Engineering tissues utilizing biodegradable polymer matrices is a promising approach to the treatment of a number of diseases. However, processing techniques utilized to fabricate these matrices typically involve organic solvents and/or high temperatures. Here we describe a process for fabricating matrices without the use of organic solvents and/or elevated temperatures. Disks comprised of polymer [e.g., poly (D,L-lactic-co-glycolic acid)] and NaCl particles were compression molded at room temperature and subsequently allowed to equilibrate with high pressure CO 2 gas (800 psi). Creation of a thermodynamic instability led to the nucleation and growth of gas pores in the polymer particles, resulting in the expansion of the polymer particles. The polymer particles fused to form a continuous matrix with entrapped salt particles. The NaCl particles subsequently were leached to yield macropores within the poly-mer matrix. The overall porosity and level of pore connectivity were regulated by the ratio of polymer/salt particles and the size of salt particles. Both the compressive modulus (159 ± 130 kPa versus 289 ± 25 kPa) and the tensile modulus (334 ± 52 kPa versus 1100 ± 236 kPa) of the matrices formed with this approach were significantly greater than those formed with a standard solvent casting/particulate leaching process. The utility of these matrices was demonstrated by engineering smooth muscle tissue in vitro with them. This novel process, a combination of high pressure gas foaming and particulate leaching techniques, allows one to fabricate matrices with a well controlled porosity and pore structure. This process avoids the potential negatives associated with the use of high temperatures and/or organic solvents in biomaterials processing.

This paper presents a novel generative model to synthesize fluid simulations from a set of reduced parameters. A convolutional neural network is trained on a collection of discrete, parameterizable fluid simulation velocity fields. Due to the capability of deep learning architectures to learn representative features of the data, our generative model is able to accurately approximate the training data set, while providing plausible interpolated in‐betweens. The proposed generative model is optimized for fluids by a novel loss function that guarantees divergence‐free velocity fields at all times. In addition, we demonstrate that we can handle complex parameterizations in reduced spaces, and advance simulations in time by integrating in the latent space with a second network. Our method models a wide variety of fluid behaviors, thus enabling applications such as fast construction of simulations, interpolation of fluids with different parameters, time re‐sampling, latent space simulations, and compression of fluid simulation data. Reconstructed velocity fields are generated up to 700× faster than re‐simulating the data with the underlying CPU solver, while achieving compression rates of up to 1300×.

There is accumulating circumstantial evidence suggesting that endothelial cell dysfunction contributes to the "no-reflow" phenomenon in postischemic kidneys. Here, we demonstrated the vulnerability of in vitro, ex vivo, and in vivo endothelial cells exposed to pathophysiologically relevant insults, such as oxidative and nitrosative stress or ischemia. All of these stimuli compromised the integrity of the endothelial lining. Next, we performed minimally invasive intravital microscopy of blood flow in peritubular capillaries, which provided direct evidence of the existence of the no-reflow phenomenon, attributable, at least in part, to endothelial injury. In an attempt to ameliorate the hemodynamic consequences of lost endothelial integrity, we transplanted endothelial cells or surrogate cells expressing endothelial nitric oxide synthase into rats subjected to renal artery clamping. Implantation of endothelial cells or their surrogates expressing functional endothelial nitric oxide synthase in the renal microvasculature resulted in a dramatic functional protection of ischemic kidneys. These observations strongly suggest that endothelial cell dysfunction is the primary cause of the no-reflow phenomenon, which, when ameliorated, results in prevention of renal injury seen in acute renal failure.

The engineering of functional smooth muscle (SM) tissue is critical if one hopes to successfully replace the large number of tissues containing an SM component with engineered equivalents. This study reports on the effects of SM cell (SMC) seeding and culture conditions on the cellularity and composition of SM tissues engineered using biodegradable matrices (5 x 5 mm, 2-mm thick) of polyglycolic acid (PGA) fibers. Cells were seeded by injecting a cell suspension into polymer matrices in tissue culture dishes (static seeding), by stirring polymer matrices and a cell suspension in spinner flasks (stirred seeding), or by agitating polymer matrices and a cell suspension in tubes with an orbital shaker (agitated seeding). The density of SMCs adherent to these matrices was a function of cell concentration in the seeding solution, but under all conditions a larger number (approximately 1 order of magnitude) and more uniform distribution of SMCs adherent to the matrices were obtained with dynamic versus static seeding methods. The dynamic seeding methods, as compared to the static method, also ultimately resulted in new tissues that had a higher cellularity, more uniform cell distribution, and greater elastin deposition. The effects of culture conditions were next studied by culturing cell-polymer constructs in a stirred bioreactor versus static culture conditions. The stirred culture of SMC-seeded polymer matrices resulted in tissues with a cell density of 6.4 +/- 0.8 x 10(8) cells/cm3 after 5 weeks, compared to 2.0 +/- 1.1 x 10(8) cells/cm3 with static culture. The elastin and collagen synthesis rates and deposition within the engineered tissues were also increased by culture in the bioreactors. The elastin content after 5-week culture in the stirred bioreactor was 24 +/- 3%, and both the elastin content and the cellularity of these tissues are comparable to those of native SM tissue. New tissues were also created in vivo when dynamically seeded polymer matrices were implanted in rats for various times. In summary, the system defined by these studies shows promise for engineering a tissue comparable in many respects to native SM. This engineered tissue may find clinical applications and provide a tool to study molecular mechanisms in vascular development.

Knowledge tracing, the act of modeling a student's knowledge through learning activities, is an extensively studied problem in the field of computer-aided education. Armed with attention mechanisms focusing on relevant information for target prediction, recurrent neural networks and Transformer-based knowledge tracing models have outperformed traditional approaches such as Bayesian knowledge tracing and collaborative filtering. However, the attention mechanisms of current state-of-the-art knowledge tracing models share two limitations. Firstly, the models fail to leverage deep self-attentive computations for knowledge tracing. As a result, they fail to capture complex relations among exercises and responses over time. Secondly, appropriate features for constructing queries, keys and values for the self-attention layer for knowledge tracing have not been extensively explored. The usual practice of using exercises and interactions (exercise-response pairs), as queries and keys/values, respectively, lacks empirical support.In this paper, we propose a novel Transformer-based model for knowledge tracing, SAINT: Separated Self-AttentIve Neural Knowledge Tracing. SAINT has an encoder-decoder structure where the exercise and response embedding sequences separately enter, respectively, the encoder and the decoder. The encoder applies self-attention layers to the sequence of exercise embeddings, and the decoder alternately applies selfattention layers and encoder-decoder attention layers to the sequence of response embeddings. This separation of input allows us to stack attention layers multiple times, resulting in an improvement in area under receiver operating characteristic curve (AUC). To the best of our knowledge, this is the first work to suggest an encoder-decoder model for knowledge tracing that applies deep self-attentive layers to exercises and responses separately.We empirically evaluate SAINT on a large-scale knowledge tracing dataset, EdNet, collected by an active mobile education

Advances in Artificial Intelligence in Education (AIEd) and the ever-growing scale of Interactive Educational Systems (IESs) have led to the rise of data-driven approaches for knowledge tracing and learning path recommendation. Unfortunately, collecting student interaction data is challenging and costly. As a result, there is no public largescale benchmark dataset reflecting the wide variety of student behaviors observed in modern IESs. Although several datasets, such as ASSISTments, Junyi Academy, Synthetic and STATICS are publicly available and widely used, they are not large enough to leverage the full potential of state-of-the-art data-driven models. Furthermore, the recorded behavior is limited to question-solving activities. To this end, we introduce EdNet, a large-scale hierarchical dataset of diverse student activities collected by Santa, a multi-platform self-study solution equipped with an artificial intelligence tutoring system. EdNet contains 131,417,236 interactions from 784,309 students collected over more than 2 years, making it the largest public IES dataset released to date. Unlike existing datasets, EdNet records a wide variety of student actions ranging from questionsolving to lecture consumption to item purchasing. Also, EdNet has a hierarchical structure which divides the student actions into 4 different levels of abstractions. The features of EdNet are domain-agnostic, allowing EdNet to be easily extended to different domains. The dataset is publicly released for research purposes. We plan to host challenges in multiple AIEd tasks with EdNet to provide a common ground for the fair comparison between different state-of-the-art models and to encourage the development of practical and effective methods.

The reconstruction of urinary tissues often employs various types of biomaterials, and adequate material biocompatibility is essential for the successful reconstruction of urinary tissues. In this study we utilized a primary normal human urothelial cell culture system to evaluate the in vitro biocompatibility of a number of naturally derived biomaterials [i.e., bladder submucosa, small intestinal submucosa, collagen, and alginate] and polymeric biomaterials [i.e., poly(glycolic acid), poly(L-lactic acid), poly(lactic-coglycolic acid), and silicone] that are either experimentally or clinically used in urinary reconstructive surgery. To determine the cytotoxic and bioactive effects of these biomaterials, the cell viability, metabolic activity, apoptotic properties, and DNA-synthesis activity were measured with four types of assays [Neutral Red, 3-(4,5-dimethylthiazol-2-yl)-2,5diphenyl tetrazolium bromide, apoptotic activity, and tritiated thymidine incorporation assays] using extract and direct contact methods. Most of the biomaterials tested did not induce significant cytotoxic effects and exhibited normal metabolic function and cell growth in vitro. This normal primary human urothelial cell culture model is suitable for in vitro biocompatibility assessments and is able to provide information on the cell-biomaterial interactions and the ability of biomaterials to support bioactive cell functions.

No abstract