Molecular Dynamics With On-The-Fly Machine Learning of QM forces
Event details
| Date | 13.04.2015 |
| Hour | 13:15 › 14:15 |
| Speaker | Prof. Alessandro De Vita, Physics Department King's College London UK |
| Location | |
| Category | Conferences - Seminars |
Abstract: Many technologically important, chemically complex phenomena are currently beyond the reach of standard or O(N) first principles molecular dynamics (MD) techniques, mostly because the necessary model system sizes are too large, and/or the required simulation times are too long. In most situations, using classical MD is not a viable alternative, as suitably general and accurate “reactive” force fields are not available, nor are fitting databases a priori guaranteed to contain the information necessary to describe all the chemical processes encountered along the dynamics. “QM/MM” techniques combining quantum and classical zones in a single calculation are also not devoid of difficulties, especially for applications on processes involving sustained mass transport into and out of the (e.g., fast moving) QM zone.
In this colloquium I will argue that this situation forces the use of MD techniques capable of incorporating accurate QM information generated at run time during the simulations [1, 2]. It also effectively creates a novel market for dynamical databases coupled with specially-tuned Machine Learning (ML) force fields which minimise the computational workload by allowing QM subroutine calls only when “chemically novel” configurations are encountered along the system’s trajectory. I will present one such “Learn On the Fly” scheme, effectively unifying First-Principles Molecular Dynamics and Machine Learning into a single, information efficient simulation scheme capable of learning/predicting atomic forces through Bayesian inference [3]. Interestingly, QM-zone partitioning and execution via any of the existing O(N3) QM engines is predicted to be a better option than using O(N) QM methods when dealing with large QM zones in QM/MM calculations running on high-end parallel platforms [4-5].
References
[1] J.R.Kermode, L.Ben-Bashat, F.Atrash, J.J.Cilliers, D.Sherman and A.De Vita, Nat. Commun. 4, 2441 (2013).
[2] A. Gleizer, G. Peralta, J. R. Kermode, A. De Vita and D. Sherman, Phys. Rev. Lett., 112, 115501 (2014).
[3] Z. Li, J. R. Kermode and A. De Vita, Phys. Rev. Lett., 114, 096405 (2015).
[4] Cf., e.g., the US-DOE INCITE on SiO2 ML-Fracture Project https://www.alcf.anl.gov/projects/sio2-fracture-chemomechanics-machine-learning-hybrid-qmmm-scheme
[5] M. Caccin, Z. Li, J. R. Kermode and A. De Vita, in preparation.
In this colloquium I will argue that this situation forces the use of MD techniques capable of incorporating accurate QM information generated at run time during the simulations [1, 2]. It also effectively creates a novel market for dynamical databases coupled with specially-tuned Machine Learning (ML) force fields which minimise the computational workload by allowing QM subroutine calls only when “chemically novel” configurations are encountered along the system’s trajectory. I will present one such “Learn On the Fly” scheme, effectively unifying First-Principles Molecular Dynamics and Machine Learning into a single, information efficient simulation scheme capable of learning/predicting atomic forces through Bayesian inference [3]. Interestingly, QM-zone partitioning and execution via any of the existing O(N3) QM engines is predicted to be a better option than using O(N) QM methods when dealing with large QM zones in QM/MM calculations running on high-end parallel platforms [4-5].
References
[1] J.R.Kermode, L.Ben-Bashat, F.Atrash, J.J.Cilliers, D.Sherman and A.De Vita, Nat. Commun. 4, 2441 (2013).
[2] A. Gleizer, G. Peralta, J. R. Kermode, A. De Vita and D. Sherman, Phys. Rev. Lett., 112, 115501 (2014).
[3] Z. Li, J. R. Kermode and A. De Vita, Phys. Rev. Lett., 114, 096405 (2015).
[4] Cf., e.g., the US-DOE INCITE on SiO2 ML-Fracture Project https://www.alcf.anl.gov/projects/sio2-fracture-chemomechanics-machine-learning-hybrid-qmmm-scheme
[5] M. Caccin, Z. Li, J. R. Kermode and A. De Vita, in preparation.
Links
Practical information
- General public
- Free
Organizer
- Fabien Sorin
Contact
- Fabien Sorin