Detecting the phase transition in a strongly-interacting Fermi gas by unsupervised machine learning

We study the critical temperature of the superfluid phase transition of strongly-interacting fermions in the crossover regime between a Bardeen-Cooper-Schrieffer (BCS) superconductor and a Bose-Einstein condensate (BEC) of dimers. To this end, we employ the technique of unsupervised machine learning using an autoencoder neural network which we directly apply to time-of-flight images of the fermions. We extract the critical temperature of the phase transition from trend changes in the data distribution revealed in the latent space of the autoencoder bottleneck.