Many-body Δ-Machine Learning brings the
accuracy of conventional force field to coupled
cluster: application to the TTM2.1 water
force field

Qu, Chen; Yu, Qi; Houston, Paul L.; Conte, Riccardo; Nandi, Apurba; Bowman, Joel M.

doi:10.21203/rs.3.rs-1847384/v1

Cited by 3 publications

(2 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The first aspect is the adaptability of the model to dynamically changing DP regions, allowing atoms to enter or leave these regions. The ∆E-MLP/MM methods employ MLPs to elevate the quality of the QM region from semi-empirical to ab initio QM levels 63,80,81 . This strategy can be viewed as a hybrid MLP+QM/MM scheme, albeit one that remains reliant on and constrained by existing QM/MM methods.…”

Section: Conclusion and Discussionmentioning

confidence: 99%

“…The entire system is also evaluated using a conventional FF uniformly during the simulations. The added DP term bridges the gap between the higher-level QM and the MM forces on zinc and zinc-coordinated atoms, which can be considered as a ∆ term [61][62][63] . In DP/MM simulations, zinc ions and their associated ligand atoms included in the DP description are selected adaptively.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

DP/MM: A Hybrid Model for Zinc-Protein Interactions in Molecular Dynamics

Ding,

Huang

2023

Preprint

View full text Add to dashboard Cite

Zinc-containing proteins are essential to a variety of biological processes, yet accurately modeling them using classical force fields is hindered by complicated polarization and charge transfer effects. This study introduces DP/MM, a hybrid force field model that combinesab initioaccuracy with MM-level efficiency for modeling zinc-protein interactions. The DP/MM scheme utilizes a deep potential model to correct the atomic forces of zinc ions and their coordinated atoms, elevating them from MM to QM levels of accuracy. The model is trained on the difference in atomic forces between MM and QM calculations across diverse zinc coordination groups. Simulations on a variety of zinccontaining proteins demonstrate that DP/MM faithfully reproduces their coordination geometry and structural characteristics, for example, the tetrahedral coordination structures for theCys4and theCys3His1groups. Furthermore, DP/MM is capable of handling exchangeable water molecules in the zinc coordination environment. With its unique blend of accuracy, efficiency, flexibility, and transferability, DP/MM not only serves as a valuable tool for studying zinc-containing proteins but also represents a pioneering approach that augments the growing landscape of machine learning potentials in molecular modeling.

show abstract

Section: Conclusion and Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

DP/MM: A Hybrid Model for Zinc-Protein Interactions in Molecular Dynamics

Ding,

Huang

2023

Preprint

View full text Add to dashboard Cite

show abstract

Kohn–Sham accuracy from orbital-free density functional theory via Δ-machine learning

Kumar,

Jing,

Pask

et al. 2023

The Journal of Chemical Physics

View full text Add to dashboard Cite

We present a Δ-machine learning model for obtaining Kohn–Sham accuracy from orbital-free density functional theory (DFT) calculations. In particular, we employ a machine-learned force field (MLFF) scheme based on the kernel method to capture the difference between Kohn–Sham and orbital-free DFT energies/forces. We implement this model in the context of on-the-fly molecular dynamics simulations and study its accuracy, performance, and sensitivity to parameters for representative systems. We find that the formalism not only improves the accuracy of Thomas–Fermi–von Weizsäcker orbital-free energies and forces by more than two orders of magnitude but is also more accurate than MLFFs based solely on Kohn–Sham DFT while being more efficient and less sensitive to model parameters. We apply the framework to study the structure of molten Al0.88Si0.12, the results suggesting no aggregation of Si atoms, in agreement with a previous Kohn–Sham study performed at an order of magnitude smaller length and time scales.

show abstract

Artificial Intelligence for Web-based Educational Systems

Wang

2022

AIST

View full text Add to dashboard Cite

Due to the global COVID-19 epidemic in the preceding two years, there has been a significant debate among different academics about how learners may be lectured through the web while maintaining a higher degree of cognitive efficiency. Students may have problems concentrating on their work because of the absence of teacher-student connection, but there are benefits to online learning that are not offered in conventional classrooms. The Adaptive and Intelligent Web-based Educational Systems (AIWES) is a platform that incorporates the design of students' online courses. RLATES is an AIWES that uses reinforcement learning to build instructional tactics. This research intends the aggregation and evaluation of the present research, model classification, and design techniques for integrated functional academic frameworks as a precondition to undertaking research in this subject, with the purpose of acting as an academic standard in the related fields to aid them obtain accessibility to fundamental materials conveniently and quickly.

show abstract

Many-body Δ-Machine Learning brings the accuracy of conventional force field to coupled cluster: application to the TTM2.1 water force field

Cited by 3 publications

References 47 publications

DP/MM: A Hybrid Model for Zinc-Protein Interactions in Molecular Dynamics

DP/MM: A Hybrid Model for Zinc-Protein Interactions in Molecular Dynamics

Kohn–Sham accuracy from orbital-free density functional theory via Δ-machine learning

Artificial Intelligence for Web-based Educational Systems

Contact Info

Product

Resources

About