimized Physics-Informed Neural Networks (NeurIPS 2024 · similarity 0.90)
- Parameterized Physics-informed Neural Networks for Parameterized PDEs (ICML 2024 · similarity 0.89)
- PINNsAgent: Automated PDE Surrogation with Large Language Models (ICML 2025 · similarity 0.89)
- Challenges in Training PINNs: A Loss Landscape Perspective (ICML 2024 · similarity 0.89)
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.
