Faculty @ MIT, US
Thursday June 30 – 17.45 BST
Machine-learning-accelerated Bose-Einstein condensation
Machine learning is emerging as a technology that can enhance physics experiment and data analysis. Here, we apply machine learning to accelerate the production of a Bose-Einstein condensate across the phase transition from a classical to a quantum gas. Starting from a room-temperature gas and optimizing laser cooling (Raman sideband cooling) using a Bayesian approach with up to 55 control parameters, we prepare a condensate of rubidium atoms in less than 0.6 seconds; the fastest condensation to date. We find that the choice of cost function for the algorithm strongly influences the trade-off between large and pure condensates. We anticipate that many other physics experiments with complex nonlinear system dynamics or involving phase transitions can be significantly enhanced by a similar machine-learning approach.