-
Considerations in the use of ML interaction potentials for free energy calculations
Authors:
Orlando A. Mendible,
Jonathan K. Whitmer,
Yamil J. Colón
Abstract:
Machine learning potentials (MLPs) offer the potential to accurately model the energy and free energy landscapes of molecules with the precision of quantum mechanics and an efficiency similar to classical simulations. This research focuses on using equivariant graph neural networks MLPs due to their proven effectiveness in modeling equilibrium molecular trajectories. A key issue addressed is the c…
▽ More
Machine learning potentials (MLPs) offer the potential to accurately model the energy and free energy landscapes of molecules with the precision of quantum mechanics and an efficiency similar to classical simulations. This research focuses on using equivariant graph neural networks MLPs due to their proven effectiveness in modeling equilibrium molecular trajectories. A key issue addressed is the capability of MLPs to accurately predict free energies and transition states by considering both the energy and the diversity of molecular configurations. We examined how the distribution of collective variables (CVs) in the training data affects MLP accuracy in determining the free energy surface (FES) of systems, using Metadynamics simulations for butane and alanine dipeptide (ADP). The study involved training forty-three MLPs, half based on classical molecular dynamics data and the rest on ab initio computed energies. The MLPs were trained using different distributions that aim to replicate hypothetical scenarios of sampled CVs obtained if the underlying FES of the system was unknown. Findings for butane revealed that training data coverage of key FES regions ensures model accuracy regardless of CV distribution. However, missing significant FES regions led to correct potential energy predictions but failed free energy reconstruction. For ADP, models trained on classical dynamics data were notably less accurate, while ab initio-based MLPs predicted potential energy well but faltered on free energy predictions. These results emphasize the challenge of assembling an all-encompassing training set for accurate FES prediction and highlight the importance of understanding the FES in preparing training data. The study points out the limitations of MLPs in free energy calculations, stressing the need for comprehensive data that encompasses the system's full FES for effective model training.
△ Less
Submitted 20 March, 2024;
originally announced March 2024.
-
Transfer Learning Facilitates the Prediction of Polymer-Surface Adhesion Strength
Authors:
Jiale Shi,
Fahed Albreiki,
Yamil J. Colón,
Samanvaya Srivastava,
Jonathan K. Whitmer
Abstract:
Machine learning (ML) accelerates the exploration of material properties and their links to the structure of the underlying molecules. In previous work [J. Shi, M. J. Quevillon, P. H. A. Valença, and J. K. Whitmer, \textit{ACS Appl. Mater. Interfaces.}, 2022, 14, 32, 37161--37169], ML models were applied to predict the adhesive free energy of polymer--surface interactions with high accuracy from t…
▽ More
Machine learning (ML) accelerates the exploration of material properties and their links to the structure of the underlying molecules. In previous work [J. Shi, M. J. Quevillon, P. H. A. Valença, and J. K. Whitmer, \textit{ACS Appl. Mater. Interfaces.}, 2022, 14, 32, 37161--37169], ML models were applied to predict the adhesive free energy of polymer--surface interactions with high accuracy from the knowledge of the sequence data, demonstrating successes in inverse-design of polymer sequence for known surface compositions. While the method was shown to be successful in designing polymers for a known surface, extensive datasets were needed for each specific surface in order to train the surrogate models. Ideally, one should be able to infer information about similar surfaces without having to regenerate a full complement of adhesion data for each new case. In the current work, we demonstrate a transfer learning (TL) technique using a deep neural network to improve the accuracy of ML models trained on small datasets by pre-training on a larger database from a related system and fine-tuning the weights of all layers with a small amount of additional data. The shared knowledge from the pre-trained model facilitates the prediction accuracy significantly on small datasets. We also explore the limits of database size on accuracy and the optimal tuning of network architecture and parameters for our learning tasks. While applied to a relatively simple coarse-grained (CG) polymer model, the general lessons of this study apply to detailed modeling studies and the broader problems of inverse materials design.
△ Less
Submitted 5 January, 2023;
originally announced January 2023.
-
Free Energy Landscape and Isomerization Rates of Au$_4$ Clusters at Finite Temperature
Authors:
Jiale Shi,
Shanghui Huang,
François Gygi,
Jonathan K. Whitmer
Abstract:
In metallic nanoparticles, the cluster geometric structures control the particle's electronic band structure, polarizability, and catalytic properties. Analyzing the structural properties is a complex problem; the structure of an assembled cluster changes from moment to moment due to thermal fluctuations. Conventional structural analyses based on spectroscopy or diffraction cannot determine the in…
▽ More
In metallic nanoparticles, the cluster geometric structures control the particle's electronic band structure, polarizability, and catalytic properties. Analyzing the structural properties is a complex problem; the structure of an assembled cluster changes from moment to moment due to thermal fluctuations. Conventional structural analyses based on spectroscopy or diffraction cannot determine the instantaneous structure exactly and can merely provide an averaged structure. Molecular simulations offer an opportunity to examine the assembly and evolution of metallic clusters, as the preferred assemblies and conformations can easily be visualized and explored. Here, we utilize the adaptive biasing force algorithm applied to first principles molecular dynamics to demonstrate exploration of a relatively simple system which permits comprehensive study of the small metal cluster $\ce{Au4}$ in both neutral and charged configurations. Our simulation work offers a quantitative understanding of these clusters' dynamic structure, which is significant for single-site catalytic reactions on metal clusters and provides a starting point for a detailed quantitative understanding of more complex pure metal and alloy clusters' dynamic properties.
△ Less
Submitted 11 October, 2021;
originally announced October 2021.
-
Predicting Adhesive Free Energies of Polymer--Surface Interactions with Machine Learning
Authors:
Jiale Shi,
Michael J. Quevillon,
Pedro H. Amorim Valença,
Jonathan K. Whitmer
Abstract:
Polymer-surface interactions are crucial to many biological processes and industrial applications. Here we propose a machine-learning method to connect a model polymer's sequence with its adhesion to decorated surfaces. We simulate the adhesive free energies of $20,000$ unique coarse-grained 1D sequential polymers interacting with functionalized surfaces and build support vector regression (SVR) m…
▽ More
Polymer-surface interactions are crucial to many biological processes and industrial applications. Here we propose a machine-learning method to connect a model polymer's sequence with its adhesion to decorated surfaces. We simulate the adhesive free energies of $20,000$ unique coarse-grained 1D sequential polymers interacting with functionalized surfaces and build support vector regression (SVR) models that demonstrate inexpensive and reliable prediction of the adhesive free energy as a function of the sequence. Our work highlights the promising integration of coarse-grained simulation with data-driven machine learning methods for the design of new functional polymers and represents an important step toward linking polymer compositions with polymer-surface interactions.
△ Less
Submitted 6 October, 2021;
originally announced October 2021.
-
Linking dynamics and structure in highly asymmetric ionic liquids
Authors:
Mariana E. Farías-Anguiano,
Ernesto C. Cortés-Morales,
Jonathan K. Whitmer,
Pedro E. Ramírez-González
Abstract:
We explore an idealized theoretical model for the transport of ions within highly asymmetric ionic liquid mixtures. A primitive model (PM)-inspired system serves as a representative for asymmetric ionic materials (such as liquid crystalline salts) which quench to form disordered, partially-arrested phases. Self-Consistent Generalized Langevin Equation (SCGLE) Theory is applied to understand the co…
▽ More
We explore an idealized theoretical model for the transport of ions within highly asymmetric ionic liquid mixtures. A primitive model (PM)-inspired system serves as a representative for asymmetric ionic materials (such as liquid crystalline salts) which quench to form disordered, partially-arrested phases. Self-Consistent Generalized Langevin Equation (SCGLE) Theory is applied to understand the connection between the size ratio of charge-matched salts and their average mobility. Within this model, we identify novel glassy states where one of the two charged species (either the macro-cation or the micro-anion) are arrested, while the other retains mobility. We discuss how this result is useful in the development of novel single-ion conducting phases in ionic liquid based materials.
△ Less
Submitted 27 August, 2021;
originally announced August 2021.
-
Automatic Determination of $n$-Cyanobiphenyl Elastic Constants from Molecular Simulation
Authors:
Hythem Sidky,
Jonathan K. Whitmer
Abstract:
New applications of liquid crystalline materials have increased the need for precise engineering of elastic properties. Recently, Sidky et al. presented methods by which the elastic coefficients of molecular models with atomistic detail can be accurately calculated, demonstrating the result for the ubiquitous mesogen 5CB. In this work, these techniques are applied to the homologous series of nCB m…
▽ More
New applications of liquid crystalline materials have increased the need for precise engineering of elastic properties. Recently, Sidky et al. presented methods by which the elastic coefficients of molecular models with atomistic detail can be accurately calculated, demonstrating the result for the ubiquitous mesogen 5CB. In this work, these techniques are applied to the homologous series of nCB materials, focusing on the standard bend, twist, and splay deformations, using an entirely automated process. Our results show strong agreement with published experimental measurements for the nCBs and present a path forward to computational molecular engineering of liquid crystal elasticity for novel molecules and mixtures.
△ Less
Submitted 28 February, 2019;
originally announced February 2019.
-
Learning Free Energy Landscapes Using Artificial Neural Networks
Authors:
Hythem Sidky,
Jonathan K. Whitmer
Abstract:
Existing adaptive bias techniques, which seek to estimate free energies and physical properties from molecular simulations, are limited by their reliance on fixed kernels or basis sets which hinder their ability to efficiently conform to varied free energy landscapes. Further, user-specified parameters are in general non-intuitive, yet significantly affect the convergence rate and accuracy of the…
▽ More
Existing adaptive bias techniques, which seek to estimate free energies and physical properties from molecular simulations, are limited by their reliance on fixed kernels or basis sets which hinder their ability to efficiently conform to varied free energy landscapes. Further, user-specified parameters are in general non-intuitive, yet significantly affect the convergence rate and accuracy of the free energy estimate. Here we propose a novel method wherein artificial neural networks (ANNs) are used to develop an adaptive biasing potential which learns free energy landscapes. We demonstrate that this method is capable of rapidly adapting to complex free energy landscapes and is not prone to boundary or oscillation problems. The method is made robust to hyperparameters and overfitting through Bayesian regularization which penalizes network weights and auto-regulates the number of effective parameters in the network. ANN sampling represents a promising innovative approach which can resolve complex free energy landscapes in less time than conventional approaches while requiring minimal user input.
△ Less
Submitted 7 December, 2017;
originally announced December 2017.
-
Orientationally Glassy Crystals of Janus Spheres
Authors:
Shan Jiang,
Jing Yan,
Jonathan K. Whitmer,
Stephen M. Anthony,
Erik Luijten,
Steve Granick
Abstract:
Colloidal Janus spheres in water (one hemisphere attractive and the other repulsive) assemble into two-dimensional hexagonal crystals with orientational order controlled by anisotropic interactions. We exploit the decoupled translational and rotational order to quantify the orientational dynamics. Via imaging experiments and Monte Carlo simulations we demonstrate that the correlations in the orien…
▽ More
Colloidal Janus spheres in water (one hemisphere attractive and the other repulsive) assemble into two-dimensional hexagonal crystals with orientational order controlled by anisotropic interactions. We exploit the decoupled translational and rotational order to quantify the orientational dynamics. Via imaging experiments and Monte Carlo simulations we demonstrate that the correlations in the orientation of individual Janus spheres exhibit glasslike dynamics that can be controlled via the ionic strength. Thus, these colloidal building blocks provide a particularly suitable model glass system for elucidating nontrivial dynamics arising from directional interactions, not captured by the consideration of just translational order.
△ Less
Submitted 10 June, 2014;
originally announced June 2014.
