“Learn on the Fly”: A Hybrid Classical and Quantum-Mechanical Molecular Dynamics Simulation

Csänyi, Gábor; Albaret, Tristan; Payne, Michael; Vita, Alessandro De

doi:10.1103/physrevlett.93.175503

Cited by 297 publications

(264 citation statements)

References 30 publications

Supporting

Mentioning

264

Contrasting

Order By: Relevance

“…[91] Recent advances in statistical data analysis methods [92][93][94][95] and applications in other areas of science and engineering, such as searching the internet, automated locomotion (self-driving cars), algorithmic trading, or brain-computer interfaces, strongly suggest that they will also play an increasingly important role in quantum chemistry. Examples of first efforts to quantitatively infer laws for atomistic simulations include ' 'Learning On The Fly,' ' [96] or ' 'force-matching.' ' [97,98] More sophisticated statistical learning methods have been applied to the training of exchange correlation functionals in DFT, [99,100] or to parameterizing interatomic force fields.…”

Section: Methodsmentioning

confidence: 99%

First principles view on chemical compound space: Gaining rigorous atomistic control of molecular properties

Lilienfeld

2013

Int J of Quantum Chemistry

138

160

View full text Add to dashboard Cite

A well‐defined notion of chemical compound space (CCS) is essential for gaining rigorous control of properties through variation of elemental composition and atomic configurations. Here, we give an introduction to an atomistic first principles perspective on CCS. First, CCS is discussed in terms of variational nuclear charges in the context of conceptual density functional and molecular grand‐canonical ensemble theory. Thereafter, we revisit the notion of compound pairs, related to each other via “alchemical” interpolations involving fractional nuclear charges in the electronic Hamiltonian. We address Taylor expansions in CCS, property nonlinearity, improved predictions using reference compound pairs, and the ounce‐of‐gold prize challenge to linearize CCS. Finally, we turn to machine learning of analytical structure property relationships in CCS. These relationships correspond to inferred, rather than derived through variational principle, solutions of the electronic Schrödinger equation. © 2013 Wiley Periodicals, Inc.

show abstract

Section: Methodsmentioning

confidence: 99%

First principles view on chemical compound space: Gaining rigorous atomistic control of molecular properties

Lilienfeld

2013

Int J of Quantum Chemistry

138

160

View full text Add to dashboard Cite

show abstract

“…Since we are eventually aiming at molecular systems, where collective motions are crucial, we decided to stick to the MD approach. Existing hybrid MD methods that concurrently couple different length scales have been developed to study solid state systems, where atomistic MD was either combined with the finite elements method [26,27,28] or it was linked to a quantum mechanical model [29]. To our knowledge, however, in pursuit of this objective no adaptive hybrid atomistic/mesoscale particlebased MD method, which would allow to dynamically adjust the level of detail, which means the adjustment of the degrees of freedom in the system, has been developed so far.…”

Section: Introductionmentioning

confidence: 99%

Adaptive resolution molecular-dynamics simulation: Changing the degrees of freedom on the fly

Praprotnik

Site

Kremer

2005

The Journal of Chemical Physics

371

514

View full text Add to dashboard Cite

We present a new adaptive resolution technique for efficient particle-based multiscale molecular dynamics (MD) simulations. The presented approach is tailor-made for molecular systems where atomistic resolution is required only in spatially localized domains whereas a lower mesoscopic level of detail is sufficient for the rest of the system. Our method allows an on-the-fly interchange between a given molecule's atomic and coarse-grained level of description, enabling us to reach large length and time scales while spatially retaining atomistic details of the system. The new approach is tested on a model system of a liquid of tetrahedral molecules. The simulation box is divided into two regions: one containing only atomistically resolved tetrahedral molecules, the other containing only one particle coarse-grained spherical molecules. The molecules can freely move between the two regions while changing their level of resolution accordingly. The coarse-grained and the atomistically resolved systems have the same statistical properties at the same physical conditions.

show abstract

“…The transition region is well defined by the here introduced generalization of the equipartition theorem for fractional dimension of phase space. While it directly applies to a scheme recently tested by the authors [19,20] it in the same way should also provide the general theoretical framework to extend other commonly used schemes, such as [12,16,17] towards a truly adaptive multiscale simulation scheme.…”

mentioning

confidence: 99%

“…Their efficiency and scope increases significantly if two or more such different approaches are combined into hybrid multiscale schemes. This is the case for the quantum based QM/MM ap-proach [4] and that of dual scale resolution techniques [5,6,7,8,9,10,11,12,13,14,15,16,17,18] aiming at bridging the atomistic and mesoscale length scale. However, the common feature and limitation of all these methods is the fact that the regions or parts of the system treated at different level of resolution are fixed and do not allow for free exchange.…”

mentioning

confidence: 99%

Adaptive molecular resolution via a continuous change of the phase space dimensionality

2007

View full text Add to dashboard Cite

For the study of complex synthetic and biological molecular systems by computer simulations one is still restricted to simple model systems or to by far too small time scales. To overcome this problem multiscale techniques are being developed for many applications. However in almost all cases, the regions of different resolution are fixed and not in a true equilibrium with each other. We here give the theoretical framework for an efficient and flexible coupling of the different regimes. The approach leads to an analog of a geometry induced phase transition and a counterpart of the equipartition theorem for fractional degrees of freedom. This provides a rather general formal basis for advanced computer simulation methods applying different levels of resolution. A long standing and often most challenging problem in condensed matter physics, up to some simple crystalline materials, is to understand the microscopic origin of macroscopic properties. While in certain cases a microscopic scale can be clearly separated from a macroscopic one, this is not the case for most experimental systems -details of the local interaction and generic/universal aspects are closely related. Already the proper determination of the fracture energy and crack propagation in crystalline materials requires a hierarchical and interrelated description, which links the breaking of the interatomic bonding in the fracture region to the response of the rest of the system on a micron scale [1]. The presence of microscopic chemical impurities in many metals and alloys changes their macroscopic mechanical behavior [2,3]. Even more complicated are synthetic and biological soft matter systems. Whether one is dealing with the morphology, the glass transition of a polymeric system, the function of a molecular assembly, e.g. for electronic applications or studies ligand-protein recognition or protein-protein interaction, in all cases the generic soft matter properties, such as matrix or chain conformation fluctuations, and details of the local chemistry apply to roughly the same length scales. For all these problems, which due to their complexity are heavily studied by computer simulations, there is a common underlying physical scenario: the number of degrees of freedom (DOFs) involved is very large and the exhaustive exploration of the related phase space is prohibitive. For many questions, however, such a deep level of detail in the description is only required locally.Theoretical methods employed to study these systems span from quantum-mechanical to macroscopic statistical approaches. Their efficiency and scope increases significantly if two or more such different approaches are combined into hybrid multiscale schemes. This is the case for the quantum based QM/MM ap- * On leave from the National

show abstract

“Learn on the Fly”: A Hybrid Classical and Quantum-Mechanical Molecular Dynamics Simulation

Cited by 297 publications

References 30 publications

First principles view on chemical compound space: Gaining rigorous atomistic control of molecular properties

First principles view on chemical compound space: Gaining rigorous atomistic control of molecular properties

Adaptive resolution molecular-dynamics simulation: Changing the degrees of freedom on the fly

Adaptive molecular resolution via a continuous change of the phase space dimensionality

Contact Info

Product

Resources

About