Figure 6
From: Machine learning quantum phases of matter beyond the fermion sign problem
An example of transfer learning in an artificial neural network. A CNN that was trained to discriminate the phases of the sign-problem free, spinful Hubbard model (4) and then applied to identify the phases and phase transition of the sign-problematic, spinless Hubbard model (10). The network is found to reliably distinguish the fermionic phases of the spinless model and provides a relatively accurate estimate for the location of the phase transition (the vertical line indicates the location of the transition in the thermodynamic limit of infinite system size).
