Research paperComputed PhononAccelerating the calculation of electron-phonon coupling by machine learning methodsYang Zhong, Zhiguo Tao, Weibin Chu, Xingao Gong et al.2023·10.48550/arxiv.2302.00439·arXiv:2302.00439AbstractElectron-phonon coupling (EPC) is important for many physical phenomena but expensive to compute. The paper derives an analytical formula for EPC in nonorthogonal atomic orbital bases in terms of the Hamiltonian and its gradients, and uses an E(3)-equivariant neural network to predict those quantities efficiently, bypassing expensive self-consistent DFT iterations. The approach is tested on a water molecule and a MoS₂ supercell and shown to reproduce DFT EPC values closely.Read more
Water molecule used as a test system for EPC prediction and validation against DFT.1 characterization2 figuresSimulatedH₂OStudied MaterialExpand
MoS₂ supercell used as a test system for EPC prediction and validation against DFT.1 characterization2 figuresSimulated Supercell DftMoS₂Studied MaterialExpand
Research paperComputed PhononAccelerating the calculation of electron-phonon coupling by machine learning methodsYang Zhong, Zhiguo Tao, Weibin Chu, Xingao Gong et al.2023·10.48550/arxiv.2302.00439·arXiv:2302.00439AbstractElectron-phonon coupling (EPC) is important for many physical phenomena but expensive to compute. The paper derives an analytical formula for EPC in nonorthogonal atomic orbital bases in terms of the Hamiltonian and its gradients, and uses an E(3)-equivariant neural network to predict those quantities efficiently, bypassing expensive self-consistent DFT iterations. The approach is tested on a water molecule and a MoS₂ supercell and shown to reproduce DFT EPC values closely.Read more
Water molecule used as a test system for EPC prediction and validation against DFT.1 characterization2 figuresSimulatedH₂OStudied MaterialExpand
MoS₂ supercell used as a test system for EPC prediction and validation against DFT.1 characterization2 figuresSimulated Supercell DftMoS₂Studied MaterialExpand
Research paperComputed PhononAccelerating the calculation of electron-phonon coupling by machine learning methodsYang Zhong, Zhiguo Tao, Weibin Chu, Xingao Gong et al.2023·10.48550/arxiv.2302.00439·arXiv:2302.00439AbstractElectron-phonon coupling (EPC) is important for many physical phenomena but expensive to compute. The paper derives an analytical formula for EPC in nonorthogonal atomic orbital bases in terms of the Hamiltonian and its gradients, and uses an E(3)-equivariant neural network to predict those quantities efficiently, bypassing expensive self-consistent DFT iterations. The approach is tested on a water molecule and a MoS₂ supercell and shown to reproduce DFT EPC values closely.Read more
Water molecule used as a test system for EPC prediction and validation against DFT.1 characterization2 figuresSimulatedH₂OStudied MaterialExpand
MoS₂ supercell used as a test system for EPC prediction and validation against DFT.1 characterization2 figuresSimulated Supercell DftMoS₂Studied MaterialExpand
Research paperComputed PhononAccelerating the calculation of electron-phonon coupling by machine learning methodsYang Zhong, Zhiguo Tao, Weibin Chu, Xingao Gong et al.2023·10.48550/arxiv.2302.00439·arXiv:2302.00439AbstractElectron-phonon coupling (EPC) is important for many physical phenomena but expensive to compute. The paper derives an analytical formula for EPC in nonorthogonal atomic orbital bases in terms of the Hamiltonian and its gradients, and uses an E(3)-equivariant neural network to predict those quantities efficiently, bypassing expensive self-consistent DFT iterations. The approach is tested on a water molecule and a MoS₂ supercell and shown to reproduce DFT EPC values closely.Read more
Water molecule used as a test system for EPC prediction and validation against DFT.1 characterization2 figuresSimulatedH₂OStudied MaterialExpand
MoS₂ supercell used as a test system for EPC prediction and validation against DFT.1 characterization2 figuresSimulated Supercell DftMoS₂Studied MaterialExpand
