Abstract:We develop a pairing-based graph neural network for simulating quantum many-body systems. Our architecture augments a BCS-type geminal wavefunction with a generalized pair amplitude parameterized by a graph neural network. Variational Monte Carlo with our neural network simultaneously provides an accurate, flexible, and scalable method for simulating many-electron systems. We apply this method to two-dimensional semiconductor electron-hole bilayers and obtain accurate results on a variety of interaction-induced phases, including the exciton Bose-Einstein condensate, electron-hole superconductor, and bilayer Wigner crystal. Our study demonstrates the potential of physically-motivated neural network wavefunctions for quantum materials simulations.

What problem does this paper attempt to address?

The main objective of this paper is to develop a paired-based Graph Neural Network (GNN) method to simulate many-body systems in quantum materials. Specifically, the research team designed an enhanced BCS (Bardeen-Cooper-Schrieffer) type two-state wave function by introducing a generalized pairing amplitude parameterized by a Graph Neural Network to enhance the expressive power of the wave function. This method can effectively describe various interaction-induced phases in electron-hole bilayer systems, including exciton Bose-Einstein condensation, electron-hole superconductors, and bilayer Wigner crystals. Compared to traditional Hartree-Fock-Bogoliubov (HFB) theory and other approximation methods, this new neural network wave function provides higher accuracy while maintaining flexibility and scalability. The research team applied this method to a two-dimensional semiconductor electron-hole bilayer system and obtained precise results under different physical parameters, confirming the effectiveness and potential of this method in handling strongly correlated systems. Additionally, the paper details the design concepts and implementation details of the proposed neural network architecture, as well as how to optimize the wave function parameters using the variational Monte Carlo method to achieve the lowest energy state. In summary, this paper proposes an innovative approach to simulate complex quantum many-body systems, especially in two-dimensional electron-hole bilayer systems, demonstrating the powerful capability of using physics-inspired neural network wave functions for quantum material simulations.

Pairing-based graph neural network for simulating quantum materials

Ion-based quantum simulation of many-body electron–electron Coulomb interaction

Neural-network Variational Quantum Algorithm for Simulating Many-Body Dynamics

Simulating moiré quantum matter with neural network

Solving quantum many-particle models with graph attention network

A Study on Quantum Graph Neural Networks Applied to Molecular Physics

Transformer neural networks and quantum simulators: a hybrid approach for simulating strongly correlated systems

Quantum Simulation of Pairing Hamiltonians with Nearest-Neighbor-Interacting Qubits

Unsupervised Deep Neural Network Approach To Solve Bosonic Systems

Symmetries and many-body excited states with neural-network quantum states

Symmetrical Graph Neural Network for Quantum Chemistry, with Dual R/K Space

Neural network approach to quasiparticle dispersions in doped antiferromagnets

Neural-network quantum states for ultra-cold Fermi gases

Quantum Embedding Method with Transformer Neural Network Quantum States for Strongly Correlated Materials

Polynomially efficient quantum enabled variational Monte Carlo for training neural-network quantum states for physico-chemical applications

Quantum neural network approach to Markovian dissipative dynamics of many-body open quantum systems

Simulating Quantum Many-Body States with Neural-Network Exponential Ansatz

Highly resolved spectral functions of two-dimensional systems with neural quantum states