Unsupervised learning in hexagonal boron nitride memristor-based spiking neural networks.

Resistive random access memory (RRAM) is an emerging non-volatile memory technology that can be used in neuromorphic computing hardware to exceed the limitations of traditional von Neumann architectures by merging processing and memory units. Two-dimensional (2D) materials with non-volatile switching behavior can be used as the switching layer of RRAMs, exhibiting superior behavior compared to conventional oxide-based devices. In this study, we investigate the electrical performance of 2D hexagonal boron nitride (h-BN) memristors towards their implementation in spiking neural networks (SNN). Based on experimental behavior of the h-BN memristors as artificial synapses, we simulate the implementation of unsupervised learning in SNN for image classification on the Modified National Institute of Standards and Technology dataset. Additionally, we propose a simple spike-timing-dependent-plasticity (STDP)-based dropout technique to enhance the recognition rate in h-BN memristor-based SNN. Our results demonstrate the viability of using 2D-material-based memristors as artificial synapses to perform unsupervised learning in SNN using hardware-friendly methods for online learning.

Decoding the hydrodynamic properties of microscale helical propellers from Brownian fluctuations.

The complex motility of bacteria, ranging from single-swimmer behaviors such as chemotaxis to collective dynamics, including biofilm formation and active matter phenomena, is driven by their microscale propellers. Despite extensive study of swimming flagellated bacteria, the hydrodynamic properties of their helical-shaped propellers have never been directly measured. The primary challenges to directly studying microscale propellers are 1) their small size and fast, correlated motion, 2) the necessity of controlling fluid flow at the microscale, and 3) isolating the influence of a single propeller from a propeller bundle. To solve the outstanding problem of characterizing the hydrodynamic properties of these propellers, we adopt a dual statistical viewpoint that connects to the hydrodynamics through the fluctuation-dissipation theorem (FDT). We regard the propellers as colloidal particles and characterize their Brownian fluctuations, described by 21 diffusion coefficients for translation, rotation, and correlated translation-rotation in a static fluid. To perform this measurement, we applied recent advances in high-resolution oblique plane microscopy to generate high-speed volumetric movies of fluorophore-labeled, freely diffusing Escherichia coli flagella. Analyzing these movies with a bespoke helical single-particle tracking algorithm, we extracted trajectories, calculated the full set of diffusion coefficients, and inferred the average propulsion matrix using a generalized Einstein relation. Our results provide a direct measurement of a microhelix's propulsion matrix and validate proposals that the flagella are highly inefficient propellers, with a maximum propulsion efficiency of less than 3%. Our approach opens broad avenues for studying the motility of particles in complex environments where direct hydrodynamic approaches are not feasible.