We propose and analyze a family of approximately symmetric neural networks for quantum spin liquid problems. These tailored architectures are parameter efficient, scalable, and significantly outperform existing symmetry-unaware neural network architectures. Utilizing the mixed-field toric code and PXP Rydberg Hamiltonian models, we demonstrate that our approach is competitive with state-of-the-art tensor network and quantum Monte Carlo methods. Moreover, at the largest system sizes ($N=480$ for toric code, $N=1584$ for Rydberg PXP), our method allows us to explore Hamiltonians with sign problems beyond the reach of both quantum Monte Carlo and finite-size matrix-product states. The network comprises an exactly symmetric block following a nonsymmetric block, which we argue learns a transformation of the ground state analogous to quasiadiabatic continuation. Our Letter paves the way toward investigating quantum spin liquid problems within interpretable neural network architectures.

Close

PrevNext

View figure in article

• Quantum correlations in quantum information
• Quantum spin liquid
• Quantum many-body systems
• Topological materials
• Artificial neural networks
• Symmetries in condensed matter