Machine learning active-nematic hydrodynamics

Abstract

Hydrodynamic theories effectively describe many-body systems out of equilibrium in terms of a few macroscopic parameters. However, such hydrodynamic parameters are difficult to derive from microscopics. Seldom is this challenge more apparent than in active matter where the energy cascade mechanisms responsible for autonomous large-scale dynamics are poorly understood. Here, we use active nematics to demonstrate that neural networks can extract the spatio-temporal variation of hydrodynamic parameters directly from experiments. Our algorithms analyze microtubule-kinesin and actin-myosin experiments as computer vision problems. Unlike existing methods, neural networks can determine how multiple parameters such as activity and elastic constants vary with ATP and motor concentration. In addition, we can forecast the evolution of these chaotic many-body systems solely from image-sequences of their past by combining autoencoder and recurrent networks with residual architecture. Our study paves the way for artificial-intelligence characterization and control of coupled chaotic fields in diverse physical and biological systems even when no knowledge of the underlying dynamics exists.

arXiv commentSI Movie 1: https://www.youtube.com/watch?v=9WzIT7OG9pY SI Movie 2: https://youtu.be/Trc4RyU7-dw SI Movie 3: https://youtu.be/Epm_P_EakH8

References in corpus (39)

• Deep Learning in Neural Networks: An Overview
• Deep Residual Learning for Image Recognition
• Machine Learning for Fluid Mechanics
• Machine learning and the physical sciences
• Turbulence Modeling in the Age of Data
• Spontaneous motion in hierarchically assembled active matter
• Machine learning phases of matter
• Emergence of macroscopic directed motion in populations of motile colloids
• Meso-scale turbulence in living fluids
• Hydrodynamic fluctuations and instabilities in ordered suspensions of self-propelled particles
• A high-bias, low-variance introduction to Machine Learning for physicists
• Learning phase transitions by confusion
• Topology and Dynamics of Active Nematic Vesicles
• Rheology of Active-Particle Suspensions
• Spontaneous flow transition in active polar gels
• Defect annihilation and proliferation in active nematics
• Orientational order of motile defects in active nematics
• Transition from turbulent to coherent flows in confined three-dimensional active fluids
• Command of active matter by topological defects and patterns
• Defect dynamics in active nematics
• Lattice Boltzmann Simulations of Liquid Crystal Hydrodynamics
• Topological structure and dynamics of three dimensional active nematics
• Mutual Information, Neural Networks and the Renormalization Group
• Turbulent dynamics of epithelial cell cultures
• Tunable structure and dynamics of active liquid crystals
• Onset of meso-scale turbulence in living fluids
• Emergence of active nematic behaviour in monolayers of isotropic cells
• Data-driven quantitative modeling of bacterial active nematics
• An exact mapping between the Variational Renormalization Group and Deep Learning
• Universal scaling of active nematic turbulence
• Defect unbinding in active nematics
• Correlation lengths in hydrodynamic models of active nematics
• Interplay of Structure, Elasticity and Dynamics in Actin-Based Nematic Materials
• Sounds and hydrodynamics of polar active fluids
• Statistical Properties of Autonomous Flows in 2D Active Nematics
• Cross-talk between topological defects in different fields revealed by nematic microfluidics
• Hydrodynamics of Active Defects: from order to chaos to defect ordering
• Microfluidic control over topological states in channel-confined nematic flows
• The geometry of thresholdless active flow in nematic microfluidics

Cited by in corpus (25)

• Topological active matter
• Time-(ir)reversibility in active matter: from micro to macro
• Learning hydrodynamic equations for active matter from particle simulations and experiments
• Topological defects in solids with odd elasticity
• Learning to Control Active Matter
• Data-driven discovery of active nematic hydrodynamics
• Machine Learning of Implicit Combinatorial Rules in Mechanical Metamaterials
• Soft Metamaterials: Adaptation and Intelligence
• The 2024 Motile Active Matter Roadmap
• Designing, Synthesizing and Modeling Active Fluids
• Variational methods and deep Ritz method for active elastic solids
• Spontaneous rotation of active droplets in two and three dimensions
• Motor crosslinking augments elasticity in active nematics
• Hydrodynamic Enhancement of $p$-atic Defect Dynamics
• Data-driven model construction for anisotropic dynamics of active matter
• Machine Eye for Defects: Machine Learning-Based Solution to Identify and Characterize Topological Defects in Textured Images of Nematic Materials
• Theory of Nonequilibrium Multicomponent Coexistence
• Machine learning topological defects in confluent tissues
• Hermitian and non-Hermitian topology in active matter
• Theory of Nonequilibrium Coexistence with Coupled Conserved and Nonconserved Order Parameters
• Nematic liquid crystal flow driven by time-varying active surface anchoring
• Active viscoelastic nematics with partial degree of order
• A Hitchhiker's Guide To Active Motion
• Predicting Real-time Scientific Experiments Using Transformer models and Reinforcement Learning
• Irregular Metamaterial Networks