arXiv subjects

B. A. Lindquist

Publications and source records attributed to B. A. Lindquist.

3 recordsLinked to original sources

arXiv2018

Unsupervised machine learning for detection of phase transitions in off-lattice systems I. Foundations

We demonstrate the utility of an unsupervised machine learning tool for the detection of phase transitions in off-lattice systems. We focus on the application of principal component analysis (PCA) to detect the freezing transitions of two-dimensional hard-disk and three-dimensional hard-sphere systems as well as liquid-gas phase separation in a patchy colloid model. As we demonstrate, PCA autonomously discovers order-parameter-like quantities that report on phase transitions, mitigating the need for a priori construction or identification of a suitable order parameter--thus streamlining the routine analysis of phase behavior. In a companion paper, we further develop the method established here to explore the detection of phase transitions in various model systems controlled by compositional demixing, liquid crystalline ordering, and non-equilibrium active forces.

physics.comp-ph↗

arXiv2018

Unsupervised machine learning for detection of phase transitions in off-lattice systems II. Applications

We outline how principal component analysis (PCA) can be applied to particle configuration data to detect a variety of phase transitions in off-lattice systems, both in and out of equilibrium. Specifically, we discuss its application to study 1) the nonequilibrium random organization (RandOrg) model that exhibits a phase transition from quiescent to steady-state behavior as a function of density, 2) orientationally and positionally driven equilibrium phase transitions for hard ellipses, and 3) compositionally driven demixing transitions in the non-additive binary Widom-Rowlinson mixture.

physics.comp-ph↗

arXiv2017

Probabilistic inverse design for self assembling materials

One emerging approach for the fabrication of complex architectures on the nanoscale is to utilize particles customized to intrinsically self-assemble into a desired structure. Inverse methods of statistical mechanics have proven particularly effective for the discovery of interparticle interactions suitable for this aim. Here we evaluate the generality and robustness of a recently introduced inverse design strategy [Lindquist et al., J. Chem. Phys. 145, 111101 (2016)] by applying this simulated-based, machine learning method to optimize for interparticle interactions that self-assemble particles into a variety of complex microstructures: cluster fluids, porous mesophases, and crystalline lattices. Using the method, we discover isotropic pair interactions that lead to self-assembly of each of the desired morphologies, including several types of potentials that were not previously understood to be capable of stabilizing such systems. One such pair potential led to assembly of the highly asymmetric truncated trihexagonal lattice and another produced a fluid containing spherical voids, or pores, of designed size via purely repulsive interactions. Through these examples, we demonstrate several advantages inherent to this particular design approach including the use of a parametrized functional form for the optimized interparticle interactions, the ability to constrain the range of said parameters, and compatibility of the inverse design strategy with a variety of simulation protocols (e.g., positional restraints).

cond-mat.mtrl-sci↗