SpaceMath v2.0 is an upgraded version of SpaceMath v1.0, now integrating Machine Learning techniques to enhance its capabilities in analyzing Flavor-Changing Neutral Current (FCNC) processes at both tree and one-loop levels. Specifically, it includes: i) radiative decays $\ell_i \to \ell_j \gamma$, ii) three-body decays $\ell_i \to \ell_j \ell_k \bar{\ell}k$ for $\ell_i = \tau, \mu$ and $\ell_{j,k} = \mu, e$ with $\ell_i \neq \ell_j \neq \ell_k$, iii) the anomalous magnetic moment of the muon $\delta a\mu$, and iv) rare decays $B^0_{s,d} \to \mu^+ \mu^-$.

SpaceMath v2.0 introduces a novel Machine Learning-based module that predicts Benchmark Points for efficient numerical evaluations of observables in particle physics. A detailed case study is presented for the type-III Two-Higgs Doublet Model (2HDM).