Showing papers by "Xi'an Jiaotong University published in 2017"

Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising

Abstract: The discriminative model learning for image denoising has been recently attracting considerable attentions due to its favorable denoising performance. In this paper, we take one step forward by investigating the construction of feed-forward denoising convolutional neural networks (DnCNNs) to embrace the progress in very deep architecture, learning algorithm, and regularization method into image denoising. Specifically, residual learning and batch normalization are utilized to speed up the training process as well as boost the denoising performance. Different from the existing discriminative denoising models which usually train a specific model for additive white Gaussian noise at a certain noise level, our DnCNN model is able to handle Gaussian denoising with unknown noise level (i.e., blind Gaussian denoising). With the residual learning strategy, DnCNN implicitly removes the latent clean image in the hidden layers. This property motivates us to train a single DnCNN model to tackle with several general image denoising tasks, such as Gaussian denoising, single image super-resolution, and JPEG image deblocking. Our extensive experiments demonstrate that our DnCNN model can not only exhibit high effectiveness in several general image denoising tasks, but also be efficiently implemented by benefiting from GPU computing.

Channel Pruning for Accelerating Very Deep Neural Networks

Abstract: In this paper, we introduce a new channel pruning method to accelerate very deep convolutional neural networks. Given a trained CNN model, we propose an iterative two-step algorithm to effectively prune each layer, by a LASSO regression based channel selection and least square reconstruction. We further generalize this algorithm to multi-layer and multi-branch cases. Our method reduces the accumulated error and enhance the compatibility with various architectures. Our pruned VGG-16 achieves the state-of-the-art results by 5× speed-up along with only 0.3% increase of error. More importantly, our method is able to accelerate modern networks like ResNet, Xception and suffers only 1.4%, 1.0% accuracy loss under 2× speedup respectively, which is significant.

A Survey on Internet of Things: Architecture, Enabling Technologies, Security and Privacy, and Applications

Abstract: Fog/edge computing has been proposed to be integrated with Internet of Things (IoT) to enable computing services devices deployed at network edge, aiming to improve the user’s experience and resilience of the services in case of failures. With the advantage of distributed architecture and close to end-users, fog/edge computing can provide faster response and greater quality of service for IoT applications. Thus, fog/edge computing-based IoT becomes future infrastructure on IoT development. To develop fog/edge computing-based IoT infrastructure, the architecture, enabling techniques, and issues related to IoT should be investigated first, and then the integration of fog/edge computing and IoT should be explored. To this end, this paper conducts a comprehensive overview of IoT with respect to system architecture, enabling technologies, security and privacy issues, and present the integration of fog/edge computing and IoT, and applications. Particularly, this paper first explores the relationship between cyber-physical systems and IoT, both of which play important roles in realizing an intelligent cyber-physical world. Then, existing architectures, enabling technologies, and security and privacy issues in IoT are presented to enhance the understanding of the state of the art IoT development. To investigate the fog/edge computing-based IoT, this paper also investigate the relationship between IoT and fog/edge computing, and discuss issues in fog/edge computing-based IoT. Finally, several applications, including the smart grid, smart transportation, and smart cities, are presented to demonstrate how fog/edge computing-based IoT to be implemented in real-world applications.

Design of N-Coordinated Dual-Metal Sites: A Stable and Active Pt-Free Catalyst for Acidic Oxygen Reduction Reaction

Abstract: We develop a host-guest strategy to construct an electrocatalyst with Fe-Co dual sites embedded on N-doped porous carbon and demonstrate its activity for oxygen reduction reaction in acidic electrolyte. Our catalyst exhibits superior oxygen reduction reaction performance, with comparable onset potential (Eonset, 1.06 vs 1.03 V) and half-wave potential (E1/2, 0.863 vs 0.858 V) than commercial Pt/C. The fuel cell test reveals (Fe,Co)/N-C outperforms most reported Pt-free catalysts in H2/O2 and H2/air. In addition, this cathode catalyst with dual metal sites is stable in a long-term operation with 50 000 cycles for electrode measurement and 100 h for H2/air single cell operation. Density functional theory calculations reveal the dual sites is favored for activation of O-O, crucial for four-electron oxygen reduction.

Channel Pruning for Accelerating Very Deep Neural Networks

Abstract: In this paper, we introduce a new channel pruning method to accelerate very deep convolutional neural networks.Given a trained CNN model, we propose an iterative two-step algorithm to effectively prune each layer, by a LASSO regression based channel selection and least square reconstruction. We further generalize this algorithm to multi-layer and multi-branch cases. Our method reduces the accumulated error and enhance the compatibility with various architectures. Our pruned VGG-16 achieves the state-of-the-art results by 5x speed-up along with only 0.3% increase of error. More importantly, our method is able to accelerate modern networks like ResNet, Xception and suffers only 1.4%, 1.0% accuracy loss under 2x speed-up respectively, which is significant. Code has been made publicly available.

Gut microbiota dysbiosis contributes to the development of hypertension.

Abstract: Recently, the potential role of gut microbiome in metabolic diseases has been revealed, especially in cardiovascular diseases. Hypertension is one of the most prevalent cardiovascular diseases worldwide, yet whether gut microbiota dysbiosis participates in the development of hypertension remains largely unknown. To investigate this issue, we carried out comprehensive metagenomic and metabolomic analyses in a cohort of 41 healthy controls, 56 subjects with pre-hypertension, 99 individuals with primary hypertension, and performed fecal microbiota transplantation from patients to germ-free mice. Compared to the healthy controls, we found dramatically decreased microbial richness and diversity, Prevotella-dominated gut enterotype, distinct metagenomic composition with reduced bacteria associated with healthy status and overgrowth of bacteria such as Prevotella and Klebsiella, and disease-linked microbial function in both pre-hypertensive and hypertensive populations. Unexpectedly, the microbiome characteristic in pre-hypertension group was quite similar to that in hypertension. The metabolism changes of host with pre-hypertension or hypertension were identified to be closely linked to gut microbiome dysbiosis. And a disease classifier based on microbiota and metabolites was constructed to discriminate pre-hypertensive and hypertensive individuals from controls accurately. Furthermore, by fecal transplantation from hypertensive human donors to germ-free mice, elevated blood pressure was observed to be transferrable through microbiota, and the direct influence of gut microbiota on blood pressure of the host was demonstrated. Overall, our results describe a novel causal role of aberrant gut microbiota in contributing to the pathogenesis of hypertension. And the significance of early intervention for pre-hypertension was emphasized.

Filling the oxygen vacancies in Co3O4 with phosphorus: an ultra-efficient electrocatalyst for overall water splitting

Abstract: It is of essential importance to design an electrocatalyst with excellent performance for both hydrogen evolution reaction (HER) and oxygen evolution reaction (OER) in water splitting. Co3O4 has been developed as a highly efficient OER electrocatalyst, but it has almost no activity for HER. In a previous study, it has been demonstrated that the formation of oxygen vacancies (VO) in Co3O4 can significantly enhance the OER activity. However, the stability of VO needs to be considered, especially under the highly oxidizing conditions of the OER process. It is a big challenge to stabilize the VO in Co3O4 while reserving the excellent activity. Filling the oxygen vacancies with heteroatoms in the VO-rich Co3O4 may be a smart strategy to stabilize the VO by compensating the coordination numbers and obtain an even surprising activity due to the modification of electronic properties by heteroatoms. Herein, we successfully transformed VO-rich Co3O4 into an HER-OER electrocatalyst by filling the in situ formed VO in Co3O4 with phosphorus (P-Co3O4) by treating Co3O4 with Ar plasma in the presence of a P precursor. The relatively lower coordination numbers in VO-Co3O4 than those in pristine Co3O4 were evidenced by X-ray adsorption spectroscopy, indicating that the oxygen vacancies were created after Ar plasma etching. On the other hand, the relatively higher coordination numbers in P-Co3O4 than those in VO-Co3O4 and nearly same coordination number as that in pristine Co3O4 strongly suggest the efficient filling of P in the vacancies by treatment with Ar plasma in the presence of a P precursor. The Co–O bonds in Co3O4 consist of octahedral Co3+(Oh)–O and tetrahedral Co2+(Td)–O (Oh, octahedral coordination by six oxygen atoms and Td, tetrahedral coordination by four oxygen atoms). More Co3+(Oh)–O are broken when oxygen vacancies are formed in VO-Co3O4, and more electrons enter the octahedral Co 3d orbital than those entering the tetrahedral Co 3d orbital. Then, with the filling of P in the vacancy site, electrons are transferred out of the Co 3d states, and more Co2+(Td) than Co3+(Oh) are left in P-Co3O4. These results suggest that the favored catalytic ability of P-Co3O4 is dominated by the Co2+(Td) site. P-Co3O4 shows superior electrocatalytic activities for HER and OER (among the best non-precious metal catalysts). Owing to its superior efficiency, P-Co3O4 can directly catalyze overall water splitting with excellent performance. The theoretical calculations illustrated that the improved electrical conductivity and intermediate binding by P-filling contributed significantly to the enhanced HER and OER activity of Co3O4.

Metal-Organic-Framework-Derived Hybrid Carbon Nanocages as a Bifunctional Electrocatalyst for Oxygen Reduction and Evolution.

Abstract: The oxygen reduction reaction (ORR) and oxygen evolution reaction (OER) are cornerstone reactions for many renewable energy technologies. Developing cheap yet durable substitutes of precious-metal catalysts, especially the bifunctional electrocatalysts with high activity for both ORR and OER reactions and their streamlined coupling process, are highly desirable to reduce the processing cost and complexity of renewable energy systems. Here, a facile strategy is reported for synthesizing double-shelled hybrid nanocages with outer shells of Co-N-doped graphitic carbon (Co-NGC) and inner shells of N-doped microporous carbon (NC) by templating against core-shell metal-organic frameworks. The double-shelled NC@Co-NGC nanocages well integrate the high activity of Co-NGC shells into the robust NC hollow framework with enhanced diffusion kinetics, exhibiting superior electrocatalytic properties to Pt and RuO2 as a bifunctional electrocatalyst for ORR and OER, and hold a promise as efficient air electrode catalysts in Zn-air batteries. First-principles calculations reveal that the high catalytic activities of Co-NGC shells are due to the synergistic electron transfer and redistribution between the Co nanoparticles, the graphitic carbon, and the doped N species. Strong yet favorable adsorption of an OOH* intermediate on the high density of uncoordinated hollow-site C atoms with respect to the Co lattice in the Co-NGC structure is a vital rate-determining step to achieve excellent bifunctional electrocatalytic activity.

Single-Junction Binary-Blend Nonfullerene Polymer Solar Cells with 12.1% Efficiency.

Abstract: A new fluorinated nonfullerene acceptor, ITIC-Th1, has been designed and synthesized by introducing fluorine (F) atoms onto the end-capping group 1,1-dicyanomethylene-3-indanone (IC). On the one hand, incorporation of F would improve intramolecular interaction, enhance the push-pull effect between the donor unit indacenodithieno[3,2-b]thiophene and the acceptor unit IC due to electron-withdrawing effect of F, and finally adjust energy levels and reduce bandgap, which is beneficial to light harvesting and enhancing short-circuit current density (JSC ). On the other hand, incorporation of F would improve intermolecular interactions through CF···S, CF···H, and CF···π noncovalent interactions and enhance electron mobility, which is beneficial to enhancing JSC and fill factor. Indeed, the results show that fluorinated ITIC-Th1 exhibits redshifted absorption, smaller optical bandgap, and higher electron mobility than the nonfluorinated ITIC-Th. Furthermore, nonfullerene organic solar cells (OSCs) based on fluorinated ITIC-Th1 electron acceptor and a wide-bandgap polymer donor FTAZ based on benzodithiophene and benzotriazole exhibit power conversion efficiency (PCE) as high as 12.1%, significantly higher than that of nonfluorinated ITIC-Th (8.88%). The PCE of 12.1% is the highest in fullerene and nonfullerene-based single-junction binary-blend OSCs. Moreover, the OSCs based on FTAZ:ITIC-Th1 show much better efficiency and better stability than the control devices based on FTAZ:PC71 BM (PCE = 5.22%).

Recent progress in layered transition metal carbides and/or nitrides (MXenes) and their composites: synthesis and applications

Abstract: Since their inception in 2011, from the inaugural synthesis of multi-layered Ti3C2Tx by etching Ti3AlC2 with hydrofluoric acid (HF), novel routes with a myriad of reducing agents, etchants and intercalants have been explored and many new members have been added to the two-dimensional (2D) material constellation. Despite being endowed with the rare combination of good electronic conductivity and hydrophilicity, their inherent low capacities, for instance, temper their prospect for application for electrodes in energy storage systems. MXene-based composites, however, with a probable synergistic effect in agglomeration prevention, facilitating electronic conductivity, improving electrochemical stability, enhancing pseudocapacitance and minimizing the shortcomings of individual components, exceed the previously mentioned capacitance ceiling. In this review, we summarise the development and progress in the synthesis of various multi-layered carbides, carbonitrides and nitrides, and intercalants, as well as the subsequent processing in order to delaminate them into single- and/or few-layered and incorporate into MXene-based composites, focusing on their performance and application as transparent conductive films, environmental remediation, electromagnetic interference (EMI) absorption and shielding, electrocatalysts, Li-ion batteries (LIBs), supercapacitors and other electrochemical storage systems.

Weighted Nuclear Norm Minimization and Its Applications to Low Level Vision

Abstract: As a convex relaxation of the rank minimization model, the nuclear norm minimization (NNM) problem has been attracting significant research interest in recent years. The standard NNM regularizes each singular value equally, composing an easily calculated convex norm. However, this restricts its capability and flexibility in dealing with many practical problems, where the singular values have clear physical meanings and should be treated differently. In this paper we study the weighted nuclear norm minimization (WNNM) problem, which adaptively assigns weights on different singular values. As the key step of solving general WNNM models, the theoretical properties of the weighted nuclear norm proximal (WNNP) operator are investigated. Albeit nonconvex, we prove that WNNP is equivalent to a standard quadratic programming problem with linear constrains, which facilitates solving the original problem with off-the-shelf convex optimization solvers. In particular, when the weights are sorted in a non-descending order, its optimal solution can be easily obtained in closed-form. With WNNP, the solving strategies for multiple extensions of WNNM, including robust PCA and matrix completion, can be readily constructed under the alternating direction method of multipliers paradigm. Furthermore, inspired by the reweighted sparse coding scheme, we present an automatic weight setting method, which greatly facilitates the practical implementation of WNNM. The proposed WNNM methods achieve state-of-the-art performance in typical low level vision tasks, including image denoising, background subtraction and image inpainting.

Functional and Biomimetic Materials for Engineering of the Three-Dimensional Cell Microenvironment

Abstract: The cell microenvironment has emerged as a key determinant of cell behavior and function in development, physiology, and pathophysiology. The extracellular matrix (ECM) within the cell microenvironment serves not only as a structural foundation for cells but also as a source of three-dimensional (3D) biochemical and biophysical cues that trigger and regulate cell behaviors. Increasing evidence suggests that the 3D character of the microenvironment is required for development of many critical cell responses observed in vivo, fueling a surge in the development of functional and biomimetic materials for engineering the 3D cell microenvironment. Progress in the design of such materials has improved control of cell behaviors in 3D and advanced the fields of tissue regeneration, in vitro tissue models, large-scale cell differentiation, immunotherapy, and gene therapy. However, the field is still in its infancy, and discoveries about the nature of cell–microenvironment interactions continue to overturn much earl...

π-Conjugated Lewis Base: Efficient Trap-Passivation and Charge-Extraction for Hybrid Perovskite Solar Cells

Abstract: A π-conjugated Lewis base is introduced into perovskite solar cells, namely, indacenodithiophene end-capped with 1.1-dicyanomethylene-3-indanone (IDIC), as a multifunctional interlayer, which combines efficient trap-passivation and electron-extraction. Perovskite solar cells with IDIC layers yield higher photovoltages and photocurrents, and 45% enhanced efficiency compared with control devices without IDIC.

Co-Saliency Detection via a Self-Paced Multiple-Instance Learning Framework

Abstract: As an interesting and emerging topic, co-saliency detection aims at simultaneously extracting common salient objects from a group of images On one hand, traditional co-saliency detection approaches rely heavily on human knowledge for designing hand-crafted metrics to possibly reflect the faithful properties of the co-salient regions Such strategies, however, always suffer from poor generalization capability to flexibly adapt various scenarios in real applications On the other hand, most current methods pursue co-saliency detection in unsupervised fashions This, however, tends to weaken their performance in real complex scenarios because they are lack of robust learning mechanism to make full use of the weak labels of each image To alleviate these two problems, this paper proposes a new SP-MIL framework for co-saliency detection, which integrates both multiple instance learning (MIL) and self-paced learning (SPL) into a unified learning framework Specifically, for the first problem, we formulate the co-saliency detection problem as a MIL paradigm to learn the discriminative classifiers to detect the co-saliency object in the “instance-level” The formulated MIL component facilitates our method capable of automatically producing the proper metrics to measure the intra-image contrast and the inter-image consistency for detecting co-saliency in a purely self-learning way For the second problem, the embedded SPL paradigm is able to alleviate the data ambiguity under the weak supervision of co-saliency detection and guide a robust learning manner in complex scenarios Experiments on benchmark datasets together with multiple extended computer vision applications demonstrate the superiority of the proposed framework beyond the state-of-the-arts

Potassium–sodium niobate based lead-free ceramics: novel electrical energy storage materials

Abstract: The development of lead-free bulk ceramics with high recoverable energy density (Wrec) is of decisive importance for meeting the requirements of advanced pulsed power capacitors toward miniaturization and integration. However, the Wrec (<2 J cm−3) of lead-free bulk ceramics has long been limited by their low dielectric breakdown strength (DBS < 200 kV cm−1) and small saturation polarization (Ps). In this work, a strategy (compositions control the grain size of lead-free ceramics to submicron scale to increase the DBS, and the hybridization between the Bi 6p and O 2p orbitals enhances the Ps) was proposed to improve the Wrec of lead-free ceramics. (K0.5Na0.5)NbO3–Bi(Me2/3Nb1/3)O3 solid solutions (where Me2+ = Mg and Zn) were designed for achieving large Ps, and high DBS and Wrec. As an example, (1 − x)(K0.5Na0.5)NbO3–xBi(Mg2/3Nb1/3)O3 (KNN–BMN) ceramics were prepared by using a conventional solid-state reaction process in this study. Large Ps (41 μC cm−2) and high DBS (300 kV cm−1) were obtained for 0.90KNN–0.10BMN ceramics, leading to large Wrec (4.08 J cm−3). The significantly enhanced Wrec is more than 2–3 times larger than that of other lead-free bulk ceramics. The findings in this study not only provide a design methodology for developing lead-free bulk ceramics with large Wrec but also could bring about the development of a series of KNN-based ceramics with significantly enhanced Wrec and DBS in the future. More importantly, this work opens a new research and application field (dielectric energy storage) for (K0.5Na0.5)NbO3-based ceramics.

Battling the Extreme: A Study on the Power System Resilience

Abstract: The electricity infrastructure is a critical lifeline system and of utmost importance to our daily lives. Power system resilience characterizes the ability to resist, adapt to, and timely recover from disruptions. The resilient power system is intended to cope with low probability, high risk extreme events including extreme natural disasters and man-made attacks. With an increasing awareness of such threats, the resilience of power systems has become a top priority for many countries. Facing the pressing urgency for resilience studies, the objective of this paper is to investigate the resilience of power systems. It summarizes practices taken by governments, utilities, and researchers to increase power system resilience. Based on a thorough review on the existing metrics system and evaluation methodologies, we present the concept, metrics, and a quantitative framework for power system resilience evaluation. Then, system hardening strategies and smart grid technologies as means to increase system resilience are discussed, with an emphasis on the new technologies such as topology reconfiguration, microgrids, and distribution automation; to illustrate how to increase system resilience against extreme events, we propose a load restoration framework based on smart distribution technology. The proposed method is applied on two test systems to validify its effectiveness. In the end, challenges to the power system resilience are discussed, including extreme event modeling, practical barriers, interdependence with other critical infrastructures, etc.

Direct 4D printing via active composite materials

Abstract: We describe an approach to print composite polymers in high-resolution three-dimensional (3D) architectures that can be rapidly transformed to a new permanent configuration directly by heating. The permanent shape of a component results from the programmed time evolution of the printed shape upon heating via the design of the architecture and process parameters of a composite consisting of a glassy shape memory polymer and an elastomer that is programmed with a built-in compressive strain during photopolymerization. Upon heating, the shape memory polymer softens, releases the constraint on the strained elastomer, and allows the object to transform into a new permanent shape, which can then be reprogrammed into multiple subsequent shapes. Our key advance, the markedly simplified creation of high-resolution complex 3D reprogrammable structures, promises to enable myriad applications across domains, including medical technology, aerospace, and consumer products, and even suggests a new paradigm in product design, where components are simultaneously designed to inhabit multiple configurations during service.

Reducing the stochasticity of crystal nucleation to enable subnanosecond memory writing

Abstract: Operation speed is a key challenge in phase-change random-access memory (PCRAM) technology, especially for achieving subnanosecond high-speed cache memory. Commercialized PCRAM products are limited by the tens of nanoseconds writing speed, originating from the stochastic crystal nucleation during the crystallization of amorphous germanium antimony telluride (Ge2Sb2Te5). Here, we demonstrate an alloying strategy to speed up the crystallization kinetics. The scandium antimony telluride (Sc0.2Sb2Te3) compound that we designed allows a writing speed of only 700 picoseconds without preprogramming in a large conventional PCRAM device. This ultrafast crystallization stems from the reduced stochasticity of nucleation through geometrically matched and robust scandium telluride (ScTe) chemical bonds that stabilize crystal precursors in the amorphous state. Controlling nucleation through alloy design paves the way for the development of cache-type PCRAM technology to boost the working efficiency of computing systems.

Dislocation mechanisms and 3D twin architectures generate exceptional strength-ductility-toughness combination in CrCoNi medium-entropy alloy

Abstract: Combinations of high strength and ductility are hard to attain in metals. Exceptions include materials exhibiting twinning-induced plasticity. To understand how the strength-ductility trade-off can be defeated, we apply in situ, and aberration-corrected scanning, transmission electron microscopy to examine deformation mechanisms in the medium-entropy alloy CrCoNi that exhibits one of the highest combinations of strength, ductility and toughness on record. Ab initio modelling suggests that it has negative stacking-fault energy at 0K and high propensity for twinning. With deformation we find that a three-dimensional (3D) hierarchical twin network forms from the activation of three twinning systems. This serves a dual function: conventional twin-boundary (TB) strengthening from blockage of dislocations impinging on TBs, coupled with the 3D twin network which offers pathways for dislocation glide along, and cross-slip between, intersecting TB-matrix interfaces. The stable twin architecture is not disrupted by interfacial dislocation glide, serving as a continuous source of strength, ductility and toughness.

Self-healing SEI enables full-cell cycling of a silicon-majority anode with a coulombic efficiency exceeding 99.9%

Abstract: Despite active developments, full-cell cycling of Li-battery anodes with >50 wt% Si (a Si-majority anode, SiMA) is rare. The main challenge lies in the solid electrolyte interphase (SEI), which when formed naturally (nSEI), is fragile and cannot tolerate the large volume changes of Si during lithiation/delithiation. An artificial SEI (aSEI) with a specific set of mechanical characteristics is henceforth designed; we enclose Si within a TiO2 shell thinner than 15 nm, which may or may not be completely hermetic at the beginning. In situ TEM experiments show that the TiO2 shell exhibits 5× greater strength than an amorphous carbon shell. Void-padded compartmentalization of Si can survive the huge volume changes and electrolyte ingression, with a self-healing aSEI + nSEI. The half-cell capacity exceeds 990 mA h g−1 after 1500 cycles. To improve the volumetric capacity, we further compress SiMA 3-fold from its tap density (0.4 g cm−3) to 1.4 g cm−3, and then run the full-cell battery tests against a 3 mA h cm−2 LiCoO2 cathode. Despite some TiO2 enclosures being inevitably broken, 2× the volumetric capacity (1100 mA h cm−3) and 2× the gravimetric capacity (762 mA h g−1) of commercial graphite anode is achieved in stable full-cell battery cycling, with a stabilized areal capacity of 1.6 mA h cm−2 at the 100th cycle. The initial lithium loss, characterized by the coulombic inefficiency (CI), is carefully tallied on a logarithmic scale and compared with the actual full-cell capacity loss. It is shown that a strong, non-adherent aSEI, even if partially cracked, facilitates an adaptive self-repair mechanism that enables full-cell cycling of a SiMA, leading to a stabilized coulombic efficiency exceeding 99.9%.

View Adaptive Recurrent Neural Networks for High Performance Human Action Recognition from Skeleton Data

Abstract: Skeleton-based human action recognition has recently attracted increasing attention due to the popularity of 3D skeleton data. One main challenge lies in the large view variations in captured human actions. We propose a novel view adaptation scheme to automatically regulate observation viewpoints during the occurrence of an action. Rather than re-positioning the skeletons based on a human defined prior criterion, we design a view adaptive recurrent neural network (RNN) with LSTM architecture, which enables the network itself to adapt to the most suitable observation viewpoints from end to end. Extensive experiment analyses show that the proposed view adaptive RNN model strives to (1) transform the skeletons of various views to much more consistent viewpoints and (2) maintain the continuity of the action rather than transforming every frame to the same position with the same body orientation. Our model achieves significant improvement over the state-of-the-art approaches on three benchmark datasets.

3D Printing of Transparent and Conductive Heterogeneous Hydrogel-Elastomer Systems.

Abstract: A hydrogel-dielectric-elastomer system, polyacrylamide and poly(dimethylsiloxane) (PDMS), is adapted for extrusion printing for integrated device fabrication. A lithium-chloride-containing hydrogel printing ink is developed and printed onto treated PDMS with no visible signs of delamination and geometrically scaling resistance under moderate uniaxial tension and fatigue. A variety of designs are demonstrated, including a resistive strain gauge and an ionic cable.

High-Performance Ternary Organic Solar Cell Enabled by a Thick Active Layer Containing a Liquid Crystalline Small Molecule Donor.

Abstract: Ternary organic solar cells (OSCs) have attracted much research attention in the past few years, as ternary organic blends can broaden the absorption range of OSCs without the use of complicated tandem cell structures. Despite their broadened absorption range, the light harvesting capability of ternary OSCs is still limited because most ternary OSCs use thin active layers of about 100 nm in thickness, which is not sufficient to absorb all photons in their spectral range and may also cause problems for future roll-to-roll mass production that requires thick active layers. In this paper, we report a highly efficient ternary OSC (11.40%) obtained by incorporating a nematic liquid crystalline small molecule (named benzodithiophene terthiophene rhodanine (BTR)) into a state-of-the-art PTB7-Th:PC71BM binary system. The addition of BTR into PTB7-Th:PC71BM was found to improve the morphology of the blend film with decreased π–π stacking distance, enlarged coherence length, and enhanced domain purity. This resulte...

Mapping Polymer Donors toward High‐Efficiency Fullerene Free Organic Solar Cells

Abstract: Five polymer donors with distinct chemical structures and different electronic properties are surveyed in a planar and narrow-bandgap fused-ring electron acceptor (IDIC)-based organic solar cells, which exhibit power conversion efficiencies of up to 11%.