OrbitAll: A Unified Quantum Mechanical Representation Deep Learning Framework for All Molecular Systems
This addresses the problem of modeling complex real-world chemical systems with varying charges, spins, and environments for researchers in quantum chemistry and materials science, representing a novel method rather than an incremental improvement.
The paper tackles the limitation of deep learning methods in quantum chemistry to neutral, closed-shell molecules by introducing OrbitAll, a framework that represents all molecular systems with electronic structure information, achieving chemical accuracy with 10 times fewer training data and a speedup of 10^3-10^4 compared to density functional theory.
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.