Research map / Probabilistic and scientific modeling / Differential equations and PDEs
Disentangled Representation Learning for Parametric Partial Differential Equations
Ning Liu, Lu Zhang, Tian Gao, Yue Yu
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 hyper-neural operator with disentangling VAE (DisentangO) | FNO · Exp 1 HGO, inverse (SC1) | test error: 14.55% -> 4.63% | Improved |
| add hypernetwork structure to MetaNO | MetaNO · Exp 1 HGO, forward | test error: 1.59% -> 1.65% | No clear effect |
| increase latent dimension from 2 to 15 (ablation) | DisentangO · Mechanical MNIST | test error: 11.49% -> 5.48% | Improved |
| increase latent dimension from 2 to 30 (ablation) | DisentangO · Unsupervised tissue | test error: 25.18% -> 5.28% | Improved |
| increase data loss weight beta_d from 1 to 100 (ablation) | DisentangO-15 · Mechanical MNIST | test error: 6.29% -> 5.48% | Improved |
| increase data loss weight beta_d beyond 100 (ablation) | DisentangO-15 · Mechanical MNIST | test error: 5.48% -> 5.80% (beta_d=1000) | No clear effect |
| increase beta_d (ablation) | DisentangO · Mechanical MNIST | mutual information between latents | Improved |
| increase classification loss weight (ablation) | DisentangO · Mechanical MNIST | mutual information between latents | Hurt |
| remove semi-supervision on latent factors (ablation) | DisentangO · Mechanical MNIST | forward test error: 12.60% at beta_d=1 | Mixed |
| use DisentangO instead of VAE | VAE · Mechanical MNIST | test error: 16.34% -> 5.48% | Improved |
| use UFNO backbone in DisentangO (ablation) | DisentangO · Exp 1 HGO | test error: 1.65% -> 6.61% | Hurt |
Closest papers by meaning
- Data-Efficient Operator Learning via Unsupervised Pretraining and In-Context Learning (NeurIPS 2024 · similarity 0.88)
- Smooth Piecewise Cutting for Neural Operator to Handle Discontinuities and Sharp Transitions (NeurIPS 2026 · similarity 0.88)
- Latent Neural Operator for Solving Forward and Inverse PDE Problems (NeurIPS 2024 · similarity 0.88)
- RIGNO: A Graph-based Framework For Robust And Accurate Operator Learning For PDEs On Arbitrary Domains (NeurIPS 2025 · similarity 0.88)
- HyPINO: Multi-Physics Neural Operators via HyperPINNs and the Method of Manufactured Solutions (NeurIPS 2025 · similarity 0.87)
- Using Uncertainty Quantification to Characterize and Improve Out-of-Domain Learning for PDEs (ICML 2024 · similarity 0.87)
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.
