Research map / Multi-agent systems and games / Human-AI and goal-conditioned
CINOC: Cardinality-Invariant Neural Operator Policies for Scalable PDE Control
Pietro Zanotta, Dibakar Roy Sarkar, Honghui Zheng, Somdatta Goswami, Jan Drgona
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 differentiable-physics training with a neural operator policy (CINOC) instead of best RL | CINOC · KS 2D | energy: 0.50 ± 0.24 -> 0.37 ± 0.19 | Improved |
| replace DPC with CINOC | CINOC · KS 2D | energy: 9.74 ± 2.32 -> 0.37 ± 0.19 | Improved |
| use CINOC instead of DPC in obstacle-constrained Heat 2D | CINOC · Heat 2D (Obs) | MSE: 2.7e-4 -> 1.2e-4 | Improved |
| use CINOC instead of best RL (MAPPO) on Heat 2D (Obs) | CINOC · Heat 2D (Obs) | MSE: 0.004 -> 1.2e-4 | Improved |
| use CINOC instead of DPC on Heat 1D | CINOC · Heat 1D | MSE: 7.9e-5 -> 2.9e-4 | Hurt |
| use CINOC instead of RL baselines on Density Control | CINOC · Density Control | MML: 0.002 vs 0.002 (DPC) | No clear effect |
| use CINOC instead of PPO on Density Control | CINOC · Density Control | MML: 0.005 -> 0.002 | Improved |
| apply control to KS 2D versus uncontrolled evolution | CINOC · KS 2D | energy: 12.16 -> 0.20 | Improved |
| use CINOC instead of MAPPO on KS 1D | CINOC · KS 1D | energy: 0.13 ± 0.13 vs 0.13 ± 0.12 | No clear effect |
| use CINOC instead of DPC on Turbulence 2D | CINOC · Turbulence 2D | enstrophy: 3.28 ± 2.19 -> 3.23 ± 2.22 | Improved |
Closest papers by meaning
- EvoControl: Multi-Frequency Bi-Level Control for High-Frequency Continuous Control (ICML 2025 · similarity 0.85)
- Stable Long-Horizon PDE Forecasting via Latent Structured Spectral Propagators (NeurIPS 2026 · similarity 0.85)
- RIGNO: A Graph-based Framework For Robust And Accurate Operator Learning For PDEs On Arbitrary Domains (NeurIPS 2025 · similarity 0.85)
- CViT: Continuous Vision Transformer for Operator Learning (ICLR 2025 · similarity 0.85)
- Learning semilinear neural operators: A unified recursive framework for prediction and data assimilation. (ICLR 2024 · similarity 0.84)
- Adaptive Coordinate Transforms for Neural Operators (NeurIPS 2026 · 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.
