AI Paper Atlas

Research map / Probabilistic and scientific modeling / Dynamical systems

GenCP: Towards Generative Modeling Paradigm of Coupled physics

Tianrun Gao, Haoren Zheng, Wenhao Deng, Haodong Feng, Tao Zhang, Ruiqi Feng, Qianyi Chen, Tailin Wu

ICLR 2026 · Probabilistic and scientific modeling › Dynamical systems · approach: Physics-informed dynamics learning · 1 citations (OpenAlex)

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 changedWhereMeasuredEffect
use GenCP operator-splitting coupled inference instead of M2PDEflow matching with FNO* backbone · Turek-Hronaverage relative L2 error: 26.77% lower average errorImproved
use GenCP instead of best baselineCNO backbone · Turek-Hronaverage relative L2 error: 12.54% lower average errorImproved
use GenCP instead of baselinesFNO* backbone · Double Cylinderaverage relative L2 error: 34.44% error reductionImproved
use GenCP instead of baselinesCNO backbone · Double Cylinderaverage relative L2 error: 42.85% error reductionImproved
use GenCP instead of M2PDE and SurrogateFNO* backbone · Nuclear-Thermal couplingaverage relative L2 error: 49.9% reduction vs M2PDE, 51.7% vs SurrogateImproved
use GenCP instead of M2PDE and SurrogateCNO backbone · Nuclear-Thermal couplingaverage relative L2 error: 58.2% over M2PDE and 78.8% over SurrogatesImproved
use GenCP instead of M2PDEflow matching · Synthetic complex distributionWasserstein-1 distance: 25450.5442 -> 0.4928Improved
use GenCP instead of M2PDEflow matching · Synthetic easy distributionWasserstein-1 distance: 0.5177 -> 0.4366Improved
use GenCP instead of M2PDE for inference costFNO* backbone · Turek-Hroninference time: 277.20s -> 19.50sImproved
use GenCP instead of Surrogate for inference costCNO backbone · Turek-Hroninference time: 300.25s -> 16.25sImproved
use GenCP decoupled training instead of Joint Training on coupled dataFNO* backbone · Double Cylinderrelative L2 error, u: 0.0110 -> 0.0522Hurt
use GenCP instead of Surrogate for SDF fieldFNO* backbone · Turek-Hron SDFrelative L2 error: 0.0112 -> 0.0081No clear effect

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy