Abstract:The many-body Green's function provides access to electronic properties beyond density functional theory level in ab inito calculations. In this manuscript, we propose a deep learning framework for predicting the finite-temperature Green's function in atomic orbital space, aiming to achieve a balance between accuracy and efficiency. By predicting the self-energy matrices in Lehmann representation using an equivariant message passing neural network, our method respects its analytical property and the $E(3)$ equivariance. The Green's function is obtained from the predicted self-energy through Dyson equation with target total number of electrons. We present proof-of-concept benchmark results for both molecules and simple periodic systems, showing that our method is able to provide accurate estimate of physical observables such as energy and density of states based on the predicted Green's function.

What problem does this paper attempt to address?

The paper aims to address the problem of efficiently and accurately predicting the Green's functions of molecules and materials. Specifically, the authors propose a deep learning framework that utilizes equivariant message-passing neural networks to predict finite-temperature Green's functions and their self-energy matrices in the atomic orbital space. The main goal of this approach is to improve computational efficiency while maintaining accuracy. The main contributions of the paper can be summarized as follows: 1. **Deep Learning Framework**: A deep learning method is proposed for predicting many-body Green's functions and self-energy matrices of molecules and periodic systems. This method balances computational accuracy and efficiency. 2. **Equivariant Message-Passing Neural Networks**: Equivariant message-passing neural networks are used to predict Green's functions and self-energy matrices, ensuring that the analytical properties of physical quantities and E(3) equivariance are satisfied. 3. **Application of Lehmann Representation**: Green's functions and self-energy matrices are constructed using the Lehmann representation, ensuring that these functions satisfy the Carathéodory constraint conditions, thereby guaranteeing causality and the conservation of probability density. 4. **Validation with Numerical Results**: The effectiveness of the proposed method is validated through numerical experiments on molecules (such as water, benzene, and ethanol) and simple periodic systems (such as diamond and silicon). The results show that the method can provide accurate estimates of physical observables such as energy and density of states. In summary, this study proposes a novel approach to predicting the electronic properties of molecules and materials and validates its applicability and accuracy at different temperatures through numerical experiments. This provides a powerful tool for understanding and simulating complex quantum systems.

Equivariant neural network for Green's functions of molecules and materials

Predicting interacting Green's functions with neural networks

General framework for E(3)-equivariant neural network representation of density functional theory Hamiltonian

Equivariant graph neural networks for fast electron density estimation of molecules, liquids, and solids

Equivariant graph neural network interatomic potential for Green-Kubo thermal conductivity in phase change materials

E(3)-equivariant graph neural networks for data-efficient and accurate interatomic potentials

A deep equivariant neural network approach for efficient hybrid density functional calculations

Lightweight equivariant interaction graph neural network for accurate and efficient interatomic potential and force predictions

Learning Equivariant Non-Local Electron Density Functionals

Machine Learning Many-Body Green's Functions for Molecular Excitation Spectra

Lightweight Equivariant Model for Efficient Machine Learning Interatomic Potentials

Unified Deep Learning Framework for Many-Body Quantum Chemistry via Green's Functions

Higher-Order Equivariant Neural Networks for Charge Density Prediction in Materials

Learning local equivariant representations for large-scale atomistic dynamics

Predicting nonequilibrium Green's function dynamics and photoemission spectra via nonlinear integral operator learning

Equivariant Graph Attention Networks for Molecular Property Prediction

Neural network representation of electronic structure from $ab$ $initio$ molecular dynamics

Efficient and Equivariant Graph Networks for Predicting Quantum Hamiltonian

Equivariant Neural Networks Utilizing Molecular Clusters for Accurate Molecular Crystal Lattice Energy Predictions