Neural Network Quantum States analysis of the Shastry-Sutherland model

• DOI: 10.48550/arxiv.2303.14108
• Publication Date: 2023-01-01
• Authors: Matĕj Mezera; Jana Menšíková; Pavel Baláž; Martin Žonda
• Keywords: Boltzmann machine; Benchmark (surveying); Heisenberg model; Restricted Boltzmann machine; Lattice (music)

We utilize neural network quantum states (NQS) to investigate the ground state properties of Heisenberg model on a Shastry-Sutherland lattice using variational Monte Carlo method. show that already relatively simple NQSs can be used approximate this in its different phases and regimes. first compare several types with each other small lattices benchmark their energies against exact diagonalization results. argue when precision, generality, computational costs are taken into account, good choice for addressing larger systems is shallow restricted Boltzmann machine NQS. then such NQS describe main zero magnetic field. Moreover, based correctly describes intriguing plateaus forming magnetization as function increasing