High-Throughput GW Calculations via Machine Learning
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866910928126083072 |
|---|---|
| author | Abdelghany, Ragab. A. Hsu, Chih-En Hsueh, Hung-Chung Tsai, Yuan-Hong Chung, Ming-Chiang |
| author_facet | Abdelghany, Ragab. A. Hsu, Chih-En Hsueh, Hung-Chung Tsai, Yuan-Hong Chung, Ming-Chiang |
| contents | We present a machine learning (ML) framework that predicts $G_0W_0$ quasiparticle energies across molecular dynamics (MD) trajectories with high accuracy and efficiency. Using only DFT-derived mean-field eigenvalues and exchange-correlation potentials, the model is trained on 25\% of MD snapshots and achieves RMSEs below 0.1 eV. It accurately reproduces k-resolved quasiparticle band structures and density of states, even for BN polymorphs excluded from the training data. This approach bypasses the computational bottlenecks of $G_0W_0$ simulations over dynamic configurations, offering a scalable route to excited-state electronic structure simulations with many-body accuracy. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_02421 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | High-Throughput GW Calculations via Machine Learning Abdelghany, Ragab. A. Hsu, Chih-En Hsueh, Hung-Chung Tsai, Yuan-Hong Chung, Ming-Chiang Materials Science Disordered Systems and Neural Networks We present a machine learning (ML) framework that predicts $G_0W_0$ quasiparticle energies across molecular dynamics (MD) trajectories with high accuracy and efficiency. Using only DFT-derived mean-field eigenvalues and exchange-correlation potentials, the model is trained on 25\% of MD snapshots and achieves RMSEs below 0.1 eV. It accurately reproduces k-resolved quasiparticle band structures and density of states, even for BN polymorphs excluded from the training data. This approach bypasses the computational bottlenecks of $G_0W_0$ simulations over dynamic configurations, offering a scalable route to excited-state electronic structure simulations with many-body accuracy. |
| title | High-Throughput GW Calculations via Machine Learning |
| topic | Materials Science Disordered Systems and Neural Networks |
| url | https://arxiv.org/abs/2505.02421 |