arXiv subjects

Gregory A. Voth

Publications and source records attributed to Gregory A. Voth.

At least 19 recordsLinked to original sources

arXiv2026

Next Generation of Ultra-Coarse-Graining: Self-Consistent Inference of Critical Internal States

Bottom-up coarse-graining expands the length and time scales accessible to molecular dynamics (MD) simulations, but information loss can hinder accurate representation of multistate phenomena in complex biomolecular dynamics. Ultra-Coarse-Graining (UCG) projects discrete "quantum-like" extended degrees of freedom, or "internal states," onto coarse-grained (CG) molecules, extending CG model expressiveness. The rapid-local-equilibrium (RLE) approximation in UCG depends on user-defined collective variables (CVs, e.g., local density) and neglects correlations between internal states within and between CG molecules. We present Self-Consistent UCG (SC-UCG), which uses the underlying UCG interactions directly to assign internal states without designing CVs in the CG ensemble. During simulation, internal state probabilities are determined self-consistently through graph message passing. We enhance the RLE Hamiltonian with the Bethe approximation and AI-based inference to represent explicit correlations between UCG beads. For force-field training, we develop Multilayer Internal State Consistency (MISC), a machine-learning method derived from relative entropy minimization that avoids iterative sampling of intermediate force fields. We apply SC-UCG to a tetramer exhibiting a second-order symmetry-breaking phase transition from a supercritical racemic fluid to subcritical D-rich and L-rich fluids. SC-UCG captures collective switching of internal states in the subcritical region and recapitulates the phase transition across temperatures, despite being trained on a single-temperature dataset.

physics.chem-ph↗

arXiv2026

A Bottom-Up Field-Theoretic Framework via Hierarchical Coarse-Graining: Generalized Mode Theory

Multiscale simulations facilitate the efficient exploration of large spatiotemporal scales in chemical and physical systems, yet particle-based simulations become prohibitively expensive at time and length scales beyond the molecular level. Field-theoretic simulations offer an attractive alternative, but most existing formulations rely on top-down approximations and are not systematically connected to atomistic interactions. Here, we present a hierarchical bottom-up framework for constructing auxiliary field representations of molecular liquids directly from microscopic models. We introduce a hierarchical coarse-graining framework that constructs field-theoretic models directly from atomistic liquids. The method first maps atomistic interactions to coarse-grained center-of-mass potentials and regularizes short-range divergences through a perturbative expansion in reciprocal space. Building on the auxiliary field formulation developed in polymer field-theoretic simulations, we then generalize the Hubbard-Stratonovich transformation to arbitrary pair potentials by separating positive and negative Fourier modes and introducing two auxiliary fields. The resulting generalized mode theory extends bottom-up field-theoretic modeling beyond positive-definite kernels and is compatible with existing field-theoretic sampling strategies. By combining formal derivations with numerical regularization and mode-truncation procedures, this work provides the theoretical foundation for scalable, bottom-up field-theoretic simulations of molecular systems.

physics.chem-ph↗

arXiv2025

Adversarial Training for Dynamics Matching in Coarse-Grained Models

Molecular dynamics (MD) simulations are essential for studying complex molecular systems, but their high computational cost limits scalability. Coarse-grained (CG) models reduce this cost by simplifying the system, yet traditional approaches often fail to maintain dynamic consistency, compromising their reliability in kinetics-driven processes. Here, we introduce an adversarial training framework that aligns CG trajectory ensembles with all-atom (AA) reference dynamics, ensuring both thermodynamic and kinetic fidelity. Our method adapts the adversarial learning paradigm, combining a physics-based generator with a neural network discriminator that differentiates between AA and CG trajectories. By adversarially optimizing CG parameters, our approach eliminates the need for predefined kinetic features. Applied to liquid water, it accurately reproduces radial and angular distribution functions as well as dynamical mean squared displacement, even extrapolating long-timescale dynamics from short training trajectories. This framework offers a new approach for bottom-up CG modeling, offering a systematic and principled way to preserve dynamic consistency in complex coarse-grained molecular systems.

physics.chem-ph↗

arXiv2025

Coarse-Graining in Quantum Mechanics: Distinguishable and Indistinguishable Particles

Bottom-up coarse-grained (CG) modeling expands the spatial and temporal scales of molecular simulation by seeking a reduced, thermodynamically consistent representation of an atomistic model. Developments in CG theory have largely focused on CG modeling of atomistic systems which behave classically, while CG modeling of quantum systems has remained largely unexplored. We present in this work two fundamental advances in particle-based, bottom-up CG theory for systems obeying quantum statistical mechanics. We first expand the bottom-up CG formalism to include indistinguishable quantum particles, including bosons and fermions. We next introduce a variational optimization procedure for CG model parameterization which is founded on the relative entropy minimization (REM) principle and then bridge the classical and quantum REM methods through a semiclassical expansion in terms of the Feynman path centroid. We provide numerical examples of REM CG models of distinguishable and indistinguishable quantum systems, including as examples a harmonically trapped bosonic system and liquid water. The theoretical results presented here constitute a means to accelerate simulating thermal quantum systems ranging from distinguishable particle systems at higher temperatures to quantum indistinguishable particle systems at lower temperatures.

physics.chem-ph↗

arXiv2025

Understanding Dynamics in Coarse-Grained Models: V. Extension of Coarse-Grained Dynamics Theory to Non-Hard Sphere Systems

Coarse-grained (CG) modeling has gained significant attention in recent years due to its wide applicability in enhancing the spatiotemporal scales of molecular simulations. While CG simulations, often performed with Hamiltonian mechanics, faithfully recapitulate structural correlations at equilibrium, they lead to ambiguously accelerated dynamics. In the first paper of this series [J. Chem. Phys. 158, 034103 (2023)], we proposed the excess entropy scaling relationship to understand the CG dynamics. Then, in the following companion paper [J. Chem. Phys. 158, 034104 (2023)], we developed a theory to map the CG system into a dynamically-consistent hard sphere system to analytically derive an expression for fast CG dynamics. However, many chemical and physical systems do not exhibit hard sphere-like behavior, limiting the extensibility of the developed theory. In this paper, we aim to generalize the theory to the non-hard sphere system based on the Weeks-Chandler-Andersen perturbation theory. Since non-hard sphere-like CG interactions affect the excess entropy term as it deviates from the hard sphere description, we explicitly account for the extra entropy to correct the non-hard sphere nature of the system. This approach is demonstrated for two different types of interactions seen in liquids, and we further provide a generalized description for any CG models using the generalized Gaussian CG models using Gaussian basis sets. Altogether, this work allows for extending the range and applicability of the hard sphere CG dynamics theory to a myriad of CG liquids.

physics.chem-ph↗

arXiv2024

PLUMED Tutorials: a collaborative, community-driven learning ecosystem

In computational physics, chemistry, and biology, the implementation of new techniques in a shared and open source software lowers barriers to entry and promotes rapid scientific progress. However, effectively training new software users presents several challenges. Common methods like direct knowledge transfer and in-person workshops are limited in reach and comprehensiveness. Furthermore, while the COVID-19 pandemic highlighted the benefits of online training, traditional online tutorials can quickly become outdated and may not cover all the software's functionalities. To address these issues, here we introduce ``PLUMED Tutorials'', a collaborative model for developing, sharing, and updating online tutorials. This initiative utilizes repository management and continuous integration to ensure compatibility with software updates. Moreover, the tutorials are interconnected to form a structured learning path and are enriched with automatic annotations to provide broader context. This paper illustrates the development, features, and advantages of PLUMED Tutorials, aiming to foster an open community for creating and sharing educational resources.

physics.ed-ph↗

arXiv2024

Understanding Dynamics in Coarse-Grained Models: IV. Connection of Fine-Grained and Coarse-Grained Dynamics with the Stokes-Einstein and Stokes-Einstein-Debye Relations

Applying an excess entropy scaling formalism to the coarse-grained (CG) dynamics of liquids, we discovered that missing rotational motions during the CG process are responsible for artificially accelerated CG dynamics. In the context of the dynamic representability between the fine-grained (FG) and CG dynamics, this work introduces the well-known Stokes-Einstein and Stokes-Einstein-Debye relations to unravel the rotational dynamics underlying FG trajectories, thereby allowing for an indirect evaluation of the effective rotations based only on the translational information at the reduced CG resolution. Since the representability issue in CG modeling limits a direct evaluation of the shear stress appearing in the Stokes-Einstein and Stokes-Einstein-Debye relations, we introduce a translational relaxation time as a proxy to employ these relations, and we demonstrate that these relations hold for the ambient conditions studied in our series of work. Additional theoretical links to our previous work are also established. First, we demonstrate that the effective hard sphere radius determined by the classical perturbation theory can approximate the complex hydrodynamic radius value reasonably well. Also, we present a simple derivation of an excess entropy scaling relationship for viscosity by estimating the elliptical integral of molecules. In turn, since the translational and rotational motions at the FG level are correlated to each other, we conclude that the "entropy-free" CG diffusion only depends on the shape of the reference molecule. Our results and analyses impart an alternative way of recovering the FG diffusion from the CG description by coupling the translational and rotational motions at the hydrodynamic level.

physics.chem-ph↗

arXiv2023

Coarse-Graining with Equivariant Neural Networks: A Path Towards Accurate and Data-Efficient Models

Machine learning has recently entered into the mainstream of coarse-grained (CG) molecular modeling and simulation. While a variety of methods for incorporating deep learning into these models exist, many of them involve training neural networks to act directly as the CG force field. This has several benefits, the most significant of which is accuracy. Neural networks can inherently incorporate multi-body effects during the calculation of CG forces, and a well-trained neural network force field outperforms pairwise basis sets generated from essentially any methodology. However, this comes at a significant cost. First, these models are typically slower than pairwise force fields even when accounting for specialized hardware which accelerates the training and integration of such networks. The second, and the focus of this paper, is the need for the considerable amount of data needed to train such force fields. It is common to use 10s of microseconds of molecular dynamics data to train a single CG model, which approaches the point of eliminating the CG models usefulness in the first place. As we investigate in this work, this data-hunger trap from neural networks for predicting molecular energies and forces can be remediated in part by incorporating equivariant convolutional operations. We demonstrate that for CG water, networks which incorporate equivariant convolutional operations can produce functional models using datasets as small as a single frame of reference data, while networks without these operations cannot.

physics.chem-ph↗

arXiv2023

Understanding Dynamics in Coarse-Grained Models: III. Roles of Rotational Motion and Translation-Rotation Coupling in Coarse-Grained Dynamics

This paper series aims to establish a complete correspondence between fine-grained (FG) and coarse-grained (CG) dynamics by way of excess entropy scaling (introduced in Paper I). While Paper II successfully captured translational motions in CG systems using a hard sphere mapping, the absence of rotational motions in single-site CG models introduces differences between FG and CG dynamics. In this third paper, our objective is to faithfully recover atomistic diffusion coefficients from CG dynamics by incorporating rotational dynamics. By extracting FG rotational diffusion, we unravel, for the first time reported to our knowledge, a universality in excess entropy scaling between the rotational and translational diffusion. Once the missing rotational dynamics are integrated into the CG translational dynamics, an effective translation-rotation coupling becomes essential. We propose two different approaches for estimating this coupling parameter: the rough hard sphere theory with acentric factor (temperature-independent) or the rough Lennard-Jones model with CG attractions (temperature-dependent). Altogether, we demonstrate that FG diffusion coefficients can be recovered from CG diffusion coefficients by (1) incorporating "entropy-free" rotational diffusion with translation-rotation coupling and (2) recapturing the missing entropy. Our findings shed light on the fundamental relationship between FG and CG dynamics in molecular fluids.

physics.chem-ph↗

arXiv2023

Understanding Dynamics in Coarse-Grained Models: I. Universal Excess Entropy Scaling Relationship

Coarse-grained (CG) models facilitate an efficient exploration of complex systems by reducing the unnecessary degrees of freedom of the fine-grained (FG) system while recapitulating major structural correlations. Unlike structural properties, assessing dynamic properties in CG modeling is often unfeasible due to the accelerated dynamics of the CG models, which allows for more efficient structural sampling. Therefore, the ultimate goal of the present series of articles is to establish a better correspondence between the FG and CG dynamics. To assess and compare dynamical properties in the FG and the corresponding CG models, we utilize the excess entropy scaling relationship. For Paper I of this series, we provide evidence that the FG and the corresponding CG counterpart follow the same universal scaling relationship. By carefully reviewing and examining the literature, we develop a new theory to calculate excess entropies for the FG and CG systems while accounting for entropy representability. We demonstrate that the excess entropy scaling idea can be readily applied to liquid water and methanol systems at both the FG and CG resolutions. For both liquids, we reveal that the scaling exponents remain unchanged from the coarse-graining process, indicating that the scaling behavior is universal for the same underlying molecular systems. Combining this finding with the concept of mapping entropy in CG models, we show that the missing entropy plays an important role in accelerating the CG dynamics.

physics.chem-ph↗

arXiv2023

Understanding Dynamics in Coarse-Grained Models: II. Coarse-Grained Diffusion Modeled Using Hard Sphere Theory

The first paper of this series [J. Chem. Phys. 158, 034103 (2023)] demonstrated that excess entropy scaling holds for both fine-grained and corresponding coarse-grained (CG) systems. Despite its universality, a more exact determination of the scaling relationship was not possible due to the semi-empirical nature. In this second paper, an analytical excess entropy scaling relation is derived for bottom-up CG systems. At the single-site CG resolution, effective hard sphere systems are constructed that yield near-identical dynamical properties as the target CG systems by taking advantage of how hard sphere dynamics and excess entropy can be analytically expressed in terms of the liquid packing fraction. Inspired by classical equilibrium perturbation theories and recent advances in constructing hard sphere models for predicting activated dynamics of supercooled liquids, we propose a new approach for understanding the diffusion of molecular liquids in the normal regime using hard sphere reference fluids. The proposed "fluctuation matching" is designed to have the same amplitude of long wavelength density fluctuations (dimensionless compressibility) as the CG system. Utilizing the Enskog theory to derive an expression for hard sphere diffusion coefficients, a bridge between the CG dynamics and excess entropy is then established. The CG diffusion coefficient can be roughly estimated using various equations of the state, and an accurate prediction of accelerated CG dynamics at different temperatures is also possible in advance of running any CG simulation. By introducing another layer of coarsening, these findings provide a more rigorous method to assess excess entropy scaling and understand the accelerated CG dynamics of molecular fluids.

physics.chem-ph↗

arXiv2022

Utilizing Machine Learning to Greatly Expand the Range and Accuracy of Bottom-Up Coarse-Grained Models Through Virtual Particles

Coarse-grained (CG) models parameterized using atomistic reference data, i.e., 'bottom up' CG models, have proven useful in the study of biomolecules and other soft matter. However, the construction of highly accurate, low resolution CG models of biomolecules remains challenging. We demonstrate in this work how virtual particles, CG sites with no atomistic correspondence, can be incorporated into CG models within the context of relative entropy minimization (REM) as latent variables. The methodology presented, variational derivative relative entropy minimization (VD-REM), enables optimization of virtual particle interactions through a gradient descent algorithm aided by machine learning. We apply this methodology to the challenging case of a solvent-free CG model of a 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) lipid bilayer and demonstrate that introduction of virtual particles captures solvent-mediated behavior and higher-order correlations which REM alone cannot capture in a more standard CG model based only on the mapping of collections of atoms to the CG sites.

physics.chem-ph↗

arXiv2022

Centroid Molecular Dynamics Can Be Greatly Accelerated Through Neural Network Learned Centroid Forces Derived from Path Integral Molecular Dynamics

For nearly the past 30 years, Centroid Molecular Dynamics (CMD) has proven to be a viable classical-like phase space formulation for the calculation of quantum dynamical properties. However, calculation of the centroid effective force remains a significant computational cost and limits the ability of CMD to be an efficient approach to study condensed phase quantum dynamics. In this paper we introduce a neural network-based methodology for first learning the centroid effective force from path integral molecular dynamics data, which is subsequently used as an effective force field to evolve the centroids directly with the CMD algorithm. This method, called Machine-Learned Centroid Molecular Dynamics (ML-CMD) is faster and far less costly than both standard on the fly CMD and ring polymer molecular dynamics (RPMD). The training aspect of ML-CMD is also straightforwardly implemented utilizing the DeepMD software kit. ML-CMD is then applied to two model systems to illustrate the approach: liquid para-hydrogen and water. The results show comparable accuracy to both CMD and RPMD in the estimation of quantum dynamical properties, including the self-diffusion constant and velocity time correlation function, but for significantly reduced overall computational cost.

physics.chem-ph↗

arXiv2021

Explaining classifiers to understand coarse-grained models

Bottom-up coarse-grained molecular dynamics models are parameterized using complex effective Hamiltonians. These models are typically optimized to approximate high dimensional data from atomistic simulations. In contrast, human validation of these models is often limited to low dimensional statistics that do not necessarily differentiate between the CG model and said atomistic simulations. We propose that explainable machine learning can directly convey high-dimensional error to scientists and use Shapley additive explanations do so in two coarse-grained protein models.

physics.chem-ph↗

arXiv2021

Understanding the Essential Nature of the Hydrated Excess Proton Through Simulation and Interpretation of Recent Spectroscopic Experiments

Two-dimensional infrared spectroscopy experiments have presented new results regarding the dynamics of the hydrated excess proton (aka hydronium cation solvated in water). It has been suggested by these experiments that the hydrated excess proton has an anisotropy reorientation timescale of 2.5 ps, which can be viewed as being somewhat long lived. Through the use of both the reactive molecular dynamics Multistate-Empirical Valence Bond method and Experiment Directed Simulation Ab Initio Molecular Dynamics we show that timescales of the same magnitude are obtained that correspond to proton transport, while also involving structural reorientations of the hydrated proton structure that correspond to the so-called special pair dance . The latter is a process predicted by prior computational studies in which the central hydrated hydronium in a distorted Eigen cation (H 9 O 4 + ) structure continually switches special pair partners with its strongly hydrogen-bonded neighboring water molecules. These dynamics are further characterized through the time-evolution of instantaneous normal modes. It is concluded that the hydrated excess proton has a spectral signature unique from the other protons in the hydrated proton complex. However, the results call into question the use of a static picture based on a simple effective one dimensional potential well to describe the hydrated excess proton in water. Instead, they more conclusively point to a distorted and dynamic Eigen cation as the most prevalent hydrated proton species in acid solutions of dilute to moderate concentrations.

physics.chem-ph↗

arXiv2020

The Hopping Mechanism of the Hydrated Excess Proton and Its Contribution to Proton Diffusion in Water

In this work a series of analyses are performed on ab initio molecular dynamics (AIMD) simulations of a hydrated excess proton in water to quantify the relative occurrence of concerted hopping events and rattling events, and thus to further elucidate the hopping mechanism of proton transport in water. Contrary to results reported in certain earlier papers, the new analysis finds that concerted hopping events do occur in all simulations, but that the majority of events are the product of proton rattling, where the excess proton will rattle between two or more waters. The results are consistent with the proposed special-pair dance model of the hydrated excess proton, wherein the acceptor water molecule for the proton transfer will quickly change (resonate between three equivalent special pairs) until a decisive proton hop occurs. To remove the misleading effect of simple rattling, a filter was applied to the trajectory such that hopping events that were followed by back hops to the original water are not counted. A steep reduction in the number of multiple hopping events is found when the filter is applied, suggesting that many multiple hopping events that occur in the unfiltered trajectory are largely the product of rattling, contrary to prior suggestions. Comparing the continuous correlation function of the filtered and unfiltered trajectories, we find agreement with experimental values for the proton hopping time and Eigen-Zundel interconversion time, respectively.

physics.chem-ph↗

arXiv2020

Density Functional Theory-based Quantum Mechanics/Coarse-grained Molecular Mechanics: Theory and Implementation

Quantum mechanics/molecular mechanics (QM/MM) is a standard computational tool for describing chemical reactivity in systems with many degrees of freedom, including polymers, enzymes, and reacting molecules in complex solvents. However, QM/MM is less suitable for systems with complex MM dynamics due to associated long relaxation times, the high computational cost of QM energy evaluations, and expensive long-range electrostatics. Recently, a systematic coarse-graining of the MM part was proposed to overcome these QM/MM limitations in the form of the quantum mechanics/coarse-grained molecular mechanics (QM/CG-MM) approach. Herein, we recast QM/CG-MM in the density functional theory formalism and, by employing the force-matching variational principle, access the method performance for two model systems: QM CCl4 in the MM CCl4 liquid and the reaction of tert-butyl hypochlorite with the benzyl radical in the MM CCl4 solvent. We find that DFT-QM/CG-MM accurately reproduces DFT-QM/MM radial distribution functions and 3-body correlations between QM and CG-MM subsystems. The free energy profile of the reaction is also described well, with an error < 1-2 kcal/mol. DFT-QM/CG-MM is a general, systematic, and computationally efficient approach to include chemical reactivity in coarse-grained molecular models.

physics.chem-ph↗

arXiv2020

Water Assisted Proton Transport in Confined Nanochannels

Hydrated excess protons under hydrophobic confinement are a critical component of charge transport behavior and reactivity in nanoporous materials and biomolecular systems. Herein excess proton confinement effects are computationally investigated for sub-2 nm hydrophobic nanopores by varying the diameters (d = 0.81, 0.95, 1.09, 1.22, 1.36, 1.63, and 1.90 nm), lengths (l ~3 and ~5 nm), curvature, and chirality of cylindrical carbon nanotube (CNT) nanopores. CNTs with a combination of different diameter segments are also explored. The spatial distribution of water molecules under confinement are diameter-dependent; however, proton solvation and transport is consistently found to occur in the water layer adjacent to the pore wall, showing an "amphiphilic" character of the hydrated excess proton hydronium-like structure. The proton transport free energy barrier also decreases significantly as the nanopore diameter increases and proton transport becomes almost barrierless in the d > 1 nm nanopores. Among the nanopores studied, the Zundel cation (${H_{5}O_{2}}^{+}$) is populated only in the d = 0.95 nm CNT (7,7) nanopore. The presence of the hydrated excess proton and $K^{+}$ inside the CNT (7,7) nanopore induces a water density increase by 40 and 20%, respectively. The $K^{+}$ transport through CNT nanopores is also consistently higher in free energy barrier than proton transport. Interestingly, the evolution of excess protonic charge defect distribution reveals a "frozen" single water wire configuration in the d = 0.81 nm CNT (6,6) nanopore (or segment), through which hydrated excess protons can only shuttle via the Grotthuss mechanism. Vehicular diffusion becomes relevant to proton transport in the "flat" free energy regions and in the wider nanopores, where protons do not primarily shuttle in the axial direction.

cond-mat.soft↗