Sho Takatori

All Publications

• Learning continuum-level closures for control of interacting active particles. The Journal of chemical physics Quah, T., Takatori, S. C., Rawlings, J. B. 2026; 164 (4)

  Abstract

  Active matter swarms-collectives of self-propelled particles that can self-assemble, ferry microscopic cargo, or endow materials with dynamic properties-remain hard to steer. In crowded systems, tracking or controlling individual agents becomes challenging, so strategies must operate on macroscopic fields like particle density. Yet predicting how density evolves is difficult because of inter-agent interactions. For model-based feedback control methods-such as Model Predictive Control (MPC)-fast, accurate, and differentiable models are crucial. Detailed agent-based simulations are too slow, necessitating coarse-grained continuum models. However, constructing accurate closures-approximations that express the effects of unresolved microscopic states (e.g., agent positions) on continuum dynamics in terms of the modeled continuum fields (e.g., density)-is challenging for active matter swarms. We present a learning-for-control framework that learns continuum closures from agent simulations, demonstrated with active Brownian particles under a controllable external field. Our Universal Differential Equation (UDE) framework represents the continuum as an advection-diffusion equation. A neural operator learns the advection term, providing closure relations for microscopic effects such as self-propulsion, interactions, and external-field responses. This UDE approach, embedding universal function approximators in differential equations, ensures adherence to physical laws (e.g., conservation) while learning complex dynamics directly from data. We embed this learned continuum model into MPC for precise agent-simulation control. We demonstrate this framework's capabilities by dynamically exchanging particle densities between two groups and by simultaneously controlling particle density and mean flux to follow a prescribed sinusoidal profile. These results highlight the framework's potential to control complex active-matter dynamics, foundational for programmable materials.

  View details for DOI 10.1063/5.0300697

  View details for PubMedID 41614966

• Motility Modulates the Partitioning of Bacteria in Aqueous Two-Phase Systems. Physical review letters Cheon, J., Choi, K. H., Modica, K. J., Mitchell, R. J., Takatori, S. C., Jeong, J. 2025; 135 (12): 128401

  Abstract

  We study the partitioning of motile bacteria in an aqueous two-phase mixture of dextran (DEX) and polyethylene glycol (PEG), which can phase separate into DEX-rich and PEG-rich phases. While nonmotile bacteria partition exclusively into the DEX-rich phase in all conditions tested, we observed that motile bacteria penetrate the soft DEX-PEG interface and partition variably among the two phases. For our model organism Bacillus subtilis, the fraction of motile bacteria in the DEX-rich phase increased from 0.58 to 1 as we increased the DEX composition within the two-phase region. We hypothesized that the chemical affinity between DEX and the bacteria cell wall acts to weakly confine the bacteria within the DEX-rich phase; however, motility can generate sufficient mechanical forces to overcome the soft confinement and propel the bacteria into the PEG-rich phase. Using optical tweezers to drag a bacterium across the DEX-PEG interface, we demonstrate that the overall bacteria partitioning is determined by a competition between the interfacial forces and bacterial propulsive forces. Our measurements are supported by a theoretical model of dilute active rods embedded within a periodic soft confinement potential.

  View details for DOI 10.1103/6gm5-cnv1

  View details for PubMedID 41046403

• Phase field model for viscous inclusions in anisotropic networks. Soft matter Gubbala, A., Jena, A. M., Arnold, D. P., Takatori, S. C. 2025

  Abstract

  The growth of viscous two-dimensional lipid domains in contact with a viscoelastic actin network was recently shown to exhibit unusual lipid domain ripening due to the geometry and anisotropy of the actin network [Arnold & Takatori. Langmuir. 40, 26570-26578 (2024)]. In this work, we interpret previous experimental results on lipid membrane-actin composites with a theoretical model that combines the Cahn-Hilliard and Landau-de Gennes liquid crystal theory. In our model, we incorporate fiber-like characteristics of actin filaments and bundles through a nematic order parameter, and elastic anisotropy through cubic nematic gradients. Numerical simulations qualitatively agree with experimental observations, by reproducing the competition between the thermodynamic forces that coarsen lipid domains versus the elastic forces generated by the surrounding actin network that resist domain coarsening. We observe a decrease in the growth of domain sizes, finding R(t) talpha with alpha < 1/4 for different actin network stiffnesses, in sharp contrast to the t1/3 scaling for diffusive growth of domains in the absence of the actin network. Our findings may serve as a foundation for future developments in modeling elastic ripening in complex systems.

  View details for DOI 10.1039/d5sm00478k

  View details for PubMedID 40692432

• Direct experimental measurement of many-body hydrodynamic interactions with optical tweezers PHYSICAL REVIEW FLUIDS Kim, D., Nagella, S. G., Choi, K., Takatori, S. C. 2025; 10 (6)

  View details for DOI 10.1103/PhysRevFluids.10.064301

  View details for Web of Science ID 001529022100003

• Feedback Control of Active Matter ANNUAL REVIEW OF CONDENSED MATTER PHYSICS Takatori, S. C., Quah, T., Rawlings, J. B. 2025; 16: 319-341

  View details for DOI 10.1146/annurev-conmatphys-042424-043926

  View details for Web of Science ID 001441665900016