MF-Box: Multi-fidelity and multi-scale emulation for the matter power spectrum

Abstract

We introduce MF-Box, an extended version of MFEmulator, designed as a fast surrogate for power spectra, trained using N-body simulation suites from various box sizes and particle loads. To demonstrate MF-Box's effectiveness, we design simulation suites that include low-fidelity suites (L1 and L2) at $256\mathrm{Mpc}/h$ and $100\mathrm{Mpc}/h$, each with $128^3$ particles, and a high-fidelity suite (HF) with $512^3$ particles at $256\mathrm{Mpc}/h$, representing a higher particle load compared to the low-fidelity suites. MF-Box acts as a probabilistic resolution correction function, learning most of the cosmological dependencies from L1 and L2 simulations and rectifying resolution differences with just 3 HF simulations using a Gaussian process. MF-Box successfully emulates power spectra from our HF testing set with a relative error of $<3\%$ up to $k\simeq 7h/\mathrm{Mpc}$ at $z\in [0,3]$, while maintaining a cost similar to our previous multi-fidelity approach, which was accurate only up to $z=1$. The addition of an extra low-fidelity node in a smaller box significantly improves emulation accuracy for MF-Box at $k>2h/\mathrm{Mpc}$, increasing it by a factor of $10$. We conduct an error analysis of MF-Box based on computational budget, providing guidance for optimizing budget allocation per fidelity node. Our proposed MF-Box enables future surveys to efficiently combine simulation suites of varying quality, effectively expanding the range of emulation capabilities while ensuring cost efficiency.

19 pages, 16 figures. Match the MNRAS version. Data available at: https://github.com/jibanCat/MFBoxData, Code and tutorials available at: https://github.com/jibanCat/matter_emu_mfbox

References in corpus (10)

• Array Programming with NumPy
• TensorFlow Distributions
• The Mira-Titan Universe II: Matter Power Spectrum Emulation
• Fast cosmological parameter estimation using neural networks
• {\sc CosmoNet}: fast cosmological parameter estimation in non-flat models using neural networks
• SP(k) -- A hydrodynamical simulation-based model for the impact of baryon physics on the non-linear matter power spectrum
• Chemulator: Fast, accurate thermochemistry for dynamical models through emulation
• A neural network emulator for the Lyman-$\alpha$ 1D flux power spectrum
• MGLenS: Modified gravity weak lensing simulations for emulation-based cosmological inference
• FLAMINGO: Calibrating large cosmological hydrodynamical simulations with machine learning

Cited by in corpus (4)

• CSST Cosmological Emulator I: Matter Power Spectrum Emulation with one percent accuracy to k = 10 h/Mpc
• Ten-dimensional neural network emulator for the nonlinear matter power spectrum
• Design and optimization of neural networks for multifidelity cosmological emulation
• Power Spectrum Emulators from Neural Networks and Tree-Based Methods