Utilizing Machine Learning for Efficient Parameterization of Coarse Grained Molecular Force Fields

McDonagh, J. E. R.; Shkurti, Ardita; Bray, David J.; Anderson, Richard L.; Pyzer‐Knapp, Edward O.

doi:10.26434/chemrxiv.8910815

Cited by 16 publications

(20 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…2,3,78 The development of high-accuracy atomistic or coarse-grained molecular models therefore should aim for a predictive error of between 0.03-1% of the experimental value. Our hand-tuned model comes close to this, but we anticipate that the accuracy could be further improved using automated methods (such as machine learning or classical parameter optimisation), 82 fitting holistically to liquid phase density data across the family of alkanes and mixtures thereof.…”

Section: Discussionmentioning

confidence: 65%

Wax Formation in Linear and Branched Alkanes with Dissipative Particle Dynamics

Bray

Anderson

Warren

et al. 2020

J. Chem. Theory Comput.

Self Cite

View full text Add to dashboard Cite

Please cite only the published version using the reference above. This is the citation assigned by the publisher at the time of issuing the AAM. Please check the publisher's website for any updates. This is the author's final, peer-reviewed manuscript as accepted for publication (AAM). The version presented here may differ from the published version, or version of record, available through the publisher's website. This version does not track changes, errata, or withdrawals on the publisher's site.

show abstract

Section: Discussionmentioning

confidence: 65%

Wax Formation in Linear and Branched Alkanes with Dissipative Particle Dynamics

Bray

Anderson

Warren

et al. 2020

J. Chem. Theory Comput.

Self Cite

View full text Add to dashboard Cite

show abstract

“…However, the challenges for a successful DPD simulation is finding robust and general methods for parameterization of the simulation system [32][33]. This is an active research area with recent approach to apply machine learning for DPD parameterization [34].…”

Section: Dissipative Particle Dynamics (Dpd)mentioning

confidence: 99%

Experimental and Computational Modeling of Microemulsion Phase Behavior

Hon¹,

Saaid²

2022

Surfactants and Detergents - Updates and New Insights

View full text Add to dashboard Cite

The phase behavior of microemulsions formed in a surfactant-brine-oil system for a chemical Enhanced Oil Recovery (EOR) application is complex and depends on a range of parameters. Phase behavior indicates a surfactant solubilization. Phase behavior tests are simple but time-consuming especially when it involves a wide range of surfactant choices at various concentrations. An efficient and insightful microemulsion formulation via computational simulation can complement phase behavior laboratory test. Computational simulation can predict various surfactant properties, including microemulsion phase behavior. Microemulsion phase behavior can be predicted predominantly using Quantitative Structure-Property Relationship (QSPR) model. QSPR models are empirical and limited to simple pure oil system. Its application domain is limited due to the model cannot be extrapolated beyond reference condition. Meanwhile, there are theoretical models based on physical chemistry of microemulsion that can predict microemulsion phase behavior. These models use microemulsion surface tension and torque concepts as well as with solution of bending rigidity of microemulsion interface with relation to surface solubilization and interface energy.

show abstract

“…Bayesian optimization has become popular recently in the machine-learning community for the efficient tuning of the hyperparameters of deep learning models, 11 but given its strengths as a global optimizer, and its powerful theoretical guarantees, 12 it has also started to find applications in an increasingly diverse set of domains. [13][14][15][16] The core application area of Bayesian optimization is when each sample of the function is expensive to acquire, either in financial cost, acquisition time, or both, thus making this approach very attractive for our goal of more efficiently navigating large ESF maps.…”

Section: ∈χ ( )mentioning

confidence: 99%

Distributed Multi-Objective Bayesian Optimization for the Intelligent Navigation of Energy Structure Function Maps For Efficient Property Discovery

Pyzer‐Knapp¹,

Day²,

Chen³

et al. 2020

Preprint

View full text Add to dashboard Cite

Energy-structure-function (ESF) maps have emerged as a powerful tool for in silico materials design, coupling crystal structure prediction techniques with property simulations to assess the potential for new candidate materials to display desirable properties. Despite continuing increases to accessible computational power, however, the computational cost of acquiring an ESF map often remains too high to allow integration into true high-throughput virtual screening techniques. In this paper, we propose the next evolution of the ESF map, which uses parallel Bayesian optimization to selectively acquire energy and property data, generating the same levels of insight at a fraction of the computational cost by limiting the expensive property calculations to a small fraction of the predicted crystal structures associated with a molecule. We utilize this approach to obtain a two orders of magnitude speedup on a previous ESF study that focused on methane capture materials, saving over 500,000 CPUh from the original protocol. Through acceleration of the acquisition of ESF-type insight, we pave the way for the use of ESF maps in automated ultra-high throughput screening pipelines. This greatly reduce the opportunity risk associated with the choice of system to calculate. For example, it will allow researchers to use ESF maps in the search for physical properties where the computational costs are currently just intractable, or to investigate orders of magnitude more systems for a given computational cost.<br>

show abstract

Utilizing Machine Learning for Efficient Parameterization of Coarse Grained Molecular Force Fields

Abstract: This work demonstrates the use of open literature data to force field paramterization via a novel approach applying Bayesian optimization. We have selected Dissipative Particle Dynamics (DPD) as the simulation method in this proof-of-concept work.

Cited by 16 publications

References 42 publications

Wax Formation in Linear and Branched Alkanes with Dissipative Particle Dynamics

Wax Formation in Linear and Branched Alkanes with Dissipative Particle Dynamics

Experimental and Computational Modeling of Microemulsion Phase Behavior

Distributed Multi-Objective Bayesian Optimization for the Intelligent Navigation of Energy Structure Function Maps For Efficient Property Discovery

Contact Info

Product

Resources

About