Showing papers by "Harbin Institute of Technology published in 2020"

Epidemiology and transmission of COVID-19 in 391 cases and 1286 of their close contacts in Shenzhen, China: a retrospective cohort study.

Abstract: Summary Background Rapid spread of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in Wuhan, China, prompted heightened surveillance in Shenzhen, China. The resulting data provide a rare opportunity to measure key metrics of disease course, transmission, and the impact of control measures. Methods From Jan 14 to Feb 12, 2020, the Shenzhen Center for Disease Control and Prevention identified 391 SARS-CoV-2 cases and 1286 close contacts. We compared cases identified through symptomatic surveillance and contact tracing, and estimated the time from symptom onset to confirmation, isolation, and admission to hospital. We estimated metrics of disease transmission and analysed factors influencing transmission risk. Findings Cases were older than the general population (mean age 45 years) and balanced between males (n=187) and females (n=204). 356 (91%) of 391 cases had mild or moderate clinical severity at initial assessment. As of Feb 22, 2020, three cases had died and 225 had recovered (median time to recovery 21 days; 95% CI 20–22). Cases were isolated on average 4·6 days (95% CI 4·1–5·0) after developing symptoms; contact tracing reduced this by 1·9 days (95% CI 1·1–2·7). Household contacts and those travelling with a case were at higher risk of infection (odds ratio 6·27 [95% CI 1·49–26·33] for household contacts and 7·06 [1·43–34·91] for those travelling with a case) than other close contacts. The household secondary attack rate was 11·2% (95% CI 9·1–13·8), and children were as likely to be infected as adults (infection rate 7·4% in children Interpretation Our data on cases as well as their infected and uninfected close contacts provide key insights into the epidemiology of SARS-CoV-2. This analysis shows that isolation and contact tracing reduce the time during which cases are infectious in the community, thereby reducing the R. The overall impact of isolation and contact tracing, however, is uncertain and highly dependent on the number of asymptomatic cases. Moreover, children are at a similar risk of infection to the general population, although less likely to have severe symptoms; hence they should be considered in analyses of transmission and control. Funding Emergency Response Program of Harbin Institute of Technology, Emergency Response Program of Peng Cheng Laboratory, US Centers for Disease Control and Prevention.

ECA-Net: Efficient Channel Attention for Deep Convolutional Neural Networks

Abstract: Recently, channel attention mechanism has demonstrated to offer great potential in improving the performance of deep convolutional neural networks (CNNs). However, most existing methods dedicate to developing more sophisticated attention modules for achieving better performance, which inevitably increase model complexity. To overcome the paradox of performance and complexity trade-off, this paper proposes an Efficient Channel Attention (ECA) module, which only involves a handful of parameters while bringing clear performance gain. By dissecting the channel attention module in SENet, we empirically show avoiding dimensionality reduction is important for learning channel attention, and appropriate cross-channel interaction can preserve performance while significantly decreasing model complexity. Therefore, we propose a local cross-channel interaction strategy without dimensionality reduction, which can be efficiently implemented via 1D convolution. Furthermore, we develop a method to adaptively select kernel size of 1D convolution, determining coverage of local cross-channel interaction. The proposed ECA module is both efficient and effective, e.g., the parameters and computations of our modules against backbone of ResNet50 are 80 vs. 24.37M and 4.7e-4 GFlops vs. 3.86 GFlops, respectively, and the performance boost is more than 2% in terms of Top-1 accuracy. We extensively evaluate our ECA module on image classification, object detection and instance segmentation with backbones of ResNets and MobileNetV2. The experimental results show our module is more efficient while performing favorably against its counterparts.

CodeBERT: A Pre-Trained Model for Programming and Natural Languages

Abstract: We present CodeBERT, a bimodal pre-trained model for programming language (PL) and nat-ural language (NL). CodeBERT learns general-purpose representations that support downstream NL-PL applications such as natural language codesearch, code documentation generation, etc. We develop CodeBERT with Transformer-based neural architecture, and train it with a hybrid objective function that incorporates the pre-training task of replaced token detection, which is to detect plausible alternatives sampled from generators. This enables us to utilize both bimodal data of NL-PL pairs and unimodal data, where the former provides input tokens for model training while the latter helps to learn better generators. We evaluate CodeBERT on two NL-PL applications by fine-tuning model parameters. Results show that CodeBERT achieves state-of-the-art performance on both natural language code search and code documentation generation tasks. Furthermore, to investigate what type of knowledge is learned in CodeBERT, we construct a dataset for NL-PL probing, and evaluate in a zero-shot setting where parameters of pre-trained models are fixed. Results show that CodeBERT performs better than previous pre-trained models on NL-PL probing.

Identifying and removing haplotypic duplication in primary genome assemblies.

Abstract: Motivation Rapid development in long-read sequencing and scaffolding technologies is accelerating the production of reference-quality assemblies for large eukaryotic genomes. However, haplotype divergence in regions of high heterozygosity often results in assemblers creating two copies rather than one copy of a region, leading to breaks in contiguity and compromising downstream steps such as gene annotation. Several tools have been developed to resolve this problem. However, they either focus only on removing contained duplicate regions, also known as haplotigs, or fail to use all the relevant information and hence make errors. Results Here we present a novel tool, purge_dups, that uses sequence similarity and read depth to automatically identify and remove both haplotigs and heterozygous overlaps. In comparison with current tools, we demonstrate that purge_dups can reduce heterozygous duplication and increase assembly continuity while maintaining completeness of the primary assembly. Moreover, purge_dups is fully automatic and can easily be integrated into assembly pipelines. Availability and implementation The source code is written in C and is available at https://github.com/dfguan/purge_dups. Supplementary information Supplementary data are available at Bioinformatics online.

Towards complete and error-free genome assemblies of all vertebrate species

Abstract: High-quality and complete reference genome assemblies are fundamental for the application of genomics to biology, disease, and biodiversity conservation. However, such assemblies are only available for a few non-microbial species. To address this issue, the international Genome 10K (G10K) consortium has worked over a five-year period to evaluate and develop cost-effective methods for assembling the most accurate and complete reference genomes to date. Here we summarize these developments, introduce a set of quality standards, and present lessons learned from sequencing and assembling 16 species representing major vertebrate lineages (mammals, birds, reptiles, amphibians, teleost fishes and cartilaginous fishes). We confirm that long-read sequencing technologies are essential for maximizing genome quality and that unresolved complex repeats and haplotype heterozygosity are major sources of error in assemblies. Our new assemblies identify and correct substantial errors in some of the best historical reference genomes. Adopting these lessons, we have embarked on the Vertebrate Genomes Project (VGP), an effort to generate high-quality, complete reference genomes for all ~70,000 extant vertebrate species and help enable a new era of discovery across the life sciences.

Room-temperature phosphorescence from organic aggregates

Abstract: Triplet excitons in organic molecules underscore a variety of processes and technologies as a result of their long lifetime and spin multiplicity Organic phosphorescence, which originates from triplet excitons, has potential for the development of a new generation of organic optoelectronic materials and biomedical agents However, organic phosphorescence is typically only observed at cryogenic temperatures and under inert conditions in solution, which severely restricts its practical applications In the past few years, room-temperature-phosphorescent systems have been obtained based on organic aggregates Rapid advances in molecular-structure design and aggregation-behaviour modulation have enabled substantial progress, but the mechanistic picture is still not fully understood because of the high sensitivity and complexity of triplet-exciton behaviour This Review analyses key photophysical processes related to triplet excitons, including intersystem crossing, radiative and non-radiative decay, and quenching processes, to illustrate the intrinsic structure–property relationships and draw clear and integrated design principles The resulting strategies for the development of efficient and persistent room-temperature-phosphorescent systems are discussed, and newly emerged applications based on these materials are highlighted Advances in molecular-structure design and modulation of the aggregation behaviour have enabled much progress in the observation of room-temperature phosphorescence from organic aggregates This Review analyses key photophysical processes related to triplet excitons, illustrating the intrinsic structure–property relationships and identifying strategies to design efficient and persistent room-temperature-phosphorescent systems

Structural transformation of highly active metal–organic framework electrocatalysts during the oxygen evolution reaction

Abstract: Metal–organic frameworks (MOFs) are increasingly being investigated as electrocatalysts for the oxygen evolution reaction (OER). Despite their promising catalytic activity, many fundamental questions concerning their structure−performance relationships—especially those regarding the roles of active species—remain to be answered. Here we show the structural transformation of a Ni0.5Co0.5-MOF-74 during the OER by operando X-ray absorption spectroscopy analysis and high-resolution transmission electron microscopy imaging. We suggest that Ni0.5Co0.5OOH0.75, with abundant oxygen vacancies and high oxidation states, forms in situ and is responsible for the high OER activity observed. The ratio of Ni to Co in the bimetallic centres alters the geometric and electronic structure of as-formed active species and in turn the catalytic activity. Based on our understanding of this system, we fabricate a Ni0.9Fe0.1-MOF that delivers low overpotentials of 198 mV and 231 mV at 10 mA cm−2 and 20 mA cm−2, respectively. Metal–organic frameworks (MOFs) are increasingly being explored for electrocatalytic oxygen evolution, which is half of the water splitting reaction. Here the authors show that, under reaction conditions, mixed metal oxyhydroxides form at the nodes of bimetallic MOFs, which are highly catalytically active.

Deep Residual Shrinkage Networks for Fault Diagnosis

Abstract: This article develops new deep learning methods, namely, deep residual shrinkage networks, to improve the feature learning ability from highly noised vibration signals and achieve a high fault diagnosing accuracy. Soft thresholding is inserted as nonlinear transformation layers into the deep architectures to eliminate unimportant features. Moreover, considering that it is generally challenging to set proper values for the thresholds, the developed deep residual shrinkage networks integrate a few specialized neural networks as trainable modules to automatically determine the thresholds, so that professional expertise on signal processing is not required. The efficacy of the developed methods is validated through experiments with various types of noise.

Voltage issue of aqueous rechargeable metal-ion batteries

Abstract: Over the past two decades, a series of aqueous rechargeable metal-ion batteries (ARMBs) have been developed, aiming at improving safety, environmental friendliness and cost-efficiency in fields of consumer electronics, electric vehicles and grid-scale energy storage. However, the notable gap between ARMBs and their organic counterparts in energy density directly hinders their practical applications, making it difficult to replace current widely-used organic lithium-ion batteries. Basically, this huge gap in energy density originates from cell voltage, as the narrow electrochemical stability window of aqueous electrolytes substantially confines the choice of electrode materials. This review highlights various ARMBs with focuses on their voltage characteristics and strategies that can effectively raise battery voltage. It begins with the discussion on the fundamental factor that limits the voltage of ARMBs, i.e., electrochemical stability window of aqueous electrolytes, which decides the maximum-allowed potential difference between cathode and anode. The following section introduces various ARMB systems and compares their voltage characteristics in midpoint voltage and plateau voltage, in relation to respective electrode materials. Subsequently, various strategies paving the way to high-voltage ARMBs are summarized, with corresponding advancements highlighted. The final section presents potential directions for further improvements and future perspectives of this thriving field.

DDcGAN: A Dual-Discriminator Conditional Generative Adversarial Network for Multi-Resolution Image Fusion

Abstract: In this paper, we proposed a new end-to-end model, termed as dual-discriminator conditional generative adversarial network (DDcGAN), for fusing infrared and visible images of different resolutions. Our method establishes an adversarial game between a generator and two discriminators. The generator aims to generate a real-like fused image based on a specifically designed content loss to fool the two discriminators, while the two discriminators aim to distinguish the structure differences between the fused image and two source images, respectively, in addition to the content loss. Consequently, the fused image is forced to simultaneously keep the thermal radiation in the infrared image and the texture details in the visible image. Moreover, to fuse source images of different resolutions, e.g. , a low-resolution infrared image and a high-resolution visible image, our DDcGAN constrains the downsampled fused image to have similar property with the infrared image. This can avoid causing thermal radiation information blurring or visible texture detail loss, which typically happens in traditional methods. In addition, we also apply our DDcGAN to fusing multi-modality medical images of different resolutions, e.g. , a low-resolution positron emission tomography image and a high-resolution magnetic resonance image. The qualitative and quantitative experiments on publicly available datasets demonstrate the superiority of our DDcGAN over the state-of-the-art, in terms of both visual effect and quantitative metrics. Our code is publicly available at https://github.com/jiayi-ma/DDcGAN .

Ultrafast control of vortex microlasers.

Abstract: The development of classical and quantum information–processing technology calls for on-chip integrated sources of structured light. Although integrated vortex microlasers have been previously demonstrated, they remain static and possess relatively high lasing thresholds, making them unsuitable for high-speed optical communication and computing. We introduce perovskite-based vortex microlasers and demonstrate their application to ultrafast all-optical switching at room temperature. By exploiting both mode symmetry and far-field properties, we reveal that the vortex beam lasing can be switched to linearly polarized beam lasing, or vice versa, with switching times of 1 to 1.5 picoseconds and energy consumption that is orders of magnitude lower than in previously demonstrated all-optical switching. Our results provide an approach that breaks the long-standing trade-off between low energy consumption and high-speed nanophotonics, introducing vortex microlasers that are switchable at terahertz frequencies.

LayoutLM: Pre-training of Text and Layout for Document Image Understanding

Abstract: Pre-training techniques have been verified successfully in a variety of NLP tasks in recent years. Despite the widespread use of pre-training models for NLP applications, they almost exclusively focus on text-level manipulation, while neglecting layout and style information that is vital for document image understanding. In this paper, we propose the LayoutLM to jointly model interactions between text and layout information across scanned document images, which is beneficial for a great number of real-world document image understanding tasks such as information extraction from scanned documents. Furthermore, we also leverage image features to incorporate words' visual information into LayoutLM. To the best of our knowledge, this is the first time that text and layout are jointly learned in a single framework for document-level pre-training. It achieves new state-of-the-art results in several downstream tasks, including form understanding (from 70.72 to 79.27), receipt understanding (from 94.02 to 95.24) and document image classification (from 93.07 to 94.42). The code and pre-trained LayoutLM models are publicly available at https://aka.ms/layoutlm.

MiniLM: Deep Self-Attention Distillation for Task-Agnostic Compression of Pre-Trained Transformers

Abstract: Pre-trained language models (e.g., BERT (Devlin et al., 2018) and its variants) have achieved remarkable success in varieties of NLP tasks. However, these models usually consist of hundreds of millions of parameters which brings challenges for fine-tuning and online serving in real-life applications due to latency and capacity constraints. In this work, we present a simple and effective approach to compress large Transformer (Vaswani et al., 2017) based pre-trained models, termed as deep self-attention distillation. The small model (student) is trained by deeply mimicking the self-attention module, which plays a vital role in Transformer networks, of the large model (teacher). Specifically, we propose distilling the self-attention module of the last Transformer layer of the teacher, which is effective and flexible for the student. Furthermore, we introduce the scaled dot-product between values in the self-attention module as the new deep self-attention knowledge, in addition to the attention distributions (i.e., the scaled dot-product of queries and keys) that have been used in existing works. Moreover, we show that introducing a teacher assistant (Mirzadeh et al., 2019) also helps the distillation of large pre-trained Transformer models. Experimental results demonstrate that our monolingual model outperforms state-of-the-art baselines in different parameter size of student models. In particular, it retains more than 99% accuracy on SQuAD 2.0 and several GLUE benchmark tasks using 50% of the Transformer parameters and computations of the teacher model. We also obtain competitive results in applying deep self-attention distillation to multilingual pre-trained models.

Revisiting Pre-Trained Models for Chinese Natural Language Processing

Abstract: Bidirectional Encoder Representations from Transformers (BERT) has shown marvelous improvements across various NLP tasks, and consecutive variants have been proposed to further improve the performance of the pre-trained language models. In this paper, we target on revisiting Chinese pre-trained language models to examine their effectiveness in a non-English language and release the Chinese pre-trained language model series to the community. We also propose a simple but effective model called MacBERT, which improves upon RoBERTa in several ways, especially the masking strategy that adopts MLM as correction (Mac). We carried out extensive experiments on eight Chinese NLP tasks to revisit the existing pre-trained language models as well as the proposed MacBERT. Experimental results show that MacBERT could achieve state-of-the-art performances on many NLP tasks, and we also ablate details with several findings that may help future research. https://github.com/ymcui/MacBERT

GraphCodeBERT: Pre-training Code Representations with Data Flow

Abstract: Pre-trained models for programming language have achieved dramatic empirical improvements on a variety of code-related tasks such as code search, code completion, code summarization, etc. However, existing pre-trained models regard a code snippet as a sequence of tokens, while ignoring the inherent structure of code, which provides crucial code semantics and would enhance the code understanding process. We present GraphCodeBERT, a pre-trained model for programming language that considers the inherent structure of code. Instead of taking syntactic-level structure of code like abstract syntax tree (AST), we use data flow in the pre-training stage, which is a semantic-level structure of code that encodes the relation of "where-the-value-comes-from" between variables. Such a semantic-level structure is neat and does not bring an unnecessarily deep hierarchy of AST, the property of which makes the model more efficient. We develop GraphCodeBERT based on Transformer. In addition to using the task of masked language modeling, we introduce two structure-aware pre-training tasks. One is to predict code structure edges, and the other is to align representations between source code and code structure. We implement the model in an efficient way with a graph-guided masked attention function to incorporate the code structure. We evaluate our model on four tasks, including code search, clone detection, code translation, and code refinement. Results show that code structure and newly introduced pre-training tasks can improve GraphCodeBERT and achieves state-of-the-art performance on the four downstream tasks. We further show that the model prefers structure-level attentions over token-level attentions in the task of code search.

U2Fusion: A Unified Unsupervised Image Fusion Network.

Abstract: This study proposes a novel unified and unsupervised end-to-end image fusion network, termed as U2Fusion, which is capable of solving different fusion problems, including multi-modal, multi-exposure, and multi-focus cases. Using feature extraction and information measurement, U2Fusion automatically estimates the importance of corresponding source images and comes up with adaptive information preservation degrees. Hence, different fusion tasks are unified in the same framework. Based on the adaptive degrees, a network is trained to preserve the adaptive similarity between the fusion result and source images. Therefore, the stumbling blocks in applying deep learning for image fusion, e.g., the requirement of ground-truth and specifically designed metrics, are greatly mitigated. By avoiding the loss of previous fusion capabilities when training a single model for different tasks sequentially, we obtain a unified model that is applicable to multiple fusion tasks. Moreover, a new aligned infrared and visible image dataset, RoadScene (available at https://github.com/hanna-xu/RoadScene), is released to provide a new option for benchmark evaluation. Qualitative and quantitative experimental results on three typical image fusion tasks validate the effectiveness and universality of U2Fusion. Our code is publicly available at https://github.com/hanna-xu/U2Fusion.

Deeply understanding the Zn anode behaviour and corresponding improvement strategies in different aqueous Zn-based batteries

Abstract: Owing to the high capacity of the metallic Zn anode and intrinsically safe aqueous electrolyte, aqueous Zn-based batteries are advanced energy storage technology alternatives beyond lithium-ion batteries, providing a cost benefit, high safety, and competitive energy density. There has been a new wave of research interest across the family of Zn batteries, but fundamental understanding of the Zn electrode and its performance improvement still remain inconclusive. Based on the pH value of the electrolyte, Zn-based batteries can be divided into two types, with one adopting alkaline electrolyte and the other mild (including slightly acidic) electrolyte. As the behavior of the Zn electrode in these two distinctive systems is different, their requirements to yield excellent performance are different. In this Review, we present a comprehensive overview of the Zn electrode and its fundamentals in both systems. First, the differences and similarities of the Zn electrode in both systems are outlined. Specific attention is paid to the working principles and technical challenges. Then, Zn electrode issues and recently proposed strategies for each system are summarized and compared. Finally, a perspective on future research directions towards practical applications of aqueous Zn batteries is included.

A Novel Method.

Abstract: Parameter selection is one of the most important parts for nearly all the control strategies. Traditionally, controller parameters are chosen by utilizing trial and error, which is always tedious and time consuming. Moreover, such method is highly dependent on the experience of researchers, which means that it is hard to be popularized. In this light, this paper proposes a novel parameter searching approach by utilizing uniform design (UD) algorithm. By which the satisfactory controller parameters under a performance index could be selected. In this end, two simulation examples are conducted to verify the effectiveness of proposed scheme. Simulation results show that this novel approach, as compared to other intelligent tuning algorithms, excels in efficiency and time saving.

Multi-Scale Progressive Fusion Network for Single Image Deraining

Abstract: Rain streaks in the air appear in various blurring degrees and resolutions due to different distances from their positions to the camera. Similar rain patterns are visible in a rain image as well as its multi-scale (or multi-resolution) versions, which makes it possible to exploit such complementary information for rain streak representation. In this work, we explore the multi-scale collaborative representation for rain streaks from the perspective of input image scales and hierarchical deep features in a unified framework, termed multi-scale progressive fusion network (MSPFN) for single image rain streak removal. For the similar rain streaks at different positions, we employ recurrent calculation to capture the global texture, thus allowing to explore the complementary and redundant information at the spatial dimension to characterize target rain streaks. Besides, we construct multi-scale pyramid structure, and further introduce the attention mechanism to guide the fine fusion of these correlated information from different scales. This multi-scale progressive fusion strategy not only promotes the cooperative representation, but also boosts the end-to-end training. Our proposed method is extensively evaluated on several benchmark datasets and achieves the state-of-the-art results. Moreover, we conduct experiments on joint deraining, detection, and segmentation tasks, and inspire a new research direction of vision task driven image deraining. The source code is available at https://github.com/kuihua/MSPFN.

Siamese Box Adaptive Network for Visual Tracking

Abstract: Most of the existing trackers usually rely on either a multi-scale searching scheme or pre-defined anchor boxes to accurately estimate the scale and aspect ratio of a target. Unfortunately, they typically call for tedious and heuristic configurations. To address this issue, we propose a simple yet effective visual tracking framework (named Siamese Box Adaptive Network, SiamBAN) by exploiting the expressive power of the fully convolutional network (FCN). SiamBAN views the visual tracking problem as a parallel classification and regression problem, and thus directly classifies objects and regresses their bounding boxes in a unified FCN. The no-prior box design avoids hyper-parameters associated with the candidate boxes, making SiamBAN more flexible and general. Extensive experiments on visual tracking benchmarks including VOT2018, VOT2019, OTB100, NFS, UAV123, and LaSOT demonstrate that SiamBAN achieves state-of-the-art performance and runs at 40 FPS, confirming its effectiveness and efficiency. The code will be available at https://github.com/hqucv/siamban.

Tailoring the d-Band Centers Endows (NixFe1–x)2P Nanosheets with Efficient Oxygen Evolution Catalysis

Abstract: The rational design of oxygen evolution reaction (OER) catalysts from the perspective of electronic structure is highly desirable to optimize electrocatalytic activity. Monometallic phosphides such...

Transparent ferroelectric crystals with ultrahigh piezoelectricity

Abstract: Transparent piezoelectrics are highly desirable for numerous hybrid ultrasound–optical devices ranging from photoacoustic imaging transducers to transparent actuators for haptic applications1–7. However, it is challenging to achieve high piezoelectricity and perfect transparency simultaneously because most high-performance piezoelectrics are ferroelectrics that contain high-density light-scattering domain walls. Here, through a combination of phase-field simulations and experiments, we demonstrate a relatively simple method of using an alternating-current electric field to engineer the domain structures of originally opaque rhombohedral Pb(Mg1/3Nb2/3)O3-PbTiO3 (PMN-PT) crystals to simultaneously generate near-perfect transparency, an ultrahigh piezoelectric coefficient d33 (greater than 2,100 picocoulombs per newton), an excellent electromechanical coupling factor k33 (about 94 per cent) and a large electro-optical coefficient γ33 (approximately 220 picometres per volt), which is far beyond the performance of the commonly used transparent ferroelectric crystal LiNbO3. We find that increasing the domain size leads to a higher d33 value for the [001]-oriented rhombohedral PMN-PT crystals, challenging the conventional wisdom that decreasing the domain size always results in higher piezoelectricity8–10. This work presents a paradigm for achieving high transparency and piezoelectricity by ferroelectric domain engineering, and we expect the transparent ferroelectric crystals reported here to provide a route to a wide range of hybrid device applications, such as medical imaging, self-energy-harvesting touch screens and invisible robotic devices. The use of alternating-current electric fields to control domain size in ferroelectric crystals affords excellent transparency, piezoelectricity and birefringence.

CodeBERT: A Pre-Trained Model for Programming and Natural Languages

Abstract: We present CodeBERT, a bimodal pre-trained model for programming language (PL) and natural language (NL). CodeBERT learns general-purpose representations that support downstream NL-PL applications such as natural language code search, code documentation generation, etc. We develop CodeBERT with Transformer-based neural architecture, and train it with a hybrid objective function that incorporates the pre-training task of replaced token detection, which is to detect plausible alternatives sampled from generators. This enables us to utilize both “bimodal” data of NL-PL pairs and “unimodal data, where the former provides input tokens for model training while the latter helps to learn better generators. We evaluate CodeBERT on two NL-PL applications by fine-tuning model parameters. Results show that CodeBERT achieves state-of-the-art performance on both natural language code search and code documentation generation. Furthermore, to investigate what type of knowledge is learned in CodeBERT, we construct a dataset for NL-PL probing, and evaluate in a zero-shot setting where parameters of pre-trained models are fixed. Results show that CodeBERT performs better than previous pre-trained models on NLPL probing.

Position Sensorless Permanent Magnet Synchronous Machine Drives—A Review

Abstract: Owing to the competitive advantages of cost reduction, system downsizing, and reliability enhancement, position sensorless control methods for permanent magnet synchronous machine drives have drawn increasing attention from academia to industrial applications. In this article, a survey of the major sensorless control techniques for a wide speed range from low to high speeds is presented. The different high frequency signal injection schemes, fundamental pulsewidth modulation excitation methods, and model-based sensorless control are displayed and compared, which is able to facilitate the sensorless control implementation.

Grain-orientation-engineered multilayer ceramic capacitors for energy storage applications.

Abstract: Dielectric ceramics are highly desired for electronic systems owing to their fast discharge speed and excellent fatigue resistance. However, the low energy density resulting from the low breakdown electric field leads to inferior volumetric efficiency, which is the main challenge for practical applications of dielectric ceramics. Here, we propose a strategy to increase the breakdown electric field and thus enhance the energy storage density of polycrystalline ceramics by controlling grain orientation. We fabricated high-quality -textured Na0.5Bi0.5TiO3–Sr0.7Bi0.2TiO3 (NBT-SBT) ceramics, in which the strain induced by the electric field is substantially lowered, leading to a reduced failure probability and improved Weibull breakdown strength, on the order of 103 MV m−1, an ~65% enhancement compared to their randomly oriented counterparts. The recoverable energy density of -textured NBT-SBT multilayer ceramics is up to 21.5 J cm−3, outperforming state-of-the-art dielectric ceramics. The present research offers a route for designing dielectric ceramics with enhanced breakdown strength, which is expected to benefit a wide range of applications of dielectric ceramics for which high breakdown strength is required, such as high-voltage capacitors and electrocaloric solid-state cooling devices. The energy density of dielectric ceramic capacitors is limited by low breakdown fields. Here, by considering the anisotropy of electrostriction in perovskites, it is shown that -textured Na0.5Bi0.5TiO3–Sr0.7Bi0.2TiO3 ceramics can sustain higher electrical fields and achieve an energy density of 21.5 J cm−3.