Research map / Probabilistic and scientific modeling / Molecules and proteins
Noether Networks: meta-learning useful conserved quantities
Ferran Alet, Dylan Doblar, Allan Zhou, Joshua B. Tenenbaum, Kenji Kawaguchi, Chelsea Finn
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 |
|---|---|---|---|
| Meta-tailor the predictor with a synthesized conservation formula | Vanilla MLP · Ideal pendulum | RMSE: 0.0563 -> 0.0423 | Improved |
| Meta-tailor the predictor with a synthesized conservation formula | Vanilla MLP · Ideal spring | RMSE: .0174 -> .0165 | Improved |
| Use an automatically discovered conservation formula instead of the true formula | Noether Nets · Ideal pendulum | RMSE: 0.0423 vs 0.0422 | No clear effect |
| Use an automatically discovered conservation formula instead of the true formula | Noether Nets · Ideal spring | RMSE: .0165 vs .0166 | No clear effect |
| Add prediction-time tailoring with a learned conserved quantity | SVG · Controlled pixel pendulum | MSE | Improved |
| Add prediction-time tailoring with a learned conserved quantity | SVG · Physics 101 ramp, 20-degree incline | LPIPS | Improved |
| Add prediction-time tailoring with a learned conserved quantity | SVG · Physics 101 ramp, 20-degree incline | MSE | Improved |
| Add prediction-time tailoring with a learned conserved quantity | SVG · Physics 101 ramp, 20-degree incline | PSNR | Improved |
| Add prediction-time tailoring with a learned conserved quantity | SVG · Physics 101 ramp, 20-degree incline | SSIM | Improved |
| Continue training SVG for 20 additional epochs instead of adding Noether tailoring | SVG · Controlled pixel pendulum | MSE | No clear effect |
| Increase the number of inner conservation-optimization steps | Noether Networks · Physics 101 ramp, 20-degree incline | Task loss | Improved |
Closest papers by meaning
- Noether's Razor: Learning Conserved Quantities (NeurIPS 2024 · similarity 0.85)
- Position: The No Free Lunch Theorem, Kolmogorov Complexity, and the Role of Inductive Biases in Machine Learning (ICML 2024 · similarity 0.85)
- Non-asymptotic Convergence of Training Transformers for Next-token Prediction (NeurIPS 2024 · similarity 0.84)
- Investigating the Overlooked Hessian Structure: From CNNs to LLMs (ICML 2025 · similarity 0.83)
- Explainable Neural Networks with Guarantee: A Sparse Estimation Approach (AAAI 2025 · similarity 0.83)
- Simplicity Bias and Optimization Threshold in Two-Layer ReLU Networks (ICML 2025 · similarity 0.83)
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.
