Showing papers by "Nanyang Technological University published in 2018"

Recent Trends in Deep Learning Based Natural Language Processing [Review Article]

Abstract: Deep learning methods employ multiple processing layers to learn hierarchical representations of data, and have produced state-of-the-art results in many domains. Recently, a variety of model designs and methods have blossomed in the context of natural language processing (NLP). In this paper, we review significant deep learning related models and methods that have been employed for numerous NLP tasks and provide a walk-through of their evolution. We also summarize, compare and contrast the various models and put forward a detailed understanding of the past, present and future of deep learning in NLP.

Perovskite light-emitting diodes with external quantum efficiency exceeding 20 per cent

Abstract: Metal halide perovskite materials are an emerging class of solution-processable semiconductors with considerable potential for use in optoelectronic devices1–3. For example, light-emitting diodes (LEDs) based on these materials could see application in flat-panel displays and solid-state lighting, owing to their potential to be made at low cost via facile solution processing, and could provide tunable colours and narrow emission line widths at high photoluminescence quantum yields4–8. However, the highest reported external quantum efficiencies of green- and red-light-emitting perovskite LEDs are around 14 per cent7,9 and 12 per cent8, respectively—still well behind the performance of organic LEDs10–12 and inorganic quantum dot LEDs13. Here we describe visible-light-emitting perovskite LEDs that surpass the quantum efficiency milestone of 20 per cent. This achievement stems from a new strategy for managing the compositional distribution in the device—an approach that simultaneously provides high luminescence and balanced charge injection. Specifically, we mixed a presynthesized CsPbBr3 perovskite with a MABr additive (where MA is CH3NH3), the differing solubilities of which yield sequential crystallization into a CsPbBr3/MABr quasi-core/shell structure. The MABr shell passivates the nonradiative defects that would otherwise be present in CsPbBr3 crystals, boosting the photoluminescence quantum efficiency, while the MABr capping layer enables balanced charge injection. The resulting 20.3 per cent external quantum efficiency represents a substantial step towards the practical application of perovskite LEDs in lighting and display. A strategy for managing the compositional distribution in metal halide perovskite light-emitting diodes enables them to surpass 20% external quantum efficiency—a step towards their practical application in lighting and displays.

ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks

Abstract: The Super-Resolution Generative Adversarial Network (SRGAN) is a seminal work that is capable of generating realistic textures during single image super-resolution. However, the hallucinated details are often accompanied with unpleasant artifacts. To further enhance the visual quality, we thoroughly study three key components of SRGAN – network architecture, adversarial loss and perceptual loss, and improve each of them to derive an Enhanced SRGAN (ESRGAN). In particular, we introduce the Residual-in-Residual Dense Block (RRDB) without batch normalization as the basic network building unit. Moreover, we borrow the idea from relativistic GAN to let the discriminator predict relative realness instead of the absolute value. Finally, we improve the perceptual loss by using the features before activation, which could provide stronger supervision for brightness consistency and texture recovery. Benefiting from these improvements, the proposed ESRGAN achieves consistently better visual quality with more realistic and natural textures than SRGAN and won the first place in the PIRM2018-SR Challenge (region 3) with the best perceptual index. The code is available at https://github.com/xinntao/ESRGAN.

Atomically dispersed Ni(i) as the active site for electrochemical CO2 reduction

Abstract: Electrochemical reduction of CO2 to chemical fuel offers a promising strategy for managing the global carbon balance, but presents challenges for chemistry due to the lack of effective electrocatalyst. Here we report atomically dispersed nickel on nitrogenated graphene as an efficient and durable electrocatalyst for CO2 reduction. Based on operando X-ray absorption and photoelectron spectroscopy measurements, the monovalent Ni(i) atomic center with a d9 electronic configuration was identified as the catalytically active site. The single-Ni-atom catalyst exhibits high intrinsic CO2 reduction activity, reaching a specific current of 350 A gcatalyst−1 and turnover frequency of 14,800 h−1 at a mild overpotential of 0.61 V for CO conversion with 97% Faradaic efficiency. The catalyst maintained 98% of its initial activity after 100 h of continuous reaction at CO formation current densities as high as 22 mA cm−2. Electrocatalysts with improved activity and stability for the conversion of CO2 to CO are being sought. Using operando spectroscopies, the authors identify atomically dispersed Ni(i) as the active site in a nitrogenated-graphene-supported catalyst with high intrinsic activity and stability over 100 hours.

Analysis of shared heritability in common disorders of the brain

Abstract: Disorders of the brain can exhibit considerable epidemiological comorbidity and often share symptoms, provoking debate about their etiologic overlap. We quantified the genetic sharing of 25 brain disorders from genome-wide association studies of 265,218 patients and 784,643 control participants and assessed their relationship to 17 phenotypes from 1,191,588 individuals. Psychiatric disorders share common variant risk, whereas neurological disorders appear more distinct from one another and from the psychiatric disorders. We also identified significant sharing between disorders and a number of brain phenotypes, including cognitive measures. Further, we conducted simulations to explore how statistical power, diagnostic misclassification, and phenotypic heterogeneity affect genetic correlations. These results highlight the importance of common genetic variation as a risk factor for brain disorders and the value of heritability-based methods in understanding their etiology.

Novel genetic associations for blood pressure identified via gene-alcohol interaction in up to 570K individuals across multiple ancestries

Abstract: Heavy alcohol consumption is an established risk factor for hypertension; the mechanism by which alcohol consumption impact blood pressure (BP) regulation remains unknown. We hypothesized that a genome-wide association study accounting for gene-alcohol consumption interaction for BP might identify additional BP loci and contribute to the understanding of alcohol-related BP regulation. We conducted a large two-stage investigation incorporating joint testing of main genetic effects and single nucleotide variant (SNV)-alcohol consumption interactions. In Stage 1, genome-wide discovery meta-analyses in ≈131K individuals across several ancestry groups yielded 3,514 SNVs (245 loci) with suggestive evidence of association (P < 1.0 x 10-5). In Stage 2, these SNVs were tested for independent external replication in ≈440K individuals across multiple ancestries. We identified and replicated (at Bonferroni correction threshold) five novel BP loci (380 SNVs in 21 genes) and 49 previously reported BP loci (2,159 SNVs in 109 genes) in European ancestry, and in multi-ancestry meta-analyses (P < 5.0 x 10-8). For African ancestry samples, we detected 18 potentially novel BP loci (P < 5.0 x 10-8) in Stage 1 that warrant further replication. Additionally, correlated meta-analysis identified eight novel BP loci (11 genes). Several genes in these loci (e.g., PINX1, GATA4, BLK, FTO and GABBR2) have been previously reported to be associated with alcohol consumption. These findings provide insights into the role of alcohol consumption in the genetic architecture of hypertension.

A library of atomically thin metal chalcogenides

Abstract: Investigations of two-dimensional transition-metal chalcogenides (TMCs) have recently revealed interesting physical phenomena, including the quantum spin Hall effect1,2, valley polarization3,4 and two-dimensional superconductivity 5 , suggesting potential applications for functional devices6–10. However, of the numerous compounds available, only a handful, such as Mo- and W-based TMCs, have been synthesized, typically via sulfurization11–15, selenization16,17 and tellurization 18 of metals and metal compounds. Many TMCs are difficult to produce because of the high melting points of their metal and metal oxide precursors. Molten-salt-assisted methods have been used to produce ceramic powders at relatively low temperature 19 and this approach 20 was recently employed to facilitate the growth of monolayer WS2 and WSe2. Here we demonstrate that molten-salt-assisted chemical vapour deposition can be broadly applied for the synthesis of a wide variety of two-dimensional (atomically thin) TMCs. We synthesized 47 compounds, including 32 binary compounds (based on the transition metals Ti, Zr, Hf, V, Nb, Ta, Mo, W, Re, Pt, Pd and Fe), 13 alloys (including 11 ternary, one quaternary and one quinary), and two heterostructured compounds. We elaborate how the salt decreases the melting point of the reactants and facilitates the formation of intermediate products, increasing the overall reaction rate. Most of the synthesized materials in our library are useful, as supported by evidence of superconductivity in our monolayer NbSe2 and MoTe2 samples21,22 and of high mobilities in MoS2 and ReS2. Although the quality of some of the materials still requires development, our work opens up opportunities for studying the properties and potential application of a wide variety of two-dimensional TMCs.

Defining “Fake News”: A typology of scholarly definitions

Abstract: This paper is based on a review of how previous studies have defined and operationalized the term “fake news.” An examination of 34 academic articles that used the term “fake news” between 2003 and...

Deep Learning Applications in Medical Image Analysis

Abstract: The tremendous success of machine learning algorithms at image recognition tasks in recent years intersects with a time of dramatically increased use of electronic medical records and diagnostic imaging. This review introduces the machine learning algorithms as applied to medical image analysis, focusing on convolutional neural networks, and emphasizing clinical aspects of the field. The advantage of machine learning in an era of medical big data is that significant hierarchal relationships within the data can be discovered algorithmically without laborious hand-crafting of features. We cover key research areas and applications of medical image classification, localization, detection, segmentation, and registration. We conclude by discussing research obstacles, emerging trends, and possible future directions.

PSANet: Point-wise Spatial Attention Network for Scene Parsing

Abstract: We notice information flow in convolutional neural networks is restricted inside local neighborhood regions due to the physical design of convolutional filters, which limits the overall understanding of complex scenes. In this paper, we propose the point-wise spatial attention network (PSANet) to relax the local neighborhood constraint. Each position on the feature map is connected to all the other ones through a self-adaptively learned attention mask. Moreover, information propagation in bi-direction for scene parsing is enabled. Information at other positions can be collected to help the prediction of the current position and vice versa, information at the current position can be distributed to assist the prediction of other ones. Our proposed approach achieves top performance on various competitive scene parsing datasets, including ADE20K, PASCAL VOC 2012 and Cityscapes, demonstrating its effectiveness and generality.

Advanced Energy Storage Devices: Basic Principles, Analytical Methods, and Rational Materials Design.

Abstract: Tremendous efforts have been dedicated into the development of high-performance energy storage devices with nanoscale design and hybrid approaches. The boundary between the electrochemical capacitors and batteries becomes less distinctive. The same material may display capacitive or battery-like behavior depending on the electrode design and the charge storage guest ions. Therefore, the underlying mechanisms and the electrochemical processes occurring upon charge storage may be confusing for researchers who are new to the field as well as some of the chemists and material scientists already in the field. This review provides fundamentals of the similarities and differences between electrochemical capacitors and batteries from kinetic and material point of view. Basic techniques and analysis methods to distinguish the capacitive and battery-like behavior are discussed. Furthermore, guidelines for material selection, the state-of-the-art materials, and the electrode design rules to advanced electrode are proposed.

Domain Generalization with Adversarial Feature Learning

Abstract: In this paper, we tackle the problem of domain generalization: how to learn a generalized feature representation for an "unseen" target domain by taking the advantage of multiple seen source-domain data. We present a novel framework based on adversarial autoencoders to learn a generalized latent feature representation across domains for domain generalization. To be specific, we extend adversarial autoencoders by imposing the Maximum Mean Discrepancy (MMD) measure to align the distributions among different domains, and matching the aligned distribution to an arbitrary prior distribution via adversarial feature learning. In this way, the learned feature representation is supposed to be universal to the seen source domains because of the MMD regularization, and is expected to generalize well on the target domain because of the introduction of the prior distribution. We proposed an algorithm to jointly train different components of our proposed framework. Extensive experiments on various vision tasks demonstrate that our proposed framework can learn better generalized features for the unseen target domain compared with state-of-the-art domain generalization methods.

Single Cobalt Atoms Anchored on Porous N-Doped Graphene with Dual Reaction Sites for Efficient Fenton-like Catalysis

Abstract: The Fenton-like process presents one of the most promising strategies to generate reactive oxygen-containing radicals to deal with the ever-growing environmental pollution. However, developing improved catalysts with adequate activity and stability is still a long-term goal for practical application. Herein, we demonstrate single cobalt atoms anchored on porous N-doped graphene with dual reaction sites as highly reactive and stable Fenton-like catalysts for efficient catalytic oxidation of recalcitrant organics via activation of peroxymonosulfate (PMS). Our experiments and density functional theory (DFT) calculations show that the CoN4 site with a single Co atom serves as the active site with optimal binding energy for PMS activation, while the adjacent pyrrolic N site adsorbs organic molecules. The dual reaction sites greatly reduce the migration distance of the active singlet oxygen produced from PMS activation and thus improve the Fenton-like catalytic performance.

Two-dimensional metal-organic framework nanosheets: synthesis and applications.

Abstract: Two-dimensional (2D) metal–organic framework (MOF) nanosheets are attracting increasing research attention due to their unique properties originating from their ultrathin thickness, large surface area and high surface-to-volume atom ratios. Many great advances have been made in the synthesis and application of 2D MOF nanosheets over the past few years. In this review, we summarize the recent advances in the synthesis of 2D MOF nanosheets by using top-down methods, e.g. sonication exfoliation, mechanical exfoliation, Li-intercalation exfoliation and chemical exfoliation, and bottom-up methods, i.e. interfacial synthesis, three-layer synthesis, surfactant-assisted synthesis, modulated synthesis, and sonication synthesis. In addition, the recent progress in 2D MOF nanosheet-based nanocomposites is also briefly introduced. The potential applications of 2D MOF nanosheets in gas separation, energy conversion and storage, catalysis, sensors and biomedicine are discussed. Finally, we give our personal insights into the challenges and opportunities for the future research of 2D MOF nanosheets and their composites.

Construction of ZnIn2S4-In2O3 Hierarchical Tubular Heterostructures for Efficient CO2 Photoreduction.

Abstract: We demonstrate the rational design and construction of sandwich-like ZnIn2S4–In2O3 hierarchical tubular heterostructures by growing ZnIn2S4 nanosheets on both inner and outer surfaces of In2O3 microtubes as photocatalysts for efficient CO2 photoreduction. The unique design integrates In2O3 and ZnIn2S4 into hierarchical one-dimensional (1D) open architectures with double-heterojunction shells and ultrathin two-dimensional (2D) nanosheet subunits. This design accelerates the separation and transfer of photogenerated charges, offers large surface area for CO2 adsorption, and exposes abundant active sites for surface catalysis. Benefiting from the structural and compositional merits, the optimized ZnIn2S4–In2O3 photocatalyst exhibits outstanding performance for reductive CO2 deoxygenation with considerable CO generation rate (3075 μmol h–1 g–1) and high stability.

Topological insulator laser: Theory

Abstract: INTRODUCTION Topological insulators emerged in condensed matter physics and constitute a new phase of matter, with insulating bulk and robust edge conductance that is immune to imperfections and disorder To date, topological protection is known to be a ubiquitous phenomenon, occurring in many physical settings, ranging from photonics and cold atoms to acoustic, mechanical, and elastic systems So far, however, most of these studies were carried out in entirely passive, linear, and conservative settings RATIONALE We propose topological insulator lasers: lasers whose lasing mode exhibits topologically protected transport without magnetic fields Extending topological physics to lasers is far from natural In fact, lasers are built on foundations that are seemingly inconsistent with the essence of topological insulators: They require gain (and thus are non-Hermitian), they are nonlinear entities because the gain must be saturable, and they are open systems because they emit light These properties, common to all lasers, cast major doubts on the possibility of harnessing topological features to make a topological insulator laser Despite this common mindset, we show that the use of topological properties leads to highly efficient lasers, robust to defects and disorder, with single-mode lasing even at conditions high above the laser threshold RESULTS We demonstrate that topological insulator lasers are theoretically possible and experimentally feasible We consider two configurations involving planar arrays of coupled active resonators The first is based on the Haldane model, archetypical for topological systems The second model, geared toward experiment, constitutes an aperiodic array architecture creating an artificial magnetic field We show that by introducing saturable gain and loss, it is possible to make these systems lase in a topological edge state In this way, the lasing mode exhibits topologically protected transport; the light propagates unidirectionally along the edges of the cavity, immune to scattering and disorder, unaffected by the shape of the edges Moreover, we show that the underlying topological properties not only make the system robust to fabrication and operational disorder and defects, they also lead to a highly efficient single-mode lasing that remains single-mode even at gain values high above the laser threshold The figure describes the geometry and features of a topological insulator laser based on the Haldane model while adding saturable gain, loss, and an output port The cavity is a planar honeycomb lattice of coupled microring resonators, pumped at the perimeter with a lossy interior We show that under these conditions, lasing occurs at the topological edge mode, which has unidirectional flux and is extended around the perimeter with almost-uniform intensity The topological cavities exhibit higher efficiency than the trivial cavity, even under strong disorder For the topological laser with a small gap, the topological protection holds as long as the disorder level is smaller than the gap size DISCUSSION The concept of the topological insulator laser alters current understanding of the interplay between disorder and lasing, and opens exciting possibilities at the interface of topological physics and laser science, such as topologically protected transport in systems with gain We show here that the laser system based on the archetypal Haldane model exhibits topologically protected transport, with features similar to those of its passive counterpart This behavior means that this system is likely to have topological invariants, despite the nonhermiticity Technologically, the topological insulator laser offers an avenue to make many semiconductor lasers operate as one single-mode high-power laser The topological insulator laser constructed from an aperiodic array of resonators was realized experimentally in an all-dielectric platform, as described in the accompanying experimental paper by Bandres et al

Construction of hierarchical Ni–Co–P hollow nanobricks with oriented nanosheets for efficient overall water splitting

Abstract: Complex nano-architectures with ordered two-dimensional (2D) building blocks are a class of promising electrocatalysts for different electrochemical technologies. In this work, a novel template-engaged strategy followed by sequential etching and phosphorization treatments is demonstrated to fabricate open and hierarchical Ni–Co–P hollow nanobricks (HNBs) via the assembly of oriented 2D nanosheets. Benefiting from the unique nano-architectures with large electrolyte-accessible surface and abundant mass diffusion pathways, the as-prepared Ni–Co–P HNBs exhibit high electrocatalytic activity, which affords the current density of 10 mA cm−2 at low overpotentials of 270 mV and 107 mV for oxygen and hydrogen evolution reactions respectively, and excellent stability in an alkaline medium. Remarkably, when used as both the anode and cathode, a low cell voltage of 1.62 V is required to reach the current density of 10 mA cm−2, making the Ni–Co–P HNBs an efficient bifunctional electrocatalyst for overall water splitting.

High phase-purity 1T′-MoS 2 - and 1T′-MoSe 2 -layered crystals

Abstract: Phase control plays an important role in the precise synthesis of inorganic materials, as the phase structure has a profound influence on properties such as conductivity and chemical stability. Phase-controlled preparation has been challenging for the metallic-phase group-VI transition metal dichalcogenides (the transition metals are Mo and W, and the chalcogens are S, Se and Te), which show better performance in electrocatalysis than their semiconducting counterparts. Here, we report the large-scale preparation of micrometre-sized metallic-phase 1T′-MoX2 (X = S, Se)-layered bulk crystals in high purity. We reveal that 1T′-MoS2 crystals feature a distorted octahedral coordination structure and are convertible to 2H-MoS2 following thermal annealing or laser irradiation. Electrochemical measurements show that the basal plane of 1T′-MoS2 is much more active than that of 2H-MoS2 for the electrocatalytic hydrogen evolution reaction in an acidic medium.

A Voltage-Boosting Strategy Enabling a Low-Frequency, Flexible Electromagnetic Wave Absorption Device.

Abstract: Nowadays, low-frequency electromagnetic interference (<2.0 GHz) remains a key core issue that plagues the effective attenuation performance of conventional absorption devices prepared via the component-morphology method (Strategy I). According to theoretical calculations, one fundamental solution is to develop a material that possesses a high e' but lower e″. Thus, it is attempted to control the dielectric values via applying an external electrical field, which inducts changes in the macrostructure toward a performance improvement (Strategy II). A sandwich-structured flexible electronic absorption device is designed using a carbon film electrode to conduct an external current. Simultaneously, an absorption layer that is highly responsive to an external voltage is selected via Strategy I. Relying on the synergistic effects from Strategies I and II, this device demonstrates an absorption value of more than 85% at 1.5-2.0 GHz with an applied voltage of 16 V while reducing the thickness to ≈5 mm. In addition, the device also shows a good absorption property at 25-150 °C. The method of utilizing an external voltage to break the intrinsic dielectric feature by modifying a traditional electronic absorption device is demonstrated for the first time and has great significance in solving the low-frequency electromagnetic interference issue.

Ambient Backscatter Communications: A Contemporary Survey

Abstract: Recently, ambient backscatter communication has been introduced as a cutting-edge technology which enables smart devices to communicate by utilizing ambient radio frequency (RF) signals without requiring active RF transmission. This technology is especially effective in addressing communication and energy efficiency problems for low-power communications systems such as sensor networks, and thus it is expected to realize numerous Internet-of-Things applications. Therefore, this paper aims to provide a contemporary and comprehensive literature review on fundamentals, applications, challenges, and research efforts/progress of ambient backscatter communications. In particular, we first present fundamentals of backscatter communications and briefly review bistatic backscatter communications systems. Then, the general architecture, advantages, and solutions to address existing issues and limitations of ambient backscatter communications systems are discussed. Additionally, emerging applications of ambient backscatter communications are highlighted. Finally, we outline some open issues and future research directions.

The influence of the molecular packing on the room temperature phosphorescence of purely organic luminogens.

Abstract: Organic luminogens with persistent room temperature phosphorescence (RTP) have attracted great attention for their wide applications in optoelectronic devices and bioimaging. However, these materials are still very scarce, partially due to the unclear mechanism and lack of designing guidelines. Herein we develop seven 10-phenyl-10H-phenothiazine-5,5-dioxide-based derivatives, reveal their different RTP properties and underlying mechanism, and exploit their potential imaging applications. Coupled with the preliminary theoretical calculations, it is found that strong π-π interactions in solid state can promote the persistent RTP. Particularly, CS-CF3 shows the unique photo-induced phosphorescence in response to the changes in molecular packing, further confirming the key influence of the molecular packing on the RTP property. Furthermore, CS-F with its long RTP lifetime could be utilized for real-time excitation-free phosphorescent imaging in living mice. Thus, our study paves the way for the development of persistent RTP materials, in both the practical applications and the inherent mechanism.

Mixed Metal Sulfides for Electrochemical Energy Storage and Conversion

Abstract: Mixed metal sulfides (MMSs) have attracted increased attention as promising electrode materials for electrochemical energy storage and conversion systems including lithium-ion batteries (LIBs), sodium-ion batteries (SIBs), hybrid supercapacitors (HSCs), metal–air batteries (MABs), and water splitting. Compared with monometal sulfides, MMSs exhibit greatly enhanced electrochemical performance, which is largely originated from their higher electronic conductivity and richer redox reactions. In this review, recent progresses in the rational design and synthesis of diverse MMS-based micro/nanostructures with controlled morphologies, sizes, and compositions for LIBs, SIBs, HSCs, MABs, and water splitting are summarized. In particular, nanostructuring, synthesis of nanocomposites with carbonaceous materials and fabrication of 3D MMS-based electrodes are demonstrated to be three effective approaches for improving the electrochemical performance of MMS-based electrode materials. Furthermore, some potential challenges as well as prospects are discussed to further advance the development of MMS-based electrode materials for next-generation electrochemical energy storage and conversion systems.

Two-Dimensional Metal Nanomaterials: Synthesis, Properties, and Applications

Abstract: As one unique group of two-dimensional (2D) nanomaterials, 2D metal nanomaterials have drawn increasing attention owing to their intriguing physiochemical properties and broad range of promising applications. In this Review, we briefly introduce the general synthetic strategies applied to 2D metal nanomaterials, followed by describing in detail the various synthetic methods classified in two categories, i.e. bottom-up methods and top-down methods. After introducing the unique physical and chemical properties of 2D metal nanomaterials, the potential applications of 2D metal nanomaterials in catalysis, surface enhanced Raman scattering, sensing, bioimaging, solar cells, and photothermal therapy are discussed in detail. Finally, the challenges and opportunities in this promising research area are proposed.

Formation of Hierarchical Co9S8@ZnIn2S4 Heterostructured Cages as an Efficient Photocatalyst for Hydrogen Evolution.

Abstract: Here we demonstrate the delicate design and construction of hierarchical Co9S8@ZnIn2S4 heterostructured cages as an efficient photocatalyst for hydrogen evolution with visible light. Two photoactive sulfide semiconductors are rationally integrated into a hierarchical hollow structure with strongly coupled heterogeneous shells and two-dimensional ultrathin subunits. The unique architecture can efficiently facilitate the separation and transfer of light-induced charges, offer large surface area, and expose rich active sites for photocatalytic redox reactions. Owing to the distinctive structural and compositional benefits, the hierarchical Co9S8@ZnIn2S4 hollow heterostructures without using any cocatalysts show remarkable activity with a hydrogen-producing rate of 6250 μmol h–1 g–1 and high stability for photocatalytic water splitting.

Machine Health Monitoring Using Local Feature-Based Gated Recurrent Unit Networks

Abstract: In modern industries, machine health monitoring systems (MHMS) have been applied wildly with the goal of realizing predictive maintenance including failures tracking, downtime reduction, and assets preservation. In the era of big machinery data, data-driven MHMS have achieved remarkable results in the detection of faults after the occurrence of certain failures (diagnosis) and prediction of the future working conditions and the remaining useful life (prognosis). The numerical representation for raw sensory data is the key stone for various successful MHMS. Conventional methods are the labor-extensive as they usually depend on handcrafted features, which require expert knowledge. Inspired by the success of deep learning methods that redefine representation learning from raw data, we propose local feature-based gated recurrent unit (LFGRU) networks. It is a hybrid approach that combines handcrafted feature design with automatic feature learning for machine health monitoring. First, features from windows of input time series are extracted. Then, an enhanced bidirectional GRU network is designed and applied on the generated sequence of local features to learn the representation. A supervised learning layer is finally trained to predict machine condition. Experiments on three machine health monitoring tasks: tool wear prediction, gearbox fault diagnosis, and incipient bearing fault detection verify the effectiveness and generalization of the proposed LFGRU.

Dye-Sensitized Solar Cells: Fundamentals and Current Status

Abstract: Dye-sensitized solar cells (DSSCs) belong to the group of thin-film solar cells which have been under extensive research for more than two decades due to their low cost, simple preparation methodology, low toxicity and ease of production. Still, there is lot of scope for the replacement of current DSSC materials due to their high cost, less abundance, and long-term stability. The efficiency of existing DSSCs reaches up to 12%, using Ru(II) dyes by optimizing material and structural properties which is still less than the efficiency offered by first- and second-generation solar cells, i.e., other thin-film solar cells and Si-based solar cells which offer ~ 20–30% efficiency. This article provides an in-depth review on DSSC construction, operating principle, key problems (low efficiency, low scalability, and low stability), prospective efficient materials, and finally a brief insight to commercialization.