Hybrid Physical-Neural ODEs for Fast N-body Simulations
This work addresses the need for more accurate and adaptable fast simulations in cosmology, though it is incremental as it builds on prior schemes like PGD.
The authors tackled the problem of small-scale inaccuracies in fast cosmological N-body simulations by introducing a hybrid physical-neural ODE scheme that adds a neural network parameterized effective force, resulting in improved cross-correlation coefficients and robustness compared to existing methods.
We present a new scheme to compensate for the small-scales approximations resulting from Particle-Mesh (PM) schemes for cosmological N-body simulations. This kind of simulations are fast and low computational cost realizations of the large scale structures, but lack resolution on small scales. To improve their accuracy, we introduce an additional effective force within the differential equations of the simulation, parameterized by a Fourier-space Neural Network acting on the PM-estimated gravitational potential. We compare the results for the matter power spectrum obtained to the ones obtained by the PGD scheme (Potential gradient descent scheme). We notice a similar improvement in term of power spectrum, but we find that our approach outperforms PGD for the cross-correlation coefficients, and is more robust to changes in simulation settings (different resolutions, different cosmologies).