BaCoN (BAyesian COsmological Network) trains and tests Bayesian Convolutional Neural Networks in order to classify dark matter power spectra as being representative of different cosmologies, as well as to compute the classification confidence. It supports the following theories: LCDM, wCDM, f(R), DGP, and a randomly generated class. Additional cosmologies can be easily added.
https://ui.adsabs.harvard.edu/abs/2022PhRvD.105b3531M and a link to https://github.com/Mik3M4n/BaCoN