Berlin 2018 – scientific programme

Parts | Days | Selection | Search | Updates | Downloads | Help

CPP: Fachverband Chemische Physik und Polymerphysik

CPP 34: Condensed Matter Simulations augmented by Advanced Statistical Methodologies II (joint session DY/CPP)

CPP 34.4: Talk

Tuesday, March 13, 2018, 14:45–15:00, BH-N 128

Identifying the relevant degrees of freedom in mesoscale models of liquid water with Bayesian formalism — •Julija Zavadlav and Petros Koumoutsakos — Computational Science and Engineering Laboratory, ETH Zurich, Zurich, CH-8092, Switzerland

Coarse-graining (CG) has become an established methodology in molecular modeling to access time and length scales that are computationally beyond the reach of the conventional atomistic simulations. However, it often involves making several a priori assumptions, which are rarely systematically addressed. Typically, these assumptions pertain to the level of coarse-graining and the model complexity. We address this issue for mesoscale models of liquid water by investigating on an equal footing a number of CG models that differ in the level of coarse-graining and in the model complexity. To this end, we deploy the classical as well as a novel Hierarchical Bayesian methods [1,2] to quantify and calibrate the uncertainty of the models and to perform the model selection using the experimental data. Furthermore, we assess the efficiency-accuracy trade-off of developed models and provide guidelines for future water model design at the mesoscopic scale.

[1] S.Wu, P. Angelikopoulos, G. Tauriello, C. Papadimitriou, and P. Koumoutsakos, J. Chem. Phys. 145, 244112, 2016.

[2] L. Kulakova, G. Arampatzis, P. Angelikopoulos, P. Hadjidoukas, C. Papadimitriou, and P. Koumoutsakos, Sci. Rep., 7, 16576, 2017