Searching ground states of the heisenberg antiferromagnet on a triangular lattice using machine learning methods

We consider an antiferromagnet with classical spins on two-dimensional triangular lattice in the external magnetic field with the single-ion anisotropy. Using a novel machine learning method, we calculate the ground states and obtain the magnetization phase diagram in dependence on the external magnetic field and anisotropy coefficient. We show that the calculation results are in good agreement with Monte-Carlo simulations. The anisotropy leads to the 1/3 magnetization plateau in the range of external field that corresponds to the uud ground state. The size of this plateau substantially increases with increasing strength of anisotropy.