Sampling lattices in semi-grand canonical ensemble with autoregressive machine learning
Source abstract
Abstract
Abstract Calculating thermodynamic potentials and observables efficiently and accurately is key for the application of statistical mechanics simulations to materials science. However, naive Monte Carlo approaches, on which such calculations are often dependent, struggle to scale to complex materials in many state-of-the-art disciplines such as the design of high entropy alloys or multi-component catalysts. To address this issue, we adapt sampling tools built upon machine learning-based generative modeling to the materials space by transforming them into the semi-grand canonical ensemble. Furthermore, we show that the resulting models are transferable across wide ranges of thermodynamic conditions and can be implemented with any internal energy model U, allowing integration into many existing materials workflows. We demonstrate the applicability of this approach to the simulation of benchmark systems (AgPd, CuAu) that exhibit diverse thermodynamic behavior in their phase diagrams. Finally, we discuss remaining challenges in model development and promising research directions for future improvements.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Machine Learning in Materials Science · Protein Structure and Dynamics · Catalysis and Oxidation Reactions
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
References (74)
Show source-provided references
1. Thomas, J. C. et al. CASM, v0.2.1. https://github.com/prisms-center/CASMcode/tree/v0.2.1 (2021).
2. Van der Ven, A., Thomas, J., Puchala, B. & Natarajan, A. First-principles statistical mechanics of multicomponent crystals. Annu. Rev. Mater. Res. 48, 27–55 (2018).10.1146/annurev-matsci-070317-124443 ↗
3. Ångqvist, M. et al. ICET—a Python library for constructing and sampling alloy cluster expansions. Adv. Theory Simul. 2, 1900015 (2019).10.1002/adts.201900015 ↗
4. Chang, J. H. et al. CLEASE: a versatile and user-friendly implementation of cluster expansion method. J. Phys. Condens. Matter 31, 325901 (2019).10.1088/1361-648X/ab1bbc ↗
5. Lerch, D., Wieckhorst, O., Hart, G. L., Forcade, R. W. & Müller, S. UNCLE: a code for constructing cluster expansions for arbitrary lattices with minimal user-input. Model. Simul. Mater. Sci. Eng. 17, 55003 (2009).10.1088/0965-0393/17/5/055003 ↗
6. van de Walle, A. & Ceder, G. Automating first-principles phase diagram calculations. J. Phase Equilib. 23, 348–359 (2002).10.1361/105497102770331596 ↗
7. van de Walle, A., Asta, M. & Ceder, G. The alloy theoretic automated toolkit: a user guide. Calphad 26, 539–553 (2002).10.1016/S0364-5916(02)80006-2 ↗
8. Bäker, M. Calculating phase diagrams with ATAT. Preprint at https://arxiv.org/abs/1907.10151 (2019).
9. Troppenz, M., Rigamonti, S. & Draxl, C. Predicting ground-state configurations and electronic properties of the thermoelectric clathrates Ba8AlxSi46−x and Sr8AlxSi46−x. Chem. Mater. 29, 2414–2424 (2017).10.1021/acs.chemmater.6b05027 ↗
10. Metropolis, N., Rosenbluth, A. W., Rosenbluth, M. N., Teller, A. H. & Teller, E. Equation of state calculations by fast computing machines. J. Chem. Phys. 21, 1087 (1953).10.1063/1.1699114 ↗
11. Swendsen, R. H. & Wang, J. S. Nonuniversal critical dynamics in Monte Carlo simulations. Phys. Rev. Lett. 58, 86–88 (1987).10.1103/PhysRevLett.58.86 ↗
12. Wolff, U. Collective Monte Carlo updating for spin systems. Phys. Rev. Lett. 62, 361–364 (1989).10.1103/PhysRevLett.62.361 ↗
13. Swendsen, R. H. & Wang, J. S. Replica Monte Carlo simulation of spin-glasses. Phys. Rev. Lett. 57, 2607–2609 (1986).10.1103/PhysRevLett.57.2607 ↗
14. Wang, F. & Landau, D. P. Efficient, multiple-range random walk algorithm to calculate the density of states. Phys. Rev. Lett. 86, 2050–2053 (2001).10.1103/PhysRevLett.86.2050 ↗
15. Widom, M. Modeling the structure and thermodynamics of high-entropy alloys. J. Mater. Res. 33, 2881–2898 (2018).10.1557/jmr.2018.222 ↗
16. Antillon, E. & Ghazisaeidi, M. Efficient determination of solid-state phase equilibrium with the multicell Monte Carlo method. Phys. Rev. E 101, 063306 (2020).10.1103/PhysRevE.101.063306 ↗
17. Niu, C., Windl, W. & Ghazisaeidi, M. Multi-cell Monte Carlo relaxation method for predicting phase stability of alloys. Scr. Mater. 132, 9–12 (2017).10.1016/j.scriptamat.2017.01.001 ↗
18. Niu, C., Rao, Y., Windl, W. & Ghazisaeidi, M. Multi-cell Monte Carlo method for phase prediction. Npj Comput. Mater. 5, 1–5 (2019).10.1038/s41524-019-0259-z ↗
19. Sadigh, B. & Erhart, P. Calculation of excess free energies of precipitates via direct thermodynamic integration across phase boundaries. Phys. Rev. B 86, 134204 (2012).10.1103/PhysRevB.86.134204 ↗
20. Takeuchi, K., Tanaka, R. & Yuge, K. New Wang-Landau approach to obtain phase diagrams for multicomponent alloys. Phys. Rev. B 96, 144202 (2017).10.1103/PhysRevB.96.144202 ↗
21. Schwalbe-Koda, D. & Gómez-Bombarelli, R. Generative Models for Automatic Chemical Design 445–467 (Lecture Notes in Physics, Vol. 968, Springer, 2020).10.1007/978-3-030-40245-7_21 ↗
22. Gómez-Bombarelli, R. et al. Automatic chemical design using a data-driven continuous representation of molecules. ACS Cent. Sci. 4, 268–276 (2018).10.1021/acscentsci.7b00572 ↗
23. Dan, Y. et al. Generative adversarial networks (GAN) based efficient sampling of chemical composition space for inverse design of inorganic materials. Npj Comput. Mater. 6, 84 (2020).10.1038/s41524-020-00352-0 ↗
24. Kim, B., Lee, S. & Kim, J. Inverse design of porous materials using artificial neural networks. Sci. Adv. 6, eaax9324 (2020).10.1126/sciadv.aax9324 ↗
25. Roy, A., Saffar, M., Vaswani, A. & Grangier, D. Efficient content-based sparse attention with routing transformers. Trans. Assoc. Comput. 9, 53–68 (2021).
26. Salimans, T., Karpathy, A., Chen, X. & Kingma, D. P. PixelCNN++: improving the PixelCNN with discretized logistic mixture likelihood and other modifications. In Proc. of the 5th International Conference on Learning Representations (2017).
27. Noé, F., Olsson, S., Köhler, J. & Wu, H. Boltzmann generators: sampling equilibrium states of many-body systems with deep learning. Science 365, eaaw1147 (2019).10.1126/science.aaw1147 ↗
28. Nicoli, K. A. et al. Asymptotically unbiased estimation of physical observables with neural samplers. Phys. Rev. E 101, 23304 (2020).10.1103/PhysRevE.101.023304 ↗
29. Albergo, M. S., Kanwar, G. & Shanahan, P. E. Flow-based generative models for markov chain monte carlo in lattice field theory. Phys. Rev. D. 100, 034515 (2019).10.1103/PhysRevD.100.034515 ↗
30. Kanwar, G. et al. Equivariant flow-based sampling for lattice gauge theory. Phys. Rev. Lett. 125, 121601 (2020).10.1103/PhysRevLett.125.121601 ↗
31. Pawlowski, J. M. & Urban, J. M. Reducing autocorrelation times in lattice simulations with generative adversarial networks. Mach. Learn.: Sci. Technol. 1, 045011 (2020).
32. Li, S. H. & Wang, L. Neural Network Renormalization Group. Phys. Rev. Lett. 121, 260601 (2018).10.1103/PhysRevLett.121.260601 ↗
33. Zhang, L., E, W. & Wang, L. Monge-Ampère flow for generative modeling. Preprint at https://arxiv.org/abs/1809.10188 (2018).
34. Wu, D., Wang, L. & Zhang, P. Solving statistical mechanics using variational autoregressive networks. Phys. Rev. Lett. 122, 080602 (2019).10.1103/PhysRevLett.122.080602 ↗
35. Mcnaughton, B., Milošević, M. V., Perali, A. & Pilati, S. Boosting Monte Carlo simulations of spin glasses using autoregressive neural networks. Phys. Rev. E 101, 53312 (2020).10.1103/PhysRevE.101.053312 ↗
36. Hibat-Allah, M., Inack, E. M., Wiersema, R., Melko, R. G. & Carrasquilla, J. Variational neural annealing. Nat. Mach. Intell. 3, 952–961 (2021).10.1038/s42256-021-00401-3 ↗
37. Singh, J., Scheurer, M. S. & Arora, V. Conditional generative models for sampling and phase transition indication in spin systems. SciPost Phys. 11, 43 (2021).10.21468/SciPostPhys.11.2.043 ↗
38. Dibak, M., Klein, L. & Noé, F. Temperature-steerable flows. In Proc. of the 34th Conference on Neural Information Processing Systems—ML4PS Workshop (2020).
39. Belardinelli, R. E. & Pereyra, V. D. Wang-Landau algorithm: a theoretical analysis of the saturation of the error. J. Chem. Phys. 127, 184105 (2007).10.1063/1.2803061 ↗
40. Belardinelli, R. E. & Pereyra, V. D. Fast algorithm to calculate density of states. Phys. Rev. E 75, 046701 (2007).10.1103/PhysRevE.75.046701 ↗
41. Haule, K. Wang-Landau algorithm for 2D Ising model. http://www.physics.rutgers.edu/h̃aule/681/src_MC/python_codes/wangLand.py (2010).
42. Kaufman, B. Crystal statistics. II. Partition function evaluated by spinor analysis. Phys. Rev. 76, 1232–1243 (1949).10.1103/PhysRev.76.1232 ↗
43. Beale, P. D. Exact distribution of energies in the two-dimensional ising model. Phys. Rev. Lett. 76, 78–81 (1996).10.1103/PhysRevLett.76.78 ↗
44. Pathria, R. K. & Beale, P. D. Statistical Mechanics 3rd edn (Elsevier Ltd, 2011).
45. Wang, W., Axelrod, S. & Gómez-Bombarelli, R. Differentiable molecular simulations for control and learning. Preprint at https://arxiv.org/abs/2003.00868 (2020).
46. Fontaine, D. D. Cluster Approach to Order-Disorder Transformations in Alloys 33–176 (Solid State Physics, Vol. 47, Academic Press, 1994).10.1016/S0081-1947(08)60639-6 ↗
47. Lu, Z. W., Wei, S. H., Zunger, A., Frota-Pessoa, S. & Ferreira, L. G. First-principles statistical mechanics of structural stability of intermetallic compounds. Phys. Rev. B 44, 512–544 (1991).10.1103/PhysRevB.44.512 ↗
48. Ozoliņš, V., Wolverton, C. & Zunger, A. Cu-Au, Ag-Au, Cu-Ag, and Ni-Au intermetallics: first-principles study of temperature-composition phase diagrams and structures. Phys. Rev. B 57, 6427–6443 (1998).10.1103/PhysRevB.57.6427 ↗
49. Zhang, Y., Kresse, G. & Wolverton, C. Nonlocal first-principles calculations in Cu-Au and other intermetallic alloys. Phys. Rev. Lett. 112, 075502 (2014).10.1103/PhysRevLett.112.075502 ↗
50. Kleivan, D., Akola, J., Peterson, A. A., Vegge, T. & Chang, J. H. Training sets based on uncertainty estimates in the cluster-expansion method. J. Phys. Energy 3, 034012 (2021).10.1088/2515-7655/abf9ef ↗
51. Jain, A. et al. The Materials Project: a materials genome approach to accelerating materials innovation. APL Mater. 1, 011002 (2013).10.1063/1.4812323 ↗
52. Ghosh, G., Kanter, C. & Olson, G. Thermodynamic modeling of the Pd-X (X=Ag, Co, Fe, Ni) systems. J. Phase Equilib. 20, 295–308 (1999).10.1361/105497199770335811 ↗
53. Dinsdale, A. et al. Atlas of Phase Diagrams for Lead-Free Soldering (Cost Action 531, European Cooperation in Science and Technology, 2008).
54. Xie, T. & Grossman, J. C. Crystal graph convolutional neural networks for an accurate and interpretable prediction of material properties. Phys. Rev. Lett. 120, 145301 (2018).10.1103/PhysRevLett.120.145301 ↗
55. Fung, V., Zhang, J., Juarez, E. & Sumpter, B. G. Benchmarking graph neural networks for materials chemistry. Npj Comput. Mater. 7, 84 (2021).10.1038/s41524-021-00554-0 ↗
56. Schütt, K. T. et al. SchNetPack: a deep learning toolbox for atomistic systems. J. Chem. Theory Comput. 15, 448–455 (2019).10.1021/acs.jctc.8b00908 ↗
57. Felzenszwalb, P. F. & Huttenlocher, D. P. Efficient graph-based image segmentation. Int. J. Comput. Vis. 59, 167–181 (2004).10.1023/B:VISI.0000022288.19776.77 ↗
58. van der Walt, S. et al. scikit-image: image processing in Python. PeerJ 2, e453 (2014).10.7717/peerj.453 ↗
59. Liu, X. et al. Monte Carlo simulation of order-disorder transition in refractory high entropy alloys: A data-driven approach. Comput. Mater. Sci. 187, 110135 (2021).10.1016/j.commatsci.2020.110135 ↗
60. Boyda, D. et al. Sampling using SU(n) gauge equivariant flows. Phys. Rev. D. 103, 074504 (2021).10.1103/PhysRevD.103.074504 ↗
61. Pan, F., Zhou, P., Zhou, H. J. & Zhang, P. Solving statistical mechanics on sparse graphs with feedback-set variational autoregressive networks. Phys. Rev. E 103, 012103 (2021).10.1103/PhysRevE.103.012103 ↗
62. Dai, H., Nazi, A., Li, Y., Dai, B. & Schuurmans, D. Scalable deep generative modeling for sparse graphs. In Proc. of the 37th International Conference on Machine Learning, PMLR 119, 2302–2312 (2020).
63. Wu, D., Rossi, R. & Carleo, G. Unbiased monte carlo cluster updates with autoregressive neural networks. Phys. Rev. Res. 3, L042024 (2021).10.1103/PhysRevResearch.3.L042024 ↗
64. Zhou, K., Endrödi, G., Pang, L.-G. & Stöcker, H. Regressive and generative neural networks for scalar field theory. Phys. Rev. D. 100, 011501 (2019).10.1103/PhysRevD.100.011501 ↗
65. Williams, R. J. Simple statistical gradient-following algorithms for connectionist reinforcement learning. Mach. Learn. 8, 229–256 (1992).10.1007/BF00992696 ↗
66. Kresse, G. & Furthmüller, J. Efficiency of ab-initio total energy calculations for metals and semiconductors using a plane-wave basis set. Comput. Mater. Sci. 6, 15–50 (1996).10.1016/0927-0256(96)00008-0 ↗
67. Kresse, G. & Furthmüller, J. Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set. Phys. Rev. B 54, 11169–11186 (1996).10.1103/PhysRevB.54.11169 ↗
68. Blöchl, P. E. Projector augmented-wave method. Phys. Rev. B 50, 17953–17979 (1994).10.1103/PhysRevB.50.17953 ↗
69. Kresse, G. & Joubert, D. From ultrasoft pseudopotentials to the projector augmented-wave method. Phys. Rev. B 59, 1758–1775 (1999).10.1103/PhysRevB.59.1758 ↗
70. Perdew, J. P., Burke, K. & Ernzerhof, M. Generalized gradient approximation made simple. Phys. Rev. Lett. 77, 3865–3868 (1996).10.1103/PhysRevLett.77.3865 ↗
71. Grimme, S., Antony, J., Ehrlich, S. & Krieg, H. A consistent and accurate ab initio parametrization of density functional dispersion correction (DFT-D) for the 94 elements H-Pu. J. Chem. Phys. 132, 154104 (2010).10.1063/1.3382344 ↗
72. Grimme, S., Ehrlich, S. & Goerigk, L. Effect of the damping function in dispersion corrected density functional theory. J. Comput. Chem. 32, 1456–1465 (2011).10.1002/jcc.21759 ↗
73. Monkhorst, H. J. & Pack, J. D. Special points for Brillouin-zone integrations. Phys. Rev. B 13, 5188–5192 (1976).10.1103/PhysRevB.13.5188 ↗
74. Towns, J. et al. XSEDE: accelerating scientific discovery. Comput. Sci. Eng. 16, 62–74 (2014).10.1109/MCSE.2014.80 ↗