A MACHINE LEARNING APPROACH FOR DYNAMICAL MASS MEASUREMENTS OF GALAXY CLUSTERS

Toward a Mobility-Preserving Coarse-Grained Model: A Data-Driven Approach

Coarse-grained molecular dynamics (MD) simulation is a promising alternative to all-atom MD simulation for the fast calculation of system properties, which is imperative in designing materials with a specific target property. There have been several coarse-graining strategies developed over the past few years that provide accurate structural properties of the system. However, these coarse-grained models share a major drawback in that they introduce an artificial acceleration in molecular mobility. In this paper, we report a data-driven approach to generate coarse-grained models that preserve the all-atom molecular mobility. We designed a machine learning model in the form of an artificial neural network, which directly predicts the simulation-ready mobility-preserving coarse-grained potential as an output given the all-atom force field (FF) parameters as inputs. As a proof of principle, we took 2,3,4-trimethylpentane as a model system and described the development of machine learning models in detail. We quantify the artificial acceleration in molecular mobility by defining the acceleration factor as the ratio of the coarse-grained and the all-atom diffusion coefficient. The predicted coarse-grained potential generated by the best machine learning model can bring down the acceleration factor to a value of ∼2, which could be otherwise as large as 7 for a typical value of 3 × 10^-9 m² s^-1 for the all-atom diffusion coefficient. We believe our method will be of interest in the community as a route to generating coarse-grained potentials with accurate dynamics.

Interaction from structure using machine learning: in and out of equilibrium

Prediction of pair potential given a typical configuration of an interacting classical system is a difficult inverse problem. There exists no exact result that can predict the potential given the structural information. We demonstrate that using machine learning (ML) one can get a quick but accurate answer to the question: "which pair potential lead to the given structure (represented by pair correlation function)?" We use artificial neural network (NN) to address this question and show that this ML technique is capable of providing very accurate prediction of pair potential irrespective of whether the system is in a crystalline, liquid or gas phase. We show that the trained network works well for sample system configurations taken from both equilibrium and out of equilibrium simulations (active matter systems) when the later is mapped to an effective equilibrium system with a modified potential. We show that the ML prediction about the effective interaction for the active system is not only useful to make prediction about the MIPS (motility induced phase separation) phase but also identifies the transition towards this state.