Adaptive network approach to exploration-exploitation trade-off in reinforcement learning.

A foundational machine-learning architecture is reinforcement learning, where an outstanding problem is achieving an optimal balance between exploration and exploitation. Specifically, exploration enables the agents to discover optimal policies in unknown domains of the environment for gaining potentially large future rewards, while exploitation relies on the already acquired knowledge to maximize the immediate rewards. We articulate an approach to this problem, treating the dynamical process of reinforcement learning as a Markov decision process that can be modeled as a nondeterministic finite automaton and defining a subset of states in the automaton to represent the preference for exploring unknown domains of the environment.

Structure-function relationship of phase-separated liposomes containing diacylglycerol analogues.

The composition and morphology of lipid-based nanoparticles can influence their overall behavior. Previously, we demonstrated that phase separation in liposomes composed of DSPC and a diacylglycerol lipid analogue (DOaG) drives the biodistribution towards a specific subset of endothelial cells in zebrafish embryos. In the absence of traditional targeting functionalities (, antibodies, ligands), this selectivity is mediated solely by the unique liposome morphology and composition, characterized by a DOaG-rich lipid droplet within the DSPC-rich phospholipid bilayer.

Phase-Separated Lipid-Based Nanoparticles: Selective Behavior at the Nano-Bio Interface.

The membrane-protein interface on lipid-based nanoparticles influences their in vivo behavior. Better understanding may evolve current drug delivery methods toward effective targeted nanomedicine. Previously, the cell-selective accumulation of a liposome formulation in vivo is demonstrated, through the recognition of lipid phase-separation by triglyceride lipases.

Model-free tracking control of complex dynamical trajectories with machine learning.

Nonlinear tracking control enabling a dynamical system to track a desired trajectory is fundamental to robotics, serving a wide range of civil and defense applications. In control engineering, designing tracking control requires complete knowledge of the system model and equations. We develop a model-free, machine-learning framework to control a two-arm robotic manipulator using only partially observed states, where the controller is realized by reservoir computing.

Phase-Separated Liposomes Hijack Endogenous Lipoprotein Transport and Metabolism Pathways to Target Subsets of Endothelial Cells In Vivo.

Plasma lipid transport and metabolism are essential to ensure correct cellular function throughout the body. Dynamically regulated in time and space, the well-characterized mechanisms underpinning plasma lipid transport and metabolism offers an enticing, but as yet underexplored, rationale to design synthetic lipid nanoparticles with inherent cell/tissue selectivity. Herein, a systemically administered liposome formulation, composed of just two lipids, that is capable of hijacking a triglyceride lipase-mediated lipid transport pathway resulting in liposome recognition and uptake within specific endothelial cell subsets is described.

Strategy for optimizing experimental settings for studying low atomic number colloidal assemblies using liquid phase scanning transmission electron microscopy.

Observing processes of nanoscale materials of low atomic number is possible using liquid phase electron microscopy (LP-EM). However, the achievable spatial resolution (d) is limited by radiation damage. Here, we examine a strategy for optimizing LP-EM experiments based on an analytical model and experimental measurements, and develop a method for quantifying image quality at ultra low electron dose D using scanning transmission electron microscopy (STEM).

From binary AB to ternary ABC supraparticles.

Control over the assembly and morphology of nanoscale functional building blocks is of great importance to hybrid and porous nanomaterials. In this paper, by combining different types of spherical nanoparticles with different size ratios in a hierarchical assembly process which allows us to control the final structure of multi-component assemblies, we discuss self-assembly of an extensive range of supraparticles, labelled as AB particles, and an extension to novel ternary particles, labelled as ABC particles. For supraparticles, the organization of small nanoparticles is known to be inherently related to the size ratio of building blocks.

Anionic Lipid Nanoparticles Preferentially Deliver mRNA to the Hepatic Reticuloendothelial System.

Lipid nanoparticles (LNPs) are the leading nonviral technologies for the delivery of exogenous RNA to target cells in vivo. As systemic delivery platforms, these technologies are exemplified by Onpattro, an approved LNP-based RNA interference therapy, administered intravenously and targeted to parenchymal liver cells. The discovery of systemically administered LNP technologies capable of preferential RNA delivery beyond hepatocytes has, however, proven more challenging.

Flower-Like Colloidal Particles through Precipitation Polymerization of Redox-Responsive Liquid Crystals.

We report on the synthesis of monodisperse, flower-like, liquid crystalline (LC) polymer particles by precipitation polymerization of a LC mixture consisting of benzoic acid-functionalized acrylates and disulfide-functionalized diacrylates. Introduction of a minor amount of redox-responsive disulfide-functionalized diacrylates (≤10 wt %) induced the formation of flower-like shapes. The shape of the particles can be tuned from flower- to disk-like to spherical by elevating the polymerization temperature.

Monodisperse Liquid Crystalline Polymer Shells with Programmable Alignment and Shape Prepared by Seeded Dispersion Polymerization.

Monodisperse, micrometer-sized liquid crystalline (LC) shells are prepared by seeded dispersion polymerization. After polymerizing LC monomer mixtures in the presence of non-crosslinked polymer seeds, hollow LC polymer shells with programmable alignment and shape are prepared by removing the seeds. The LC alignment in the LC polymer shells can be easily manipulated by the polymer seeds, as a radial alignment is observed with amorphous poly(phenyl methacrylate) seeds and a bipolar alignment is observed with bipolar LC polymer seeds.

Building Reversible Nanoraspberries.

The adsorption mechanism of small positively charged silica nanoparticles (SiO NPs) onto larger polystyrene latex nanoparticles (PSL NPs) forming hybrid particles was studied. CryoTEM showed the morphology of these supraparticles to be raspberry-like. After surface modification of the SiO NPs, the optimum pH regime to initiate the formation of nanoraspberries was determined.

Spontaneous organization of supracolloids into three-dimensional structured materials.

Periodic nano- or microscale structures are used to control light, energy and mass transportation. Colloidal organization is the most versatile method used to control nano- and microscale order, and employs either the enthalpy-driven self-assembly of particles at a low concentration or the entropy-driven packing of particles at a high concentration. Nonetheless, it cannot yet provide the spontaneous three-dimensional organization of multicomponent particles at a high concentration.

Designing stable, hierarchical peptide fibers from block co-polypeptide sequences.

Natural materials, such as collagen, can assemble with multiple levels of organization in solution. Achieving a similar degree of control over morphology, stability and hierarchical organization with equilibrium synthetic materials remains elusive. For the assembly of peptidic materials the process is controlled by a complex interplay between hydrophobic interactions, electrostatics and secondary structure formation.

Morphology Control of Liposome - RAFT Oligomer Precursors to Complex Polymer Nanostructures.

Different types of butyl acrylate (BA)--acrylic acid (AA) oligomers were synthesized via reversible addition-fragmentation chain transfer (RAFT) polymerization and mixed with extruded 200 nm dimethyldioctadecylammonium bromide vesicles. The resulting precursor structures form the basis for subsequent vesicle-templated polymerizations. Systematic variations in temperature, pH, oligomer length, and oligomer composition and their effects on precursor morphology were studied.

Polypeptide Nanoparticles Obtained from Emulsion Polymerization of Amino Acid -Carboxyanhydrides.

Polypeptide nanoparticles were obtained by the miniemulsion polymerization of -(-nitrobenzyl)-l-cysteine (NBC) -carboxyanhydride (NCA). Through process optimization, reaction conditions were identified that allowed the polymerization of the water sensitive NCA to yield nanoparticles of about 220 nm size. Subsequent UV-irradiation of the nanoparticle emulsions caused the removal of the nitrobenzyl group and particle cross-linking through disulfide bond formation accompanied by the shrinkage of the particles.

Native Chemical Ligation for Cross-Linking of Flower-Like Micelles.

In this study, native chemical ligation (NCL) was used as a selective cross-linking method to form core-cross-linked thermosensitive polymeric micelles for drug delivery applications. To this end, two complementary ABA triblock copolymers having polyethylene glycol (PEG) as midblock were synthesized by atom transfer radical polymerization (ATRP). The thermosensitive poly isopropylacrylamide (PNIPAM) outer blocks of the polymers were copolymerized with either N-(2-hydroxypropyl)methacrylamide-cysteine (HPMA-Cys), P(NIPAM- co-HPMA-Cys)-PEG-P(NIPAM- co-HPMA-Cys) (PNC) or N-(2-hydroxypropyl)methacrylamide-ethylthioglycolate succinic acid (HPMA-ETSA), P(NIPAM- co-HPMA-ETSA)-PEG-P(NIPAM- co-HPMA-ETSA) (PNE).

CryoTEM as an Advanced Analytical Tool for Materials Chemists.

Morphology plays an essential role in chemistry through the segregation of atoms and/or molecules into different phases, delineated by interfaces. This is a general process in materials synthesis and exploited in many fields including colloid chemistry, heterogeneous catalysis, and functional molecular systems. To rationally design complex materials, we must understand and control morphology evolution.

Bimodal Latex Effect on Spin-Coated Thin Conductive Polymer-Single-Walled Carbon Nanotube Layers.

We synthesize two differently sized poly(methyl methacrylate-co-tert-butyl acrylate) latexes by emulsion polymerization and mix these with a sonicated single-walled carbon nanotube (SWCNT) dispersion, in order to prepare 3% SWCNT composite mixtures. We spin-coat these mixtures at various spin-speed rates and spin times over a glass substrate, producing a thin, transparent, solid, conductive layer. Keeping the amount of SWCNTs constant, we vary the weight fraction of our smaller 30-nm latex particles relative to the larger 70-nm-sized ones.