Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
119 tokens/sec
GPT-4o
56 tokens/sec
Gemini 2.5 Pro Pro
43 tokens/sec
o3 Pro
6 tokens/sec
GPT-4.1 Pro
47 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Active learning of Boltzmann samplers and potential energies with quantum mechanical accuracy (2401.16487v2)

Published 29 Jan 2024 in physics.chem-ph, physics.comp-ph, and stat.ML

Abstract: Extracting consistent statistics between relevant free-energy minima of a molecular system is essential for physics, chemistry and biology. Molecular dynamics (MD) simulations can aid in this task but are computationally expensive, especially for systems that require quantum accuracy. To overcome this challenge, we develop an approach combining enhanced sampling with deep generative models and active learning of a machine learning potential (MLP). We introduce an adaptive Markov chain Monte Carlo framework that enables the training of one Normalizing Flow (NF) and one MLP per state, achieving rapid convergence towards the Boltzmann distribution. Leveraging the trained NF and MLP models, we compute thermodynamic observables such as free-energy differences or optical spectra. We apply this method to study the isomerization of an ultrasmall silver nanocluster, belonging to a set of systems with diverse applications in the fields of medicine and catalysis.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (64)
  1. Accelerated chemical science with ai. Digital Discovery, 3(1):23–33, 2024.
  2. Metals to combat antimicrobial resistance. Nature Reviews Chemistry, 7(3):202–224, February 2023.
  3. Iron oxide based nanoparticles for multimodal imaging and magnetoresponsive therapy. Chemical Reviews, 115(19):10637–10689, 2015. PMID: 26250431.
  4. An unexpected all-metal aromatic tetranuclear silver cluster in human copper chaperone atox1. Chemical Science, 13:7269–7275, 6 2022.
  5. Boron delivery agents for neutron capture therapy of cancer. Cancer Communications, 38(1):35, 2018.
  6. Michele Ceriotti. Beyond potentials: Integrated machine learning models for materials. MRS Bulletin, 47(10):1045–1053, October 2022.
  7. Christopher J Cramer. Essentials of computational chemistry. John Wiley & Sons, Chichester, England, 2 edition, September 2004.
  8. The conformational free energy landscape of β𝛽\betaitalic_β-d-glucopyranose. implications for substrate preactivation in β𝛽\betaitalic_β-glucoside hydrolases. Journal of the American Chemical Society, 129(35):10686–10693, 2007. PMID: 17696342.
  9. Ab initio molecular dynamics with enhanced sampling in heterogeneous catalysis. Catalysis Science & Technology, 12(1):12–37, 2022.
  10. Zooming in on the o–o bond formation—an ab initio molecular dynamics study applying enhanced sampling techniques. Journal of chemical theory and computation, 16(4):2436–2449, 2020.
  11. Accurate and efficient multilevel free energy simulations with neural network-assisted enhanced sampling. Journal of Chemical Theory and Computation, 19(16):5394–5406, 2023.
  12. Combining machine learning approaches and accurate ab initio enhanced sampling methods for prebiotic chemical reactions in solution. Journal of Chemical Theory and Computation, 18(9):5410–5421, 2022.
  13. Ab initio metadynamics determination of temperature-dependent free-energy landscape in ultrasmall silver clusters. The Journal of Chemical Physics, 156(15):154301, 04 2022.
  14. Jörg Behler. Four Generations of High-Dimensional Neural Network Potentials. Chemical Reviews, 121(16):10037–10072, August 2021. Publisher: American Chemical Society.
  15. Development of multimodal machine learning potentials: Toward a physics-aware artificial intelligence. Accounts of Chemical Research, 54(7):1575–1585, 2021.
  16. Ultra-fast interpretable machine-learning potentials. npj Computational Materials, 9(1):162, 2023.
  17. Machine learning potentials for complex aqueous systems made simple. Proceedings of the National Academy of Sciences, 118(38):e2110077118, 2021.
  18. A fourth-generation high-dimensional neural network potential with accurate electrostatics including non-local charge transfer. Nature communications, 12(1):398, 2021.
  19. Machine learning potentials for extended systems: a perspective. The European Physical Journal B, 94:1–11, 2021.
  20. Constructing custom thermodynamics using deep learning. Nature Computational Science, pages 1–20, 2023.
  21. Thermodynamics of diamond formation from hydrocarbon mixtures in planets. Nature Communications, 14(1):1104, February 2023. Number: 1 Publisher: Nature Publishing Group.
  22. Machine-learning-accelerated simulations to enable automatic surface reconstruction. Nature Computational Science, pages 1–11, 2023.
  23. Deep learning the slow modes for rare events sampling. Proceedings of the National Academy of Sciences, 118(44):e2113533118, 2021.
  24. Reweighted autoencoded variational bayes for enhanced sampling (rave). The Journal of chemical physics, 149(7), 2018.
  25. Machine learning approaches for analyzing and enhancing molecular dynamics simulations. Current opinion in structural biology, 61:139–145, 2020.
  26. Machine-guided path sampling to discover mechanisms of molecular self-organization. Nature Computational Science, pages 1–12, 2023.
  27. Accurate transition state generation with an object-aware equivariant elementary reaction diffusion model. Nature Computational Science, 2023.
  28. Flow-based generative models for Markov chain Monte Carlo in lattice field theory. Physical Review D, 100(3):034515, August 2019.
  29. Solving Statistical Mechanics Using Variational Autoregressive Networks. Physical Review Letters, 122(8):080602, February 2019. arXiv:1809.10606 [cond-mat, stat].
  30. Boltzmann generators: Sampling equilibrium states of many-body systems with deep learning. Science, 365(6457):eaaw1147, September 2019.
  31. Adaptive Monte Carlo augmented with normalizing flows. Proceedings of the National Academy of Sciences, 119(10):e2109420119, March 2022. arXiv:2105.12603 [cond-mat, physics:physics].
  32. Targeted free energy estimation via learned mappings. The Journal of Chemical Physics, 153(14):144112, October 2020.
  33. Computing Absolute Free Energy with Deep Generative Models. The Journal of Physical Chemistry B, November 2020. Publisher: American Chemical Society.
  34. Normalizing flows for atomic solids. Machine Learning: Science and Technology, 3(2):025009, June 2022. arXiv:2111.08696 [cond-mat, physics:physics, stat].
  35. C. Jarzynski. Targeted free energy perturbation. Physical Review E, 65(4):046122, April 2002. Publisher: American Physical Society.
  36. Charles H. Bennett. Efficient estimation of free energy differences from Monte Carlo data. Journal of Computational Physics, 22(2):245–268, 1976.
  37. He Jia and Uroš Seljak. Normalizing Constant Estimation with Gaussianized Bridge Sampling, December 2019. arXiv:1912.06073 [astro-ph, stat].
  38. Peripheral nerves directly mediate the transneuronal translocation of silver nanomaterials from the gut to central nervous system. Science Advances, 9(27):eadg2252, 2023.
  39. Cellular binding, uptake and biotransformation of silver nanoparticles in human t lymphocytes. Nature Nanotechnology, 16(8):926–932, May 2021.
  40. Silver nanoparticle enhanced metal-organic matrix with interface-engineering for efficient photocatalytic hydrogen evolution. Nature Communications, 14(1), February 2023.
  41. Silver enhances antibiotic activity against gram-negative bacteria. Science Translational Medicine, 5, 6 2013.
  42. Engineering of Hole Transporting Interface by Incorporating the Atomic-Precision Ag6 Nanoclusters for High-Efficiency Blue Perovskite Light-Emitting Diodes. Nano Letters, 23(4):1582–1590, February 2023. Publisher: American Chemical Society.
  43. A water-soluble silver(I) formulation as an effective disinfectant of contact lenses cases. Materials Science and Engineering: C, 93:902–910, December 2018.
  44. Automated handling of complex chemical structures in z-matrix coordinates—the chemcoord library. Journal of Computational Chemistry, 44(5):710–726, 2023.
  45. Ultraviolet-visible absorption of small silver clusters in neon: Agn𝑛{}_{n}start_FLOATSUBSCRIPT italic_n end_FLOATSUBSCRIPT (n = 1–9). The Journal of Chemical Physics, 134:184504, 5 2011.
  46. R. Fournier. Theoretical study of the structure of silver clusters. The Journal of Chemical Physics, 115:2165–2177, 8 2001.
  47. Validation of methods for computational catalyst design: Geometries, structures, and energies of neutral and charged silver clusters. Journal of Physical Chemistry C, 119:9617–9626, 4 2015.
  48. Rank-normalization, folding, and localization: An improved R^^𝑅\widehat{R}over^ start_ARG italic_R end_ARG for assessing convergence of MCMC. Bayesian Analysis, 16(2), June 2021. arXiv:1903.08008 [stat].
  49. Physics-Inspired Structural Representations for Molecules and Materials. Chemical Reviews, 121(16):9759–9815, August 2021. Publisher: American Chemical Society.
  50. The inefficiency of re-weighted sampling and the curse of system size in high-order path integration. Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, 468(2137):2–17, January 2012.
  51. Ab Initio Molecular Dynamics: Basic Theory and Advanced Methods. Cambridge University Press, 2009.
  52. The atomic simulation environment—a python library for working with atoms. Journal of Physics: Condensed Matter, 29(27):273002, 2017.
  53. Real-space grid implementation of the projector augmented wave method. Phys. Rev. B, 71(3):035109, 2005.
  54. Electronic structure calculations with GPAW: a real-space implementation of the projector augmented-wave method. J. Phys.: Condens. Matter, 22(25):253202, 2010.
  55. Generalized gradient approximation made simple. Physical Review Letters, 77:3865, 10 1996.
  56. Density-functional theory for time-dependent systems. Phys. Rev. Lett., 52:997–1000, Mar 1984.
  57. Mark E. Casida. Time-dependent density functional response theory for molecules, November 1995.
  58. Time-dependent density-functional theory in the projector augmented-wave method. J. Chem. Phys., 128(24):244101, 2008.
  59. Localized surface plasmon resonance in silver nanoparticles: Atomistic first-principles time-dependent density-functional theory calculations. Phys. Rev. B, 91(11):115431, 2015.
  60. Kohn–sham decomposition in real-time time-dependent density-functional theory: An efficient tool for analyzing plasmonic excitations. J. Chem. Theory Comput., 13(10):4779–4790, 2017.
  61. Normalizing flows for probabilistic modeling and inference. The Journal of Machine Learning Research, 22:1–64, 1 2021.
  62. Density estimation using real nvp. 2017.
  63. ArviZ a unified library for exploratory analysis of Bayesian models in Python. Journal of Open Source Software, 4(33):1143, January 2019.
  64. pocomc: A python package for accelerated bayesian inference in astronomy and cosmology. arXiv preprint arXiv:2207.05660, 2022.
User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (4)
  1. Ana Molina-Taborda (2 papers)
  2. Pilar Cossio (17 papers)
  3. Olga Lopez-Acevedo (8 papers)
  4. Marylou Gabrié (26 papers)
Citations (1)
View on Semantic Scholar

Summary

An Active Learning Approach for Quantum Mechanical Accuracy in Molecular Simulations

The paper "Active learning of Boltzmann samplers and potential energies with quantum mechanical accuracy" introduces a novel framework, ab-flowMC, designed to enhance the efficiency and accuracy of molecular simulations requiring quantum-level precision. This method synergizes enhanced sampling techniques with deep generative models and an innovative active learning strategy to train Machine Learning Potentials (MLPs), aiming to surpass the computational limitations traditionally associated with Molecular Dynamics (MD) at quantum levels.

Methodology Overview

The central focus of the research is the challenge posed by high-dimensional free-energy landscapes, common in molecular systems where relevant states are separated by substantial free-energy barriers. The ab-flowMC framework integrates the sampling capacity of Normalizing Flows (NFs) with the predictive accuracy of MLPs, facilitating the exploration of molecular configurations with computational efficiency.

This method leverages an adaptive Markov Chain Monte Carlo (MCMC) procedure that employs NFs to propose samples around free-energy minima. The configurations' energies are then evaluated with Density Functional Theory (DFT) for a subset, while the MLP predicts the remainder, progressively refining its accuracy through active learning. This iterative mechanism enables concurrent training and sampling, aimed at achieving convergence to a Boltzmann distribution, thereby providing a statistically rigorous understanding of thermodynamic observables like free-energy differences and optical spectra.

Application and Results

The framework is applied to the isomerization paper of an ultrasmall silver nanocluster, Ag6Ag_6, which has various applications in medicine and catalysis. The authors provide a detailed comparison of configurations sampled via MD, ab-flowMC, and a variant excluding MLPs, ensuring a comprehensive evaluation of the model's performance in terms of potential energy distribution and acceptance rates within the MCMC simulations.

Statistically significant findings include the accurate prediction of optical spectra and free-energy differences across isomer states, attained with fewer computational resources than traditional ab initio methods. Key performance metrics suggest that ab-flowMC advances specimen sampling and energy predictions in both computational efficiency and accuracy.

Implications and Future Directions

This paper underscores the potential of active learning frameworks in quantum chemistry and molecular physics, heralding a substantial shift in efficiently managing complex systems with quantum mechanical considerations. By minimizing the reliance on costly DFT evaluations, this method provides a scalable solution to simulating larger systems and accessing intricate computational landscapes.

The implications for future research in AI involve developing more sophisticated ML architectures handling intricate potential energy surfaces and extending these approaches to a broader class of molecular systems and reactions. As computational chemistry prepares for models capable of tackling thousands of atoms, ab-flowMC sets a precedent for integrating AI-driven methodologies with foundational physical simulations, pushing the envelope of feasible molecular insights.

In summary, the research proposes a pragmatic yet significant advancement in molecular simulations, promoting an adaptive learning paradigm to balance computational requirements with rigorous quantum-level precision. This innovation fuels both theoretical exploration in free-energy calculations and practical applications in nanotechnology and biochemistry, setting the stage for future breakthroughs in AI-assisted scientific discovery.

File Document Download Save Streamline Icon: https://streamlinehq.com PDF File Document Download Save Streamline Icon: https://streamlinehq.com Markdown