Transformer Wave Function for the Shastry-Sutherland Model: emergence of a Spin-Liquid Phase

Quantum magnetism in two-dimensional systems represents a lively branch of modern condensed-matter physics. In the presence of competing super-exchange couplings, magnetic order is frustrated and can be suppressed down to zero temperature, leading to exotic ground states. The Shastry-Sutherland model, describing S=1/2 degrees of freedom interacting in a two-dimensional lattice, portrays a simple example of highly-frustrated magnetism, capturing the low-temperature behavior of SrCu_2(BO_3)_2 with its intriguing properties. Here, we investigate this problem by using a Vision Transformer to define an extremely accurate variational wave function. From a technical side, a pivotal achievement relies on using a deep neural network with real-valued parameters, parametrized with a Transformer, to map physical spin configurations into a high-dimensional feature space. Within this abstract space, the determination of the ground-state properties is simplified, requiring only a single output layer with complex-valued parameters. From the physical side, we supply strong evidence for the stabilization of a spin-liquid between the plaquette and antiferromagnetic phases. Our findings underscore the potential of Neural-Network Quantum States as a valuable tool for probing uncharted phases of matter, opening opportunities to establish the properties of many-body systems.