A swampland conjecture DESIderátum?

Abstract: A key result of the string phenomenology framework is the two swampland conjectures (SCs), namely that the values of the scalar field excursion and gradient of the potential have to be $\mathcal{O}(1)$ in Planck units. Using the first data release from the Dark Energy Spectroscopic Instrument (DESI) survey and model-independent reconstructions of the SCs at late times, via a particular machine learning approach known as the genetic algorithms (GA), we find that the reconstructions of the SCs are, in fact, in agreement with their theoretical values of $\mathcal{O}(1)$. Specifically, the reconstruction of the second SC is in fact several sigmas away from zero, thus indicating a very steep potential in contrast to recent model-specific analyses, assuming exponential potentials. Our approach is based solely on model-independent reconstructions of cosmological observables, such as the angular diameter distance and the Hubble expansion history, thus it is readily applicable to the forthcoming data from Stage IV surveys and will help settle in the near future the issue on the observational falsification, or not, of the two conjectures.