Learning to Control Active Matter

Abstract

The study of active matter has revealed novel non-equilibrium collective behaviors, illustrating their potential as a new materials platform. However, most works treat active matter as unregulated systems with uniform microscopic energy input, which we refer to as activity. In contrast, functionality in biological materials results from regulating and controlling activity locally over space and time, as has only recently become experimentally possible for engineered active matter. Designing functionality requires navigation of the high dimensional space of spatio-temporal activity patterns, but brute force approaches are unlikely to be successful without system-specific intuition. Here, we apply reinforcement learning to the task of inducing net transport in a specific direction for a simulated system of Vicsek-like self-propelled disks using a spotlight that increases activity locally. The resulting time-varying patterns of activity learned exploit the distinct physics of the strong and weak coupling regimes. Our work shows how reinforcement learning can reveal physically interpretable protocols for controlling collective behavior in non-equilibrium systems.

arXiv comment11 pages, 7 figures

References in corpus (29)

• Novel type of phase transition in a system of self-driven particles
• Motility-Induced Phase Separation
• Microswimmers in Patterned Environments
• The 2019 Motile Active Matter Roadmap
• Active Brownian Motion Tunable by Light
• Flow Navigation by Smart Microswimmers via Reinforcement Learning
• Phase transitions in systems of self-propelled agents and related network models
• Dynamic density shaping of light driven bacteria
• Synchronised Swimming of Two Fish
• Controlling Organization and Forces in Active Matter Through Optically-Defined Boundaries
• Structuring Stress for Active Materials Control
• Learning to school in the presence of hydrodynamic interactions
• Light-induced self-assembly of active rectification devices
• Drift without flux: Brownian walker with a space dependent diffusion coefficient
• Optimal run-and-tumble based transportation of a Janus particle with active steering
• Machine learning active-nematic hydrodynamics
• Shape and Displacement Fluctuations in Soft Vesicles Filled by Active Particles
• Cooperation of dual modes of cell motility promotes epithelial stress relaxation to accelerate wound healing
• Optimal Control of Active Nematics
• Hydrodynamics of Active Defects: from order to chaos to defect ordering
• Chemotaxis of artificial microswimmers in active density waves
• Learning to flock through reinforcement
• Swarming in the Dirt: Ordered Flocks with Quenched Disorder
• Breakdown of Ergodicity and Self-Averaging in Polar Flocks with Quenched Disorder
• Smart Inertial Particles
• Vesicles with internal active filaments: self-organized propulsion controls shape, motility, and dynamical response
• Taxis of Artificial Swimmers in a Spatio-Temporally Modulated Activation Medium
• Boids in a Loop: Self-Propelled particles within a Flexible Boundary
• Coordinated Crawling via Reinforcement Learning

Cited by in corpus (17)

• Learning without neurons in physical systems
• Active matter under control: Insights from response theory
• Machine Learning of Implicit Combinatorial Rules in Mechanical Metamaterials
• Design rules for controlling active topological defects
• Designing, Synthesizing and Modeling Active Fluids
• Probing the Theoretical and Computational Limits of Dissipative Design
• Hunting active Brownian particles: Learning optimal behavior
• Inertial self-propelled particles in anisotropic environments
• Optimal Control in Soft and Active Matter
• Dynamic control of self-assembly of quasicrystalline structures through reinforcement learning
• Closed-loop control of active nematic flows
• Mechanochemical Topological Defects in an Active Nematic
• Kinetic theory of decentralized learning for smart active matter
• Reinforcement Learning for Active Matter
• Tailoring the overlap distribution in driven mean-field spin models
• Active Matter as a framework for living systems-inspired Robophysics
• Physics-informed Dyna-Style Model-Based Deep Reinforcement Learning for Dynamic Control