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 self-propelled particles to schools of fish, flocks of birds, 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 two-body 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.

16 pages, 9 figures

References in corpus (52)

• Discovering governing equations from data: Sparse identification of nonlinear dynamical systems
• Geometric deep learning: going beyond Euclidean data
• Collective motion
• Crystal Graph Convolutional Neural Networks for an Accurate and Interpretable Prediction of Material Properties
• The Mechanics and Statistics of Active Matter
• Flocks, herds, and schools: A quantitative theory of flocking
• Molecular Contrastive Learning of Representations via Graph Neural Networks
• Long-lived Giant Number Fluctuations in a Swarming Granular Nematic
• SISSO: a compressed-sensing method for identifying the best low-dimensional descriptor in an immensity of offered candidates
• Motility-driven glass and jamming transitions in biological tissues
• Identifying structural flow defects in disordered solids using machine learning methods
• Physical Models of Collective Cell Migration
• Phase behaviour of active Brownian particles: The role of dimensionality
• Activity-induced phase separation and self-assembly in mixtures of active and passive particles
• Emergent vortices in populations of colloidal rollers
• Analysis of data sets of stochastic systems
• Self-propulsion of a catalytically active particle near a planar wall: from reflection to sliding and hovering
• Inertial effects of self-propelled particles: from active Brownian to active Langevin motion
• Dynamics of a self-propelled particle in a harmonic trap
• Combinatorial Optimization with Physics-Inspired Graph Neural Networks
• Living clusters and crystals from low density suspensions of active colloids
• Optimized Interactions for Targeted Self-Assembly: Application to Honeycomb Lattice
• An active approach to colloidal self-assembly
• Interrupted Motility Induced Phase Separation in Aligning Active Colloids
• Long-range Ordering of Vibrated Polar Disks
• Mesoscopic Turbulence and Local Order in Janus Particles Self-Propelling under an AC Electric Field
• Learning to school in the presence of hydrodynamic interactions
• Active phase separation by turning toward regions of higher density
• DeepMoD: Deep learning for Model Discovery in noisy data
• Light-switchable propulsion of active particles with reversible interactions
• Flagellar Dynamics of Chains of Active Janus Particles Fueled by an AC electric field
• Waterlike thermodynamic anomalies in a repulsive-step potential system
• Feedback-Controlled Active Brownian Colloids with Space-Dependent Rotational Dynamics
• Inferring the dynamics of underdamped stochastic systems
• Learning the dynamics of cell-cell interactions in confined cell migration
• Geometric deep learning reveals the spatiotemporal fingerprint of microscopic motion
• Learning force fields from stochastic trajectories
• Activity-controlled Annealing of Colloidal Monolayers
• Dynamical Clustering Interrupts Motility Induced Phase Separation in Chiral Active Brownian Particles
• Inverse Statistical Mechanics: Probing the Limitations of Isotropic Pair Potentials to Produce Ground-State Structural Extremes
• Phase behaviour and dynamical features of a two-dimensional binary mixture of active/passive spherical particles
• Robust Prediction of Force Chains in Jammed Solids using Graph Neural Networks
• Comparing machine learning techniques for predicting glassy dynamics
• An Introduction to Motility-Induced Phase Separation
• Rheology of active fluids
• Microrobots from Toposelective Nanoparticle Attachment
• Quantifying the link between local structure and cellular rearrangements using information in models of biological tissues
• Tracking Janus microswimmers in 3D with Machine Learning
• Characterizing Different Motility Induced Regimes in Active Matter with Machine Learning and Noise
• Machine learning phases of active matter
• Interaction of a migrating cell monolayer with a flexible fiber
• Chiral Active Matter

Cited by in corpus (8)

• Roadmap on machine learning glassy dynamics
• Re-entrant percolation in active Brownian hard disks
• Shape matters: Inferring the motility of confluent cells from static images
• A Graph Neural Network Simulation of Dispersed Systems
• Effective interactions in active Brownian particles
• Active Matter as a framework for living systems-inspired Robophysics
• Direct Boltzmann inversion method from particle configurations at arbitrary state points
• Learning general pair interactions between self-propelled particles