Fig. 2.

Download original image
Flowchart describing the emulation process (from left to right). The N-body simulations first need to be run to provide the matter density distribution for the given DDM parameters and redshift (f, vk, Γ, z). Next, we extracted power spectra and calculated the resulting power suppression PDDM/PΛCDM. Additionally, we performed PCA on PDDM/PΛCDM data, and using five principal components , we recovered the original power suppression. Once the training data set was prepared, we trained the SIREN-like neural network. We used (f, vk, Γ, z) as the network input and output five PCA components. The PCA components obtained were used to reconstruct the power spectrum suppressions, which were then compared to the input ones. During the training process, the MSE loss between the input and output power suppression curves was minimised.
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.