Abstract:Predicting the phase diagram of interacting quantum many-body systems is a central problem in condensed matter physics and related fields. A variety of quantum many-body systems, ranging from unconventional superconductors to spin liquids, exhibit complex competing phases whose theoretical description has been the focus of intense efforts. Here, we show that neural network quantum states can be combined with a Lee-Yang theory of quantum phase transitions to predict the critical points of strongly-correlated spin lattices. Specifically, we implement our approach for quantum phase transitions in the transverse-field Ising model on different lattice geometries in one, two, and three dimensions. We show that the Lee-Yang theory combined with neural network quantum states yields predictions of the critical field, which are consistent with large-scale quantum many-body methods. As such, our results provide a starting point for determining the phase diagram of more complex quantum many-body systems, including frustrated Heisenberg and Hubbard models.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to predict the phase diagrams of interacting quantum many - body systems, especially the critical points of strongly - correlated spin lattices. Specifically, the author combines Neural Network Quantum States (NQS) and Lee - Yang theory to predict the critical fields of the Transverse - field Ising model under different dimensions and lattice geometries. ### Core of the problem 1. **Predicting phase diagrams**: Predicting the phase diagrams of interacting quantum many - body systems is a core problem, especially in condensed - matter physics and related fields. Many quantum many - body systems exhibit complex competing phases, and their theoretical descriptions have always been the focus of research. 2. **Handling complex systems**: For high - dimensional systems, the Hilbert space grows exponentially with the system size, making most practical models difficult to handle. Although the one - dimensional transverse - field Ising model can be solved analytically, systems in higher dimensions are still challenging. 3. **Utilizing neural network quantum states**: Neural network quantum states, as variational states, can effectively parameterize and find the ground states of interacting quantum many - body systems. By optimizing the parameters of the neural network to minimize its energy, an accurate approximation of the ground state can be found. ### Research methods 1. **Transverse - field Ising model**: - The model Hamiltonian is: \[ \hat{H}=-J \sum_{\langle i,j \rangle} \hat{\sigma}_i^z \hat{\sigma}_j^z - h \sum_i \hat{\sigma}_i^x \] - Here, \( J \) is the coupling strength between nearest neighbors, and \( h \) is the transverse magnetic field. 2. **Neural network quantum states**: - Use a neural network to represent the wave function \( \psi_{\vec{\theta}}(\vec{\sigma})=\langle \vec{\sigma}|\psi_{\vec{\theta}}\rangle \), where \( \vec{\theta} \) are the variational parameters. - Minimize the energy through stochastic reconfiguration to find the ground state. 3. **Lee - Yang theory**: - Utilize the moment - generating function \( \chi(s)=\langle e^{s \hat{M}_z}\rangle \) in Lee - Yang theory to describe the fluctuations of magnetic susceptibility. - Determine the critical point by analyzing the behavior of the complex zeros of the moment - generating function in the thermodynamic limit. ### Main results 1. **Prediction of critical fields**: - For different dimensions and lattice geometries (such as linear chains, two - dimensional square lattices, three - dimensional cubic lattices, etc.), the critical fields are successfully predicted. - For example, the critical field of a one - dimensional chain is \( h_c = 1.00J \), that of a two - dimensional square lattice is \( h_c = 3.05J \), and that of a three - dimensional cubic lattice is \( h_c = 5.16J \). 2. **Comparison with mean - field theory**: - As the dimension and coordination number increase, the prediction results gradually approach the mean - field theory. - The mean - field theory ignores quantum fluctuations, so the actual critical field is usually smaller than the mean - field prediction. ### Conclusion This research shows that the method of combining Lee - Yang theory with neural network quantum states has potential in predicting the phase diagrams of complex quantum many - body systems. This method is not only applicable to two - dimensional and three - dimensional systems but can also be extended to other complex quantum models, such as the Heisenberg model and the fermionic Hubbard model.

Lee-Yang theory of quantum phase transitions with neural network quantum states

Lee-Yang formalism for phase transitions of interacting fermions using tensor networks

Learning Spin Liquids on a Honeycomb Lattice with Artificial Neural Networks.

Probing Quantum Critical Phase from Neural Network Wavefunction

Learning topological defects formation with neural networks in a quantum phase transition

A neural network study of the phase transitions of the two-dimensional antiferromagnetic $q$-state Potts models on the square lattice

Extracting Critical Exponents by Finite-Size Scaling with Convolutional Neural Networks

Experimental Investigation of Lee-Yang Criticality Using Non-Hermitian Quantum System

Machine Learning Z2 Quantum Spin Liquids with Quasiparticle Statistics

Quantum phase transitions in antiferromagnetic planar cubic lattices

A Neural Networks study of the phase transitions of Potts model

Neural-network quantum-state study of the long-range antiferromagnetic Ising chain

Machine Learning Phases of Strongly Correlated Fermions

Interpreting Machine Learning of Topological Quantum Phase Transitions

Machine learning quantum criticality in the spin-1/2 quantum antiferromagnets on the square lattice with plaquette structure

Quantum Classical Algorithm for the Study of Phase Transitions in the Hubbard Model via Dynamical Mean-Field Theory

Neural-network quantum state study of the long-range antiferromagnetic Ising chain

Quantum Phase Transitions in Anti-ferromagnetic Planar Cubic Lattices

Deep learning lattice gauge theories

In-Depth Investigation of Phase Transition Phenomena in Network Models Derived from Lattice Models