Abstract
Determining which microstates generated by a thermodynamic simulation are representative of the ensemble for which sampling is desired is a ubiquitous, underspecified problem. Artificial neural networks are one type of machine learning algorithm that can provide a reproducible way to apply pattern recognition heuristics to underspecified problems. Here we use the open-source TensorFlow machine learning library and apply it to the problem of identifying which hypothetical observation sequences from a computer simulation are “equilibrated” and which are not. We generate training populations and test populations of observation sequences with embedded linear and exponential correlations. We train a two-neuron artificial network to distinguish the correlated and uncorrelated sequences. We find that this simple network is good enough for > 98% accuracy in identifying exponentially-decaying energy trajectories from molecular simulations.
Highlights
Computer simulations that model physical systems in equilibrium are tools that have been essential to understanding an enormous variety of phenomena ranging from liquid argon[1], the self-assembly of colloids[2], criticality in spin glasses[3, 4], lipid self-assembly[5], DNA-origami[6], protein folding[7, 8], and the segregation of neighborhoods[9]
Artificial neural networks are a promising advance for automating the identification of “equilibrated” measurements from thermodynamic simulations
A few minutes of training time on a laptop graphics processing units (GPUs) is sufficient to train a two-neuron artificial network for identifying an exponential decay hidden in normally-distributed data more than 96% of the time
Summary
Computer simulations that model physical systems in equilibrium are tools that have been essential to understanding an enormous variety of phenomena ranging from liquid argon[1], the self-assembly of colloids[2], criticality in spin glasses[3, 4], lipid self-assembly[5], DNA-origami[6], protein folding[7, 8], and the segregation of neighborhoods[9]. For the modeling of material systems, Metropolis Monte Carlo (MC)[10] and molecular dynamics (MD)[11] are the two most popular tools for sampling equilibrium ensembles of microstates. Both techniques generate sequences of microstates and the main cost of performing an MD or MC simulation is waiting for this sequence to converge to one that is representative of the thermodynamic ensemble at equilibrium, and sampling enough microstates to perform ensemble averages with high precision. Sampling the equilibrium distribution of microstates at a particular thermodynamic state point is challenging because both MD and MC techniques are inefficient at overcoming free energy barriers that separate local minima in the multidimensional configurational landscape[12]. Every researcher performing MD or MC simulations has encountered this equilibration problem when asking themselves “Have I run my simulation for long enough?”
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
