Biomaterial scaffolds are central to many regenerative strategies as they create a space for infiltration of host tissue and provide a platform to deliver growth factors and progenitor cells. However, biomaterial implantation results in an unavoidable inflammatory response, which can impair tissue regeneration and promote loss or dysfunction of transplanted cells. We investigated localized TGF-β1 delivery to modulate this immunological environment around scaffolds and transplanted cells. TGF-β1 was delivered from layered scaffolds, with protein entrapped within an inner layer and outer layers designed for cell seeding and host tissue integration. Scaffolds were implanted into the epididymal fat pad, a site frequently used for cell transplantation. Expression of cytokines TNF-a, IL-12, and MCP-1 were decreased by at least 40% for scaffolds releasing TGF-β1 relative to control scaffolds. This decrease in inflammatory cytokine production corresponded to a 60% decrease in leukocyte infiltration. Transplantation of islets into diabetic mice on TGF-β1 scaffolds significantly improved the ability of syngeneic islets to control blood glucose levels within the first week of transplant and delayed rejection of allogeneic islets. Together, these studies emphasize the ability of localized TGF-β1 delivery to modulate the immune response to biomaterial implants and enhance cell function in cell-based therapies.

Current approaches to single-cell transcriptomic analysis are computationally intensive and require assay-specific modeling, which limits their scope and generality. We propose a novel method that compares and clusters cells based on their transcript-compatibility read counts rather than on the transcript or gene quantifications used in standard analysis pipelines. In the reanalysis of two landmark yet disparate single-cell RNA-seq datasets, we show that our method is up to two orders of magnitude faster than previous approaches, provides accurate and in some cases improved results, and is directly applicable to data from a wide variety of assays.

No abstract

Neural networks have been used prominently in several machine learning and statistics applications. In general, the underlying optimization of neural networks is non-convex which makes their performance analysis challenging. In this paper, we take a novel approach to this problem by asking whether one can constrain neural network weights to make its optimization landscape have good theoretical properties while at the same time, be a good approximation for the unconstrained one. For two-layer neural networks, we provide affirmative answers to these questions by introducing Porcupine Neural Networks (PNNs) whose weight vectors are constrained to lie over a finite set of lines. We show that most local optima of PNN optimizations are global while we have a characterization of regions where bad local optimizers may exist. Moreover, our theoretical and empirical results suggest that an unconstrained neural network can be approximated using a polynomially-large PNN.

Standardized evaluation measures have aided in the progress of machine learning approaches in disciplines such as computer vision and machine translation. In this paper, we make the case that robotic learning would also benefit from benchmarking, and present the "REPLAB" platform for benchmarking vision-based manipulation tasks. REPLAB is a reproducible and self-contained hardware stack (robot arm, camera, and workspace) that costs about 2000 USD, occupies a cuboid of size 70x40x60 cm, and permits full assembly within a few hours. Through this low-cost, compact design, REPLAB aims to drive wide participation by lowering the barrier to entry into robotics and to enable easy scaling to many robots. We envision REPLAB as a framework for reproducible research across manipulation tasks, and as a step in this direction, we define a template for a grasping benchmark consisting of a task definition, evaluation protocol, performance measures, and a dataset of 92k grasp attempts. We implement, evaluate, and analyze several previously proposed grasping approaches to establish baselines for this benchmark. Finally, we also implement and evaluate a deep reinforcement learning approach for 3D reaching tasks on our REPLAB platform. Project page with assembly instructions, code, and videos: https://goo.gl/5F9dP4.