基于改进TextRank的科技文本关键词抽取方法

针对科技文本关键词抽取任务中抽取出现次数少但能较好表达文本主旨的词语效果差的问题,提出一种基于改进TextRank的关键词抽取方法。首先,利用词语的词频-逆文档频率(TF-IDF)统计特征和位置特征优化共现图中词语间的概率转移矩阵,通过迭代计算得到词语的初始得分;然后,利用K-Core(K-Core decomposition)算法挖掘KCore子图得到词语的层级特征,利用平均信息熵特征衡量词语的主题表征能力;最后,在词语初始得分的基础上融合层级特征和平均信息熵特征,从而确定关键词。实验结果表明,在公开数据集上,与TextRank方法和OTextRank(Optimized TextRank)方法相比,所提方法在抽取不同关键词数量的实验中,F1均值分别提高了6.5和3.3个百分点;在科技服务项目数据集上,与TextRank方法和OTextRank方法相比,所提方法在抽取不同关键词数量的实验中,F1均值分别提高了7.4和3.2个百分点。实验结果验证了所提方法抽取出现频率低但较好表达文本主旨关键词的有效性。

一种利用词典扩展数据库模式信息的Text2SQL方法

现有Text2SQL方法严重依赖表名和列名在自然语言查询中的显式提及,在同物异名的实际应用场景中准确率急剧下降.此外,这些方法仅仅依赖数据库模式捕捉数据库建模的领域知识,而数据库模式作为结构化的元数据,其表达领域知识的能力是非常有限的,即使有经验的程序员也很难仅从数据库模式完全领会该数据库建模的领域知识,因此程序员必须依赖详细的数据库设计文档才能构造SQL语句以正确地表达特定的查询.为此,本文提出一种利用词典扩展数据库模式信息的Text2SQL方法,该方法从数据库表名和列名解析出其中的单词或短语,查询词典获取这些单词或短语的语义解释,将这些解释看成是相应表名或列名的扩展内容,与表名、列名及其他数据库模式信息(主键、外键等)相结合,作为模型的输入,从而使模型能够更全面地学习数据库建模的应用领域知识.在Spider-syn和Spider数据集上进行的实验说明了所提出方法的有效性,即使自然语言查询中使用的表名和列名与数据库模式中对应的表名和列名完全不同,本文方法也能够得到较好的SQL翻译结果,明显优于最新提出的抗同义词替换攻击的方法.

In-situ interfacial passivation and self-adaptability synergistically stabilizing all-solid-state lithium metal batteries

The function of solid electrolytes and the composition of solid electrolyte interphase(SEI)are highly significant for inhibiting the growth of Li dendrites.Herein,we report an in-situ interfacial passivation combined with self-adaptability strategy to reinforce Li_(0.33)La_(0.557)TiO_(3)(LLTO)-based solid-state batteries.Specifically,a functional SEI enriched with LiF/Li_(3)PO_(4) is formed by in-situ electrochemical conversion,which is greatly beneficial to improving interface compatibility and enhancing ion transport.While the polarized dielectric BaTiO_(3)-polyamic acid(BTO-PAA,BP)film greatly improves the Li-ion transport kinetics and homogenizes the Li deposition.As expected,the resulting electrolyte offers considerable ionic conductivity at room temperature(4.3 x 10~(-4)S cm^(-1))and appreciable electrochemical decomposition voltage(5.23 V)after electrochemical passivation.For Li-LiFePO_(4) batteries,it shows a high specific capacity of 153 mA h g^(-1)at 0.2C after 100 cycles and a long-term durability of 115 mA h g^(-1)at 1.0 C after 800 cycles.Additionally,a stable Li plating/stripping can be achieved for more than 900 h at 0.5 mA cm^(-2).The stabilization mechanisms are elucidated by ex-situ XRD,ex-situ XPS,and ex-situ FTIR techniques,and the corresponding results reveal that the interfacial passivation combined with polarization effect is an effective strategy for improving the electrochemical performance.The present study provides a deeper insight into the dynamic adjustment of electrode-electrolyte interfacial for solid-state lithium batteries.

Interplay of laser power and pore characteristics in selective laser melting of ZK60 magnesium alloys:A study based on in-situ monitoring and image analysis

This study offers significant insights into the multi-physics phenomena of the SLM process and the subsequent porosity characteristics of ZK60 Magnesium(Mg)alloys.High-speed in-situ monitoring was employed to visualise process signals in real-time,elucidating the dynamics of melt pools and vapour plumes under varying laser power conditions specifically between 40 W and 60 W.Detailed morphological analysis was performed using Scanning-Electron Microscopy(SEM),demonstrating a critical correlation between laser power and pore formation.Lower laser power led to increased pore coverage,whereas a denser structure was observed at higher laser power.This laser power influence on porosity was further confirmed via Optical Microscopy(OM)conducted on both top and cross-sectional surfaces of the samples.An increase in laser power resulted in a decrease in pore coverage and pore size,potentially leading to a denser printed part of Mg alloy.X-ray Computed Tomography(XCT)augmented these findings by providing a 3D volumetric representation of the sample internal structure,revealing an inverse relationship between laser power and overall pore volume.Lower laser power appeared to favour the formation of interconnected pores,while a reduction in interconnected pores and an increase in isolated pores were observed at higher power.The interplay between melt pool size,vapour plume effects,and laser power was found to significantly influence the resulting porosity,indicating a need for effective management of these factors to optimise the SLM process of Mg alloys.

基于DAN与FastText的藏文短文本分类研究

随着藏文信息不断融入社会生活,越来越多的藏文短文本数据存在网络平台上。针对传统分类方法在藏文短文本上分类性能低的问题,文中提出了一种基于DAN-FastText的藏文短文本分类模型。该模型使用FastText网络在较大规模的藏文语料上进行无监督训练获得预训练的藏文音节向量集,使用预训练的音节向量集将藏文短文本信息转化为音节向量,把音节向量送入DAN(Deep Averaging Networks)网络并在输出阶段融合经过FastText网络训练的句向量特征,最后通过全连接层和softmax层完成分类。在公开的TNCC(Tibetan News Classification Corpus)新闻标题数据集上所提模型的Macro-F1是64.53%,比目前最好评测结果TiBERT模型的Macro-F1得分高出2.81%,比GCN模型的Macro-F1得分高出6.14%,融合模型具有较好的藏文短文本分类效果。

In-situ coating and surface partial protonation co-promoting performance of single-crystal nickel-rich cathode in all-solid-state batteries

The poor electrochemical performance of all-solid-state batteries(ASSBs),which is assemblied by Ni-rich cathode and poly(ethylene oxide)(PEO)-based electrolytes,can be attributed to unstable cathodic interface and poor crystal structure stability of Ni-rich cathode.Several coating strategies are previously employed to enhance the stability of the cathodic interface and crystal structure for Ni-rich cathode.However,these methods can hardly achieve simplicity and high efficiency simultaneously.In this work,polyacrylic acid(PAA)replaced traditional PVDF as a binder for cathode,which can achieve a uniform PAA-Li(LixPAA(0<x≤1))coating layer on the surface of single-crystal LiNi_(0.83)Co_(0.12)Mn_(0.05)O_(2)(SC-NCM83)due to H^(+)/Li^(+)exchange reaction during the initial charging-discharging process.The formation of PAA-Li coating layer on cathode can promote interfacial Li^(+)transport and enhance the stability of the cathodic interface.Furthermore,the partially-protonated surface of SC-NCM83 casued by H^(+)/Li^(+)exchange reaction can restrict Ni ions transport to enhance the crystal structure stability.The proposed SC-NCM83-PAA exhibits superior cycling performance with a retention of 92%compared with that(57.3%)of SC-NCM83-polyvinylidene difluoride(PVDF)after 200 cycles.This work provides a practical strategy to construct high-performance cathodes for ASSBs.

基于BERT和TextCNN的智能制造成熟度评估方法

随着智能制造2025目标的临近,企业为了解自身能力水平纷纷加入到智能制造成熟度评估的行列中。然而,由于智能制造成熟度评估标准的复杂性,企业缺乏其对行业水平的了解,导致企业贸然申请,浪费自身时间的同时又占用大量评估资源。鉴于此,设计了一种新的评估流程,采用文本处理算法对整个评估过程进行了重构,通过利用国标文件中智能制造成熟度评估标准,将其作为训练集,采用基于预训练语言模型与文本神经网络(BERT+TextCNN)相结合的智能评估算法代替人工评估。在真实的企业智能制造数据集上的验证表明,当BERT+TextCNN评估模型在卷积核为[2,3,4]、迭代次数为6次、学习率为3e-5时,对智能制造成熟度进行评估,准确率达到85.32%。这表明所设计的评估方法能够较准确地帮助企业完成智能制造成熟度自评估,有助于企业了解自身智能制造能力水平,制定正确的发展方向。

树立行业发展新方向——Techtextil&Texprocess 2024亮点回顾

为期4天的法兰克福国际产业用纺织品及非织造布展览会及国际纺织品及柔性材料缝制加工展览会(Techtextil&Texprocess 2024)吸引了来自全球53个国家和地区的1700家领先企业参展和来自102个国家和地区的38000名观众,展会规模再创新高,充分彰显了纺织行业蓬勃的生命力与持续的创新力。

Identifying multidisciplinary problems from scientific publications based on a text generation method

Purpose:A text generation based multidisciplinary problem identification method is proposed,which does not rely on a large amount of data annotation.Design/methodology/approach:The proposed method first identifies the research objective types and disciplinary labels of papers using a text classification technique;second,it generates abstractive titles for each paper based on abstract and research objective types using a generative pre-trained language model;third,it extracts problem phrases from generated titles according to regular expression rules;fourth,it creates problem relation networks and identifies the same problems by exploiting a weighted community detection algorithm;finally,it identifies multidisciplinary problems based on the disciplinary labels of papers.Findings:Experiments in the“Carbon Peaking and Carbon Neutrality”field show that the proposed method can effectively identify multidisciplinary research problems.The disciplinary distribution of the identified problems is consistent with our understanding of multidisciplinary collaboration in the field.Research limitations:It is necessary to use the proposed method in other multidisciplinary fields to validate its effectiveness.Practical implications:Multidisciplinary problem identification helps to gather multidisciplinary forces to solve complex real-world problems for the governments,fund valuable multidisciplinary problems for research management authorities,and borrow ideas from other disciplines for researchers.Originality/value:This approach proposes a novel multidisciplinary problem identification method based on text generation,which identifies multidisciplinary problems based on generative abstractive titles of papers without data annotation required by standard sequence labeling techniques.

CVTD: A Robust Car-Mounted Video Text Detector

Text perception is crucial for understanding the semantics of outdoor scenes,making it a key requirement for building intelligent systems for driver assistance or autonomous driving.Text information in car-mounted videos can assist drivers in making decisions.However,Car-mounted video text images pose challenges such as complex backgrounds,small fonts,and the need for real-time detection.We proposed a robust Car-mounted Video Text Detector(CVTD).It is a lightweight text detection model based on ResNet18 for feature extraction,capable of detecting text in arbitrary shapes.Our model efficiently extracted global text positions through the Coordinate Attention Threshold Activation(CATA)and enhanced the representation capability through stacking two Feature Pyramid Enhancement Fusion Modules(FPEFM),strengthening feature representation,and integrating text local features and global position information,reinforcing the representation capability of the CVTD model.The enhanced feature maps,when acted upon by Text Activation Maps(TAM),effectively distinguished text foreground from non-text regions.Additionally,we collected and annotated a dataset containing 2200 images of Car-mounted Video Text(CVT)under various road conditions for training and evaluating our model’s performance.We further tested our model on four other challenging public natural scene text detection benchmark datasets,demonstrating its strong generalization ability and real-time detection speed.This model holds potential for practical applications in real-world scenarios.

From text to image:challenges in integrating vision into ChatGPT for medical image interpretation

Large language models(LLMs),such as ChatGPT developed by OpenAI,represent a significant advancement in artificial intelligence(AI),designed to understand,generate,and interpret human language by analyzing extensive text data.Their potential integration into clinical settings offers a promising avenue that could transform clinical diagnosis and decision-making processes in the future(Thirunavukarasu et al.,2023).This article aims to provide an in-depth analysis of LLMs’current and potential impact on clinical practices.Their ability to generate differential diagnosis lists underscores their potential as invaluable tools in medical practice and education(Hirosawa et al.,2023;Koga et al.,2023).

Method to Remove Handwritten Texts Using Smart Phone

To remove handwritten texts from an image of a document taken by smart phone,an intelligent removal method was proposed that combines dewarping and Fully Convolutional Network with Atrous Convolutional and Atrous Spatial Pyramid Pooling(FCN-AC-ASPP).For a picture taken by a smart phone,firstly,the image is transformed into a regular image by the dewarping algorithm.Secondly,the FCN-AC-ASPP is used to classify printed texts and handwritten texts.Lastly,handwritten texts can be removed by a simple algorithm.Experiments show that the classification accuracy of the FCN-AC-ASPP is better than FCN,DeeplabV3+,FCN-AC.For handwritten texts removal effect,the method of combining dewarping and FCN-AC-ASPP is superior to FCN-AC-ASP alone.

Assessing the conservation impact of Chinese indigenous chicken populations between ex-situ and in-situ using genome-wide SNPs

Conservation programs require rigorous evaluation to ensure the preservation of genetic diversity and viability of conservation populations. In this study, we conducted a comparative analysis of two indigenous Chinese chicken breeds, Gushi and Xichuan black-bone, using whole-genome SNPs to understand their genetic diversity, track changes over time and population structure. The breeds were divided into five conservation populations(GS1, 2010, ex-situ;GS2, 2019, ex-situ;GS3, 2019, in-situ;XB1, 2010, in-situ;and XB2, 2019, in-situ) based on conservation methods and generations. The genetic diversity indices of three conservation populations of Gushi chicken showed consistent trends, with the GS3 population under in-situ strategy having the highest diversity and GS2 under ex-situ strategy having the lowest. The degree of inbreeding of GS2 was higher than that of GS1 and GS3. Conserved populations of Xichuan black-bone chicken showed no obvious changes in genetic diversity between XB1 and XB2. In terms of population structure, the GS3 population were stratified relative to GS1 and GS2. According to the conservation priority, GS3 had the highest contribution to the total gene and allelic diversity in GS breed, whereas the contribution of XB1 and XB2 were similar. We also observed that the genetic diversity of GS2 was lower than GS3, which were from the same generation but under different conservation programs(in-situ and ex-situ). While XB1 and XB2 had similar levels of genetic diversity. Overall, our findings suggested that the conservation programs performed in ex-situ could slow down the occurrence of inbreeding events, but could not entirely prevent the loss of genetic diversity when the conserved population size was small, while in-situ conservation populations with large population size could maintain a relative high level of genetic diversity.

In-situ AFM and quasi-in-situ studies for localized corrosion in Mg-9Al-1Fe-(Gd) alloys under 3.5 wt.% NaCl environment

Revealing the localized corrosion process of Mg alloy is considered as one of the most significant ways for improving its corrosion resistance.The reliable monitor should be high distinguishability and real-time in liquid environment.Herein,Mg-9Al-1Fe and Mg-9Al-1Fe-1Gd alloys were designed to highlight the impact of intermetallic on the corrosion behaviour.In-situ AFM with a special electrolyte circulation system and quasi-in-situ SEM observation were used to monitor the corrosion process of the designed alloys.SEM-EDS and TEM-SAED were applied to identify the intermetallic in the designed alloys,and their volta potentials were measured by SKPFM.According to the real-time and real-space in-situ AFM monitor,the corrosion process consisted of dissolution of anodicα-Mg phase,accumulation of corrosion products around cathodic phase and shedding of some fine cathodic phase.Then,the localized corrosion process of Mg alloy was revealed combined with the results of the monitor of corrosion process and Volta potential difference.

An in-situ study of static recrystallization in Mg using high temperature EBSD

It has been a common method to improve the mechanical properties of metals by manipulating their microstructures via static recrystallization,i.e.,through heat treatment.Therefore,the knowledge of recrystallization and grain growth is critical to the success of the technique.In the present work,by using in-situ high temperature EBSD,the mechanisms that control recrystallization and grain growth of an extruded pure Mg were studied.The experimental results revealed that the grains of priority for dynamic recrystallization exhibit fading competitiveness under static recrystallization.It is also found that grain boundary movement or grain growth is likely to show an inverse energy gradient effect,i.e.,low energy grains tend to swallow or grow into high energy grains,and grain boundaries of close to 30°exhibit superior growth advantage to others.Another finding is that{10-12}tensile twin boundaries are sites of hardly observed for recrystallization,and are finally swallowed by adjacent recrystallized grains.The above findings may give comprehensive insights of static recrystallization and grain growth of Mg,and may guide the design of advanced materials processing in microstructural engineering.

Mongolian Text Recognition Based on CRNN Algorithm

Optical Character Recognition(OCR)algorithm is a technology that converts text images from paper documents into a digital format using electronic devices such as scanners and digital cameras.This process transforms the captured text images into editable and searchable versions using text recognition technology.As advancements in deep learning,Al models have increasingly become pivotal in applications requiring operation on mobile devices without network connectivity,including small underwater devices,high-altitude environments,and license plate recognition systems in front-end cameras.Despite the maturity of general OCR models,there is a notable scarcity of OCR algorithms that are compatible with embedded single-chip microcomputers.These models,capable of functioning autonomously at the front-end without network support,are particularly crucial for remote applications.However,virtually no models for single-chip systems currently support the recognition of the Mongolian language.This study focuses on the development of an OCR system designed for single-chip microcomputers operating without network connectivity.The system is engineered to perform character recognition for Mongolian,English,and Chinese scripts,thereby expanding the utility of front-end single-chip devices.Specifically,the research introduces a novel approach to the recognition of modern Mongolian characters,broadening the scope of OCR system in linguistically diverse contexts.

A continuous and long-term in-situ stress measuring method based on fiber optic. Part I: Theory of inverse differential strain analysis

A method for in-situ stress measurement via fiber optics was proposed. The method utilizes the relationship between rock mass elastic parameters and in-situ stress. The approach offers the advantage of long-term stress measurements with high spatial resolution and frequency, significantly enhancing the ability to measure in-situ stress. The sensing casing, spirally wrapped with fiber optic, is cemented into the formation to establish a formation sensing nerve. Injecting fluid into the casing generates strain disturbance, establishing the relationship between rock mass properties and treatment pressure.Moreover, an optimization algorithm is established to invert the elastic parameters of formation via fiber optic strains. In the first part of this paper series, we established the theoretical basis for the inverse differential strain analysis method for in-situ stress measurement, which was subsequently verified using an analytical model. This paper is the fundamental basis for the inverse differential strain analysis method.

Text-to-SQL文本信息处理技术研究综述

信号与信息处理的需求日益增加,离不开数据处理技术,数据处理需要数据库的支持,然而没有经过训练的使用者会因为不熟悉数据库操作产生诸多问题。文本转结构化查询语言(Text to Structured Query Language,Text-to-SQL)的出现,使用户无需掌握结构化查询语言(Structured Query Language,SQL)也能够熟练操作数据库。介绍Text-to-SQL的研究背景及面临的挑战;介绍Text-to-SQL关键技术、基准数据集、模型演变及最新研究进展,关键技术包括Transformer等主流技术,用于模型训练的基准数据集包括WikiSQL和Spider;介绍Text-to-SQL不同阶段模型的特点,详细阐述Text-to-SQL最新研究成果的工作原理,包括模型构建、解析器设计及数据集生成;总结Text-to-SQL未来的发展方向及研究重点。

YOLOv5ST:A Lightweight and Fast Scene Text Detector

Scene text detection is an important task in computer vision.In this paper,we present YOLOv5 Scene Text(YOLOv5ST),an optimized architecture based on YOLOv5 v6.0 tailored for fast scene text detection.Our primary goal is to enhance inference speed without sacrificing significant detection accuracy,thereby enabling robust performance on resource-constrained devices like drones,closed-circuit television cameras,and other embedded systems.To achieve this,we propose key modifications to the network architecture to lighten the original backbone and improve feature aggregation,including replacing standard convolution with depth-wise convolution,adopting the C2 sequence module in place of C3,employing Spatial Pyramid Pooling Global(SPPG)instead of Spatial Pyramid Pooling Fast(SPPF)and integrating Bi-directional Feature Pyramid Network(BiFPN)into the neck.Experimental results demonstrate a remarkable 26%improvement in inference speed compared to the baseline,with only marginal reductions of 1.6%and 4.2%in mean average precision(mAP)at the intersection over union(IoU)thresholds of 0.5 and 0.5:0.95,respectively.Our work represents a significant advancement in scene text detection,striking a balance between speed and accuracy,making it well-suited for performance-constrained environments.

An in-situ self-etching enabled high-power electrode for aqueous zinc-ion batteries

Sluggish storage kinetics is considered as the main bottleneck of cathode materials for fast-charging aqueous zinc-ion batteries(AZIBs).In this report,we propose a novel in-situ self-etching strategy to unlock the Palm tree-like vanadium oxide/carbon nanofiber membrane(P-VO/C)as a robust freestanding electrode.Comprehensive investigations including the finite element simulation,in-situ X-ray diffraction,and in-situ electrochemical impedance spectroscopy disclosed it an electrochemically induced phase transformation mechanism from VO to layered Zn_(x)V_(2)O_5·nH_(2)O,as well as superior storage kinetics with ultrahigh pseudocapacitive contribution.As demonstrated,such electrode can remain a specific capacity of 285 mA h g^(-1)after 100 cycles at 1 A g^(-1),144.4 mA h g^(-1)after 1500 cycles at 30 A g^(-1),and even 97 mA h g^(-1)after 3000 cycles at 60 A g^(-1),respectively.Unexpectedly,an impressive power density of 78.9 kW kg^(-1)at the super-high current density of 100 A g^(-1)also can be achieved.Such design concept of in-situ self-etching free-standing electrode can provide a brand-new insight into extending the pseudocapacitive storage limit,so as to promote the development of high-power energy storage devices including but not limited to AZIBs.