Machine-learning-accelerated Bose-Einstein condensation

Machine learning is emerging as a technology that can enhance physics experiment execution and data analysis. Here, we apply machine to accelerate the production of Bose-Einstein condensate (BEC) $^{87}\mathrm{Rb}$ atoms by Bayesian optimization up 55 control parameters. This approach enables us prepare BECs $2.8\ifmmode\times\else\texttimes\fi{}{10}^{3}$ optically trapped from room-temperature gas in 575 ms. The algorithm achieves fast BEC preparation applying highly efficient Raman cooling near quantum degeneracy, followed brief final evaporation. We anticipate many other experiments with complex nonlinear system dynamics be significantly enhanced similar machine-learning approach.