Coupled Cluster Molecular Dynamics of Condensed Phase Systems Enabled by Machine Learning Potentials: Liquid Water Benchmark

Daru, János; Forbert, Harald; Behler, Jörg; Marx, Dominik

doi:10.1103/physrevlett.129.226001

Cited by 51 publications

(49 citation statements)

References 97 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…15−18 Other cluster-based models for water have gone on to explicitly include 4-body terms 20−22 and also train on larger water clusters. 23 MLPs fit to periodic electronic structure offer the opportunity to readily capture many-body electronic structure effects, since these are naturally included in the electronic structure calculation. Recent advances have opened the door to efficiently calculating the properties of periodic condensedphase systems using higher-level methods like coupled cluster singles and doubles without and with perturbative triples (CCSD and CCSD(T)) 28,29 and phaseless auxiliary-field quantum Monte Carlo (AFQMC).…”

Section: Introductionmentioning

confidence: 99%

Data-Efficient Machine Learning Potentials from Transfer Learning of Periodic Correlated Electronic Structure Methods: Liquid Water at AFQMC, CCSD, and CCSD(T) Accuracy

Chen

Lee

et al. 2023

J. Chem. Theory Comput.

View full text Add to dashboard Cite

Obtaining the atomistic structure and dynamics of disordered condensed-phase systems from first-principles remains one of the forefront challenges of chemical theory. Here we exploit recent advances in periodic electronic structure and provide a dataefficient approach to obtain machine-learned condensed-phase potential energy surfaces using AFQMC, CCSD, and CCSD(T) from a very small number (≤200) of energies by leveraging a transfer learning scheme starting from lower-tier electronic structure methods. We demonstrate the effectiveness of this approach for liquid water by performing both classical and path integral molecular dynamics simulations on these machine-learned potential energy surfaces. By doing this, we uncover the interplay of dynamical electron correlation and nuclear quantum effects across the entire liquid range of water while providing a general strategy for efficiently utilizing periodic correlated electronic structure methods to explore disordered condensed-phase systems.

show abstract

Section: Introductionmentioning

confidence: 99%

Data-Efficient Machine Learning Potentials from Transfer Learning of Periodic Correlated Electronic Structure Methods: Liquid Water at AFQMC, CCSD, and CCSD(T) Accuracy

Chen

Lee

et al. 2023

J. Chem. Theory Comput.

View full text Add to dashboard Cite

show abstract

“…So, it would be interesting to see if these NN-TL force fields produced the level of accuracy seen here or in MB-pol for the hexamer isomer energies. These were not reported in the two recent NN-TL force fields, 68,69 and so we are only making an educated speculation here.…”

Section: Discussionmentioning

confidence: 72%

“…Very recent reports using transfer-learning 67 have led to CCSD(T)-transfer-learned FFs for water. 68,69 These are both atom-centered NN fits, and in both cases were trained on DFT-based samples of condensed phase water consisting of 32 or 64 monomers. Several condensed phase quantities, i.e., radial distribution functions and diffusion constants, were reported using NVT simulations and shown to in very good agreement with experiment.…”

Section: Discussionmentioning

confidence: 99%

“…A similar, at least in spirit, NN-TL water potential was reported by Marx and coworkers. 68 That potential made use of long-range fixed-charged electrostatic interactions, that are damped as usual to avoid the Coulomb singularity, as well as short range repulsive Yukawa potentials. The charges are -0.8 for O and 0.4 for H, exactly the ones in the first generation (1981) fixed charged model, TIP.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Interfacing q-AQUA with a Polarizable Force Field: The Best of Both Worlds

Bowman

et al. 2023

Preprint

View full text Add to dashboard Cite

Polarizable force fields are pervasive in the fields of computational chemistry and biochemistry; however, their empirical or semi-empirical nature gives them both weaknesses and strengths. Here, we have developed a hybrid water potential, named q-AQUA-pol, by combining our recent ab initio q-AQUA potential with the TTM3-F water potential. The new potential demonstrates unprecedented accuracy ranging from gas-phase clusters, e.g., the eight low-lying isomers of the water hexamer, to the condensed phase, e.g., radial distribution functions, the self-diffusion coefficient, triplet OOO distribution, and the temperature dependence of the density. This represents a significant advancement in the field of polarizable machine learning potential and computational modeling.

show abstract

“…Moreover, to achieve that accuracy, a surprisingly small training data set was required which foreshadows that it might even be possible to train an APTNN on more expensive electronic structure calculations, such as hybrid DFT or even beyond. Machine learned MD simulations have already been performed using hybrid DFT 71 or even CCSD(T) 72 and allow one to sample nanoseconds of MD trajectories at that level of theory. Such high level MD trajectories can easily be postprocessed by the APTNN model to obtain well converged vibrational spectra, which are clearly computationally prohibitive otherwise.…”

mentioning

confidence: 99%

Spectroscopy from Machine Learning by Accurately Representing the Atomic Polar Tensor

Schienbein

2023

J. Chem. Theory Comput.

View full text Add to dashboard Cite

Vibrational spectroscopy is a key technique to elucidate microscopic structure and dynamics. Without the aid of theoretical approaches, it is, however, often difficult to understand such spectra at a microscopic level. Ab initio molecular dynamics has repeatedly proved to be suitable for this purpose; however, the computational cost can be daunting. Here, the E(3)-equivariant neural network is used to fit the atomic polar tensor of liquid water a posteriori on top of existing molecular dynamics simulations. Notably, the introduced methodology is general and thus transferable to any other system as well. The target property is most fundamental and gives access to the IR spectrum, and more importantly, it is a highly powerful tool to directly assign IR spectral features to nuclear motiona connection which has been pursued in the past but only using severe approximations due to the prohibitive computational cost. The herein introduced methodology overcomes this bottleneck. To benchmark the machine learning model, the IR spectrum of liquid water is calculated, indeed showing excellent agreement with the explicit reference calculation. In conclusion, the presented methodology gives a new route to calculate accurate IR spectra from molecular dynamics simulations and will facilitate the understanding of such spectra on a microscopic level.

show abstract

Coupled Cluster Molecular Dynamics of Condensed Phase Systems Enabled by Machine Learning Potentials: Liquid Water Benchmark

Cited by 51 publications

References 97 publications

Data-Efficient Machine Learning Potentials from Transfer Learning of Periodic Correlated Electronic Structure Methods: Liquid Water at AFQMC, CCSD, and CCSD(T) Accuracy

Data-Efficient Machine Learning Potentials from Transfer Learning of Periodic Correlated Electronic Structure Methods: Liquid Water at AFQMC, CCSD, and CCSD(T) Accuracy

Interfacing q-AQUA with a Polarizable Force Field: The Best of Both Worlds

Spectroscopy from Machine Learning by Accurately Representing the Atomic Polar Tensor

Contact Info

Product

Resources

About