GenCP: Towards Generative Modeling Paradigm of Coupled Physics

Real-world physical systems are inherently complex, often involving the coupling of multiple physics, making their simulation both highly valuable and challenging. Many mainstream approaches face challenges when dealing with decoupled data. Besides, they also suffer from low efficiency and fidelity in strongly coupled spatio-temporal physical systems. Here we propose GenCP, a novel and elegant generative paradigm for coupled multiphysics simulation. By formulating coupled-physics modeling as a probability modeling problem, our key innovation is to integrate probability density evolution in generative modeling with iterative multiphysics coupling, thereby enabling training on data from decoupled simulation and inferring coupled physics during sampling. We also utilize operator-splitting theory in the space of probability evolution to establish error controllability guarantees for this “conditional-to-joint” sampling scheme. We evaluate our paradigm on a synthetic setting and three challenging multi-physics scenarios to demonstrate both principled insight and superior application performance of GenCP. Code is available at this repo: github.com/AI4Science-WestlakeU/GenCP.

💡 Research Summary

GenCP (Generative Coupled Physics) introduces a principled framework for multiphysics simulation that learns from decoupled data yet performs coupled inference during sampling. The authors observe that most data‑driven surrogate or neural‑operator approaches require coupled solutions for training, which are costly to obtain, especially for high‑dimensional, stochastic, or strongly interacting systems. To overcome this, GenCP reframes coupled‑physics modeling as a probability‑density evolution problem in the functional space of the physical fields.

The method proceeds in two stages. First, conditional flow models are trained separately for each field (e.g., f and g) using only decoupled solver outputs. For the f‑field, data pairs (f₁, \bar{g}) are drawn from a dataset where g is held fixed while f evolves; similarly for g. A reference Gaussian distribution provides latent “noise” samples (z_f, z_g). Linear interpolation between the reference and the target field yields a trajectory (f_t, g_t) for t∈