In condensed matter physics, one of the goals of machine learning is the classification of phases of matter. The consideration of a system's symmetries can significantly assist the machine in this goal. We demonstrate the ability of an unsupervised machine learning protocol, the Principal Component Analysis method, to detect hidden quenched gauge symmetries introduced via the so-called Mattis gauge transformation. Our work reveals that unsupervised machine learning can identify hidden properties of a model and may therefore provide new insights into the models themselves.
D. Lozano-G'omez, Darren Pereira, M. Gingras
arXiv: Disordered Systems and Neural Networks