Machine learning cosmic inflation by Ahana Kamerkar et al. on Tuesday 29 NovemberWe present a machine-learning approach, based on the genetic algorithms (GA),that can be used to reconstruct the inflationary potential directly fromcosmological data. We create a pipeline consisting of the GA, a primordial codeand a Boltzmann code used to calculate the theoretical predictions, and CosmicMicrowave Background (CMB) data. As a proof of concept, we apply ourmethodology to the Planck CMB data and explore the functional space ofsingle-field inflationary potentials in a non-parametric, yet analytical way.We show that the algorithm easily improves upon the vanilla model of quadraticinflation and proposes slow-roll potentials better suited to the data, while weconfirm the robustness of the Starobinsky inflation model (and othersmall-field models). Moreover, using unbinned CMB data, we perform a firstconcrete application of the GA by searching for oscillatory features in thepotential in an agnostic way, and find very significant improvements upon thebest featureless potentials, $\Delta \chi^2 < -20$. These encouragingpreliminary results motivate the search for resonant features in the primordialpower spectrum with a multimodal distribution of frequencies. We stress thatour pipeline is modular and can easily be extended to other CMB data sets andinflationary scenarios, like multifield inflation or theories with higher-orderderivatives.arXiv: http://arxiv.org/abs/http://arxiv.org/abs/2211.14142v1