Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2507.03853 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866916826479329280 |
|---|---|
| author | Kang, Beom Seok Bhethanabotla, Vignesh C. Tavakoli, Amin Hanisch, Maurice D. Goddard III, William A. Anandkumar, Anima |
| author_facet | Kang, Beom Seok Bhethanabotla, Vignesh C. Tavakoli, Amin Hanisch, Maurice D. Goddard III, William A. Anandkumar, Anima |
| contents | Despite the success of deep learning methods in quantum chemistry, their representational capacity is most often confined to neutral, closed-shell molecules. However, real-world chemical systems often exhibit complex characteristics, including varying charges, spins, and environments. We introduce OrbitAll, a geometry- and physics-informed deep learning framework that can represent all molecular systems with electronic structure information. OrbitAll utilizes spin-polarized orbital features from the underlying quantum mechanical method, and combines it with graph neural networks satisfying SE(3)-equivariance. The resulting framework can represent and process any molecular system with arbitrary charges, spins, and environmental effects. OrbitAll demonstrates superior performance and generalization on predicting charged, open-shell, and solvated molecules, while also robustly extrapolating to molecules significantly larger than the training data by leveraging a physics-informed architecture. OrbitAll achieves chemical accuracy using 10 times fewer training data than competing AI models, with a speedup of approximately $10^3$ - $10^4$ compared to density functional theory. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2507_03853 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | OrbitAll: A Unified Quantum Mechanical Representation Deep Learning Framework for All Molecular Systems Kang, Beom Seok Bhethanabotla, Vignesh C. Tavakoli, Amin Hanisch, Maurice D. Goddard III, William A. Anandkumar, Anima Machine Learning Chemical Physics Despite the success of deep learning methods in quantum chemistry, their representational capacity is most often confined to neutral, closed-shell molecules. However, real-world chemical systems often exhibit complex characteristics, including varying charges, spins, and environments. We introduce OrbitAll, a geometry- and physics-informed deep learning framework that can represent all molecular systems with electronic structure information. OrbitAll utilizes spin-polarized orbital features from the underlying quantum mechanical method, and combines it with graph neural networks satisfying SE(3)-equivariance. The resulting framework can represent and process any molecular system with arbitrary charges, spins, and environmental effects. OrbitAll demonstrates superior performance and generalization on predicting charged, open-shell, and solvated molecules, while also robustly extrapolating to molecules significantly larger than the training data by leveraging a physics-informed architecture. OrbitAll achieves chemical accuracy using 10 times fewer training data than competing AI models, with a speedup of approximately $10^3$ - $10^4$ compared to density functional theory. |
| title | OrbitAll: A Unified Quantum Mechanical Representation Deep Learning Framework for All Molecular Systems |
| topic | Machine Learning Chemical Physics |
| url | https://arxiv.org/abs/2507.03853 |