Routes to independent vessel networks: In air-breathing vertebrates, the circulatory and pulmonary systems contain separate networks of channels that intertwine but do not intersect with each other. Recreating such structures within cell-compatible materials has been a major challenge; even a single vasculature system can be a burden to create. Grigoryan et al. show that natural and synthetic food dyes can be used as photoabsorbers that enable stereolithographic production of hydrogels containing intricate and functional vascular architectures. Using this approach, they demonstrate functional vascular topologies for studies of fluid mixers, valves, intervascular transport, nutrient delivery, and host engraftment. Science , this issue p. 458

Bidirectional Encoder Representations from Transformers (BERT) and BERT-based approaches are the current state-of-the-art in many natural language processing (NLP) tasks; however, their application to document classification on long clinical texts is limited. In this work, we introduce four methods to scale BERT, which by default can only handle input sequences up to approximately 400 words long, to perform document classification on clinical texts several thousand words long. We compare these methods against two much simpler architectures - a word-level convolutional neural network and a hierarchical self-attention network - and show that BERT often cannot beat these simpler baselines when classifying MIMIC-III discharge summaries and SEER cancer pathology reports. In our analysis, we show that two key components of BERT - pretraining and WordPiece tokenization - may actually be inhibiting BERT's performance on clinical text classification tasks where the input document is several thousand words long and where correctly identifying labels may depend more on identifying a few key words or phrases rather than understanding the contextual meaning of sequences of text.

The behaviour of a single droplet in an immiscible external fluid, submitted to shear flow is investigated using numerical simulations. The surface of the droplet is modelled by a Boussinesq–Scriven constitutive law involving the interfacial viscosities and a constant surface tension. A numerical method using Loop subdivision surfaces to represent droplet interface is introduced. This method couples boundary element method for fluid flows and finite element method to take into account the stresses due to the surface dilational and shear viscosities and surface tension. Validation of the numerical scheme with respect to previous analytic and computational work is provided, with particular attention to the viscosity contrast and the shear and dilational viscosities characterized both by a Boussinesq number $B_{q}$ . Then, influence of equal surface viscosities on steady-state characteristics of a droplet in shear flow are studied, considering both small and large deformations and for a large range of bulk viscosity contrast. We find that small deformation analysis is surprisingly predictive at moderate and high surface viscosities. Equal surface viscosities decrease the Taylor deformation parameter and tank-treading angle and also strongly modify the dynamics of the droplet: when the Boussinesq number (surface viscosity) is large relative to the capillary number (surface tension), the droplet displays damped oscillations prior to steady-state tank-treading, reminiscent from the behaviour at large viscosity contrast. In the limit of infinite capillary number $Ca$ , such oscillations are permanent. The influence of surface viscosities on breakup is also investigated, and results show that the critical capillary number is increased. A diagram $(B_{q};Ca)$ of breakup is established with the same inner and outer bulk viscosities. Additionally, the separate roles of shear and dilational surface viscosity are also elucidated, extending results from small deformation analysis. Indeed, shear (dilational) surface viscosity increases (decreases) the stability of drops to breakup under shear flow. The steady-state deformation (Taylor parameter) varies nonlinearly with each Boussinesq number or a linear combination of both Boussinesq numbers. Finally, the study shows that for certain combinations of shear and dilational viscosities, drop deformation for a given capillary number is the same as in the case of a clean surface while the inclination angle varies.

Genesis of atherosclerotic lesions in the human arterial system is critically influenced by the fluid mechanics. Applying computational fluid dynamic tools based on accurate coronary physiology derived from conventional biplane angiogram data may be useful in guiding percutaneous coronary interventions. The primary objective of this study is to build and validate a computational framework for accurate personalized 3-dimensional hemodynamic simulation across the complete coronary arterial tree and demonstrate the influence of side branches on coronary hemodynamics by comparing shear stress between coronary models with and without these included. The proposed novel computational framework based on biplane angiography enables significant arterial circulation analysis. This study shows that models that take into account flow through all side branches are required for precise computation of shear stress and pressure gradient whereas models that have only a subset of side branches are inadequate for biomechanical studies as they may overestimate volumetric outflow and shear stress. This study extends the ongoing computational efforts and demonstrates that models based on accurate coronary physiology can improve overall fidelity of biomechanical studies to compute hemodynamic risk-factors.

Inspired by the evolution of biological systems, genetic algorithms have been applied to generate solutions for optimization problems in a variety of scientific and engineering disciplines. For a given problem, a suitable genome representation must be defined along with a mutation operator to generate subsequent generations. Unlike natural systems, which display a variety of complex rearrangements (e.g., mobile genetic elements), mutation for genetic algorithms commonly utilizes only random pointwise changes. Furthermore, generalizing beyond pointwise mutations poses a key difficulty as useful genome rearrangements depend on the representation and problem domain. To move beyond the limitations of manually defined pointwise changes, here we propose the use of techniques from masked language models to automatically generate mutations. As a first step, common subsequences within a given population are used to generate a vocabulary. The vocabulary is then used to tokenize each genome. A masked language model is trained on the tokenized data in order to generate possible rearrangements (i.e., mutations). In order to illustrate the proposed strategy, we use string representations of molecules and use a genetic algorithm to optimize for drug-likeness and synthesizability. Our results show that moving beyond random pointwise mutations accelerates genetic algorithm optimization.

We present a boundary condition scheme for the lattice Boltzmann method that has significantly improved stability for modeling turbulent flows while maintaining excellent parallel scalability. Simulations of a three-dimensional lid-driven cavity flow are found to be stable up to the unprecedented Reynolds number $\mathrm{Re}=5\times {10}^4$ for this setup. Excellent agreement with energy balance equations, computational and experimental results are shown. We quantify rises in the production of turbulence and turbulent drag, and determine peak locations of turbulent production.