Discovering dynamic laws from observations: the case of self-propelled, interacting colloids

Abstract

Active matter spans a wide range of time and length scales, from groups of cells and synthetic selfpropelled particles to schools of fish or even human crowds. The theoretical framework describing these systems has shown tremendous success at finding universal phenomenology. However, further progress is often burdened by the difficulty of determining the forces that control the dynamics of the individual elements within each system. Accessing this local information is key to understanding the physics dominating the system and to create the models that can explain the observed collective phenomena. In this work, we present a machine learning model, a graph neural network, that uses the collective movement of the system to learn the active and two-body forces controlling the individual dynamics of the particles. We verify our approach using numerical simulations of active brownian particles, considering different interaction potentials and levels of activity. Finally, we apply our model to experiments of electrophoretic Janus particles, extracting the active and twobody forces that control the dynamics of the colloids. Due to this, we can uncover the physics dominating the behavior of the system. We extract an active force that depends on the electric field and also area fraction. We also discover a dependence of the two-body interaction with the electric field that leads us to propose that the dominant force between these colloids is a screened electrostatic interaction with a constant length scale. We expect that this methodology can open a new avenue for the study and modeling of experimental systems of active particles.

Learning dynamical models of single and collective cell migration: a review

“…In that case, all position dependence simply becomes a function of the radial distance r ij of each pair of cells. A further promising approach is the inclusion of small convolutional neural networks as the basis functions of the expansion (equation ( 16)), which has been applied to the case of interacting active colloids [308]. An advantage of this approach is that it may reduce the risk of overfitting and provide a flexible basis for complex interaction functions.…”

A Neural-Network-Optimized Hydrogen Peroxide Pairwise Additive Model for Classical Simulations

“…Our approach aims to optimize the parameters by considering several properties of pure H 2 O 2 and water mixtures simultaneously, targeting 22 different values. Notably, ANNs have recently been employed for optimization purposes of pairwise additive molecular models, 29−32 extracting forces and dynamic laws from active-matter trajectories 33,34 and even serving as complex potential functions for molecular simulations. 35−41 We present several liquid and vapor properties of the obtained pairwise additive model as a function of the weight percentage of the water− peroxide mixture and temperature.…”

Dead or alive: Distinguishing active from passive particles using supervised learning (a)

“…These findings firmly establish a correlation between local structure and the propensity of passive particles to move in a crowded environment. While ML has also been applied with considerable success in purely active systems [25][26][27][28][29][30][31][32][33], in practice and particularly in biological settings the system of interest will generally contain actors whose activity parameters are different, and distributed. In tumor tissue, for instance, there may be a distribution of epithelial (more stationary, passive) phenotypes and mesenchymal (more motile, active) ones, with the presence of mesenchymal cells being generally associated with greater metastatic potential [5,34].…”