Neuroevolution machine learning potentials: Combining high accuracy and low cost in atomistic simulations and application to heat transport

Fan, Zheyong; Zeng, Zhaoliang; Zhang, Cunzhi; Wang, Yanzhou; Song, Keke; Dong, Hai-Kuan; Chen, Yue; Ala-Nissilä, Tapio

doi:10.1103/physrevb.104.104309

Cited by 131 publications

(136 citation statements)

References 90 publications

Supporting

Mentioning

131

Contrasting

Order By: Relevance

“…In summary, we have presented the various features of the open-source gpumd package, with a focus on recent developments that have enabled the generation and use of accurate and efficient NEP MLPs. 41,42 Two improvements on the atomic-environment descriptor have been introduced: one is to change the radial functions from Chebyshev basis functions to linear combinations of the basis functions, and the other is to extend the angular descriptor components by considering some 4-body and 5-body contributions as in the ACE approach. 46 Both of these extensions are shown to improve the accuracy of NEP models further.…”

Section: Discussionmentioning

confidence: 99%

GPUMD: A package for constructing accurate machine-learned potentials and performing highly efficient atomistic simulations

Fan,

Wang,

Ying

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

We present our latest advancements of machine-learned potentials (MLPs) based on the neuroevolution potential (NEP) framework introduced in [Fan et al., Phys. Rev. B 104, 104309 (2021)] and their implementation in the open-source package gpumd. We increase the accuracy of NEP models both by improving the radial functions in the atomic-environment descriptor using a linear combination of Chebyshev basis functions and by extending the angular descriptor with some four-body and five-body contributions as in the atomic cluster expansion approach. We also detail our efficient implementation of the NEP approach in graphics processing units as well as our workflow for the construction of NEP models, and we demonstrate their application in large-scale atomistic simulations. By comparing to state-of-the-art MLPs, we show that the NEP approach not only achieves above-average accuracy but also is far more computationally efficient. These results demonstrate that the gpumd package is a promising tool for solving challenging problems requiring highly accurate, large-scale atomistic simulations. To enable the construction of MLPs using a minimal training set, we propose an active-learning scheme based on the latent space of a pre-trained NEP model. Finally, we introduce two separate Python packages, gpyumd and calorine, which enable the integration of gpumd into Python workflows.

show abstract

Section: Discussionmentioning

confidence: 99%

GPUMD: A package for constructing accurate machine-learned potentials and performing highly efficient atomistic simulations

Fan,

Wang,

Ying

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…However, thermal transport usually involves large length and long time scales and GAP18 is not currently efficient enough for this purpose. The NEP model as implemented in the gpumd package [27,28], on the other hand, can reach a computational speed of about 5 × 10 6 atom-step per second for a-Si by using a single GPU card such as Tesla V100, which is about three orders of magnitude faster than GAP18 using 72 Xeon-Gold 6240 central processing unit (CPU) cores [18].…”

Section: Training a Machine-learned Potential For A-simentioning

confidence: 99%

“…A NEP model has already been trained previously [18] (we call it NEP21), but we here re-train it by changing the relative weight of virial from 1 to 0.1 and keeping all the other hyperparameters as used in Ref. 18 unchanged.…”

Section: Training a Machine-learned Potential For A-simentioning

confidence: 99%

“…to create a nonzero heat current. Here F e is the driving force parameter with the dimension of inverse length and W i is the 3 × 3 per-atom virial tensor (not necessarily symmetric for many-body potentials) [18,38,39] defined as…”

Section: Heat Transport In Amorphous Siliconmentioning

confidence: 99%

“…where T is the temperature and V is the volume. Here the instant heat current is calculated based on the definition [18,38,39]…”

Section: Heat Transport In Amorphous Siliconmentioning

confidence: 99%

See 2 more Smart Citations

Quantum-corrected thickness-dependent thermal conductivity in amorphous silicon predicted by machine-learning molecular dynamics simulations

Wang¹,

Fan²,

Qian³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Amorphous silicon (a-Si) is an important thermal-management material and also serves as an ideal playground for studying heat transport in strongly disordered materials. Theoretical prediction of the thermal conductivity of a-Si in a wide range of temperatures and sample sizes is still a challenge. Herein we present a systematic investigation of the thermal transport properties of a-Si by employing large-scale molecular dynamics (MD) simulations with an accurate and efficient machine-learned neuroevolution potential (NEP) trained against abundant reference data calculated at the quantummechanical density-functional-theory level. The high efficiency of NEP allows us to study the effects of finite size and quenching rate in the formation of a-Si in great detail. We find that it requires a simulation cell up to 64, 000 atoms (a cubic cell with a linear size of 11 nm) and a quenching rate down to 10 11 K s −1 for convergent thermal conductivity. Structural properties, including short-and medium-range order as characterized by the pair correlation function, angular distribution function, coordination number, ring statistics and structure factor are studied to demonstrate the accuracy of NEP and to further evaluate the role of quenching rate. Using both the heterogeneous and the homogeneous nonequilibrium MD methods and the related spectral decomposition techniques, we calculate the temperature-and thickness-dependent thermal conductivity values of a-Si and show that they agree well with available experimental results from 10 K to room temperature. Our results also highlight the importance of quantum effects in the calculated thermal conductivity and support the quantum correction method based on the spectral thermal conductivity.

show abstract

Decoding Electrochemical Processes of Lithium‐Ion Batteries by Classical Molecular Dynamics Simulations

Tan,

Chen,

Zhang

et al. 2024

Advanced Energy Materials

View full text Add to dashboard Cite

Lithium‐ion batteries (LIBs) have played an essential role in the energy storage industry and dominated the power sources for consumer electronics and electric vehicles. Understanding the electrochemistry of LIBs at the molecular scale is significant for improving their performance, stability, lifetime, and safety. Classical molecular dynamics (MD) simulations could directly capture the atomic and molecular motions and thus provide dynamic insights into the electrochemical processes and ion transport in LIBs during charging and discharging that are usually challenging to observe experimentally, which is momentous in developing LIBs with superb performance. This review discusses developments in MD approaches using non‐reactive force fields, reactive force fields, and machine learning potential for modeling chemical reactions and transport of reactants in the electrodes, electrolytes, and electrode‐electrolyte interfaces. It also comprehensively discusses how molecular interactions, structures, transport, and reaction processes affect electrode stability, energy capacity, and interfacial properties. Finally, the remaining challenges and envisioned future routes are commented on for high‐fidelity, effective simulation methods to decode the invisible interactions and chemical reactions in LIBs.

show abstract

Neuroevolution machine learning potentials: Combining high accuracy and low cost in atomistic simulations and application to heat transport

Cited by 131 publications

References 90 publications

GPUMD: A package for constructing accurate machine-learned potentials and performing highly efficient atomistic simulations

GPUMD: A package for constructing accurate machine-learned potentials and performing highly efficient atomistic simulations

Quantum-corrected thickness-dependent thermal conductivity in amorphous silicon predicted by machine-learning molecular dynamics simulations

Decoding Electrochemical Processes of Lithium‐Ion Batteries by Classical Molecular Dynamics Simulations

Contact Info

Product

Resources

About