Learning Conformal Field Theory with Symbolic Regression: Recovering the Symbolic Expressions for the Energy Spectrum

In this paper, we employ the pySR Symbolic Regression (SR) package to recover symbolic formulae for the primary conformal weights and central charges of two-dimensional rational conformal field theories (2D-RCFTs) given only limited information about the energy spectra of these theories. We find symbolic expressions for these quantities in two well-understood classes of 2D-RCFT, minimal models and Wess-Zumino-Witten (WZW) $\hat{su}(2)_k$ models, and discuss future applications to theories about which less is known.