Abstract:Machine learning methods that predict initial guesses from molecular geometry can reduce this cost, but matrix-prediction models fail when extrapolating to larger molecules, degrading rather than accelerating convergence [Liu et al. 2025]. We show that this failure is a supervision problem, not an extrapolation problem: models trained on ground-state targets fit those targets well out of distribution, yet produce initial guesses that slow convergence. Solver-Aligned Initialization Learning (SAIL) resolves this for both Hamiltonian and density matrix models by differentiating through the SCF solver end-to-end. We introduce the Effective Relative Iteration Count (ERIC), a correction to the commonly used RIC that accounts for hidden Fock-build overhead. On QM40, containing molecules up to 4$\times$ larger than the training distribution, SAIL reduces ERIC by 37% (PBE), 33% (SCAN), and 27% (B3LYP), more than doubling the previous state-of-the-art reduction on B3LYP (10%). On QMugs molecules 10$\times$ the training size, SAIL delivers a 1.25$\times$ wall-time speedup at the hybrid level of theory, extending ML SCF acceleration to large drug-like molecules.
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2604.21657 [cs.LG] |
| (or arXiv:2604.21657v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2604.21657 arXiv-issued DOI via DataCite (pending registration) |
From: Eike Eberhard [view email]
[v1]
Thu, 23 Apr 2026 13:20:34 UTC (1,352 KB)
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。