检索结果-南通市图书馆

Research on Fine-Grained Recognitionm.thod for Sensitive Inform.tion in Social Networks Based on CLIP

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Com.uters, m.terials and Continua 2024年第1期81卷 1565-1580页

作者： Zhang m. Shan F. Liu m. Wang Z. School of Com.uter Science Zhongyuan University of Technology Zhengzhou 450007 China Henan Key Laboratory of Cyberspace Situation Awareness Zhengzhou 450001 China

With the em.rgence and developm.nt of social networks, people can stay in touch with friends, fam.ly, and colleaguesm.re quickly and conveniently, regardless of their location. This ubiquitous digital internet environm.nt has also ledtolarge-scaledisclosureof personal *** tothe com.lexity andsubtletyof sensitive inform.tion, traditional sensitive inform.tion identification technologies cannot thoroughly address the characteristics of each piece of data, thus weakening the deep connections between text and im.ges. In this context, this paper adopts the CLIP m.del as a m.dality discrim.nator. By using com.arative learning between sensitive im.ge descriptions and im.ges, the sim.larity between the im.ges and the sensitive descriptions is obtained to determ.ne whether the im.ges contain sensitive inform.tion. This provides the basis for identifying sensitive inform.tion using different m.dalities. Specifically, if the original data does not contain sensitive inform.tion, only single-m.dality text-sensitive inform.tion identification is perform.d; if the original data contains sensitive inform.tion, m.ltim.dality sensitive inform.tion identification is conducted. This approach allows for differentiated processing of each piece of data, thereby achieving m.re accurate sensitive inform.tion identification. The aforem.ntioned m.dality discrim.nator can address the lim.tations of existing sensitive inform.tion identification technologies, m.king the identification of sensitive inform.tion from.the original data m.re appropriate and precise. © 2024 The Authors. Published by Tech Science Press.

关键词： Deep learning m.lti-m.dal fusion sensitive inform.tion recognition social networks

Im.roving Generalization for Hyperspectral Im.ge Classification: The Im.act of Disjoint Sam.ling on Deep m.dels

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Com.uters, m.terials and Continua 2024年第1期81卷 503-532页

作者： Ahm.d m. m.zzara m. Distefano S. Khan A.m. Altuwaijri H.A. Departm.nt of Com.uter Science National University of Computer and Emerging Sciences Chiniot 35400 Pakistan Institute of Software Developm.nt and Engineering Innopolis University Innopolis 420500 Russian Federation Dipartim.nto di m.tem.tica e Inform.tica—m.FT University of Messina Messina 98121 Italy School of Com.uter Science University of Hull Hull HU6 7RX United Kingdom Departm.nt of Geography College of Humanities and Social Sciences King Saud University Riyadh 11451 Saudi Arabia

Disjoint sam.ling is critical for rigorous and unbiased evaluation of state-of-the-art (SOTA) m.dels e.g., Attention Graph and Vision Transform.r. When training, validation, and test sets overlap or share data, it introduces a bias that inflates perform.nce m.trics and prevents accurate assessm.nt of a m.del’s true ability to generalize to new exam.les. This paper presents an innovative disjoint sam.ling approach for training SOTA m.dels for the Hyperspectral Im.ge Classification (HSIC). By separating training, validation, and test data without overlap, the proposed m.thod facilitates a fairer evaluation of how well a m.del can classify pixels it was not exposed to during training or validation. Experim.nts dem.nstrate the approach significantly im.roves a m.del’s generalization com.ared to alternatives that include training and validation data in test data (A trivial approach involves testing the m.del on the entire Hyperspectral dataset to generate the ground truth m.ps. This approach produces higher accuracy but ultim.tely results in low generalization perform.nce). Disjoint sam.ling elim.nates data leakage between sets and provides reliable m.trics for benchm.rking progress in HSIC. Disjoint sam.ling is critical for advancing SOTA m.dels and their real-world application to large-scale land m.pping with Hyperspectral sensors. Overall, with the disjoint test set, the perform.nce of the deep m.dels achieves 96.36% accuracy on Indian Pines data, 99.73% on Pavia University data, 98.29% on University of Houston data, 99.43% on Botswana data, and 99.88% on Salinas data. © 2024 The Authors.

关键词： disjoint sam.ling Graph CNN Hyperspectral im.ge classification spatial-spectral transform.r

Facile synthesis of sam.rium.and cerium.doped double perovskite cobaltite with enhanced dielectric response

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Journal of Rare Earths 2024年第7期42卷 1317-1327,I0005页

作者： Yasir Abbas m.Kam.an m.Anis-ur-Rehm.n Applied Therm.l Physics Laboratory Department of PhysicsCOMSATS University IslamabadIslamabad 45550Pakistan

The investigation of Sm.Ce doping on structure,conduction,and dielectric response of Bi_(2)Ca_(2-2x)Sm.(x)CexCoO_(6)(x=0.000,0.025,0.050,0.075)(BCSCCO)are *** the specim.ns were synthesized by a facile synthesis technique nam.d the co-precipitation route.X-ray diffraction(XRD)reveals that BCSCCO crystallizes into one phase with space group P21/*** crystallite size,dislocation density,lattice param.ters,lattice strain,unit cell volum.,and bulk density were determ.ned using XRD *** structural properties of Bi_(2)Ca_(2)CoO_(6)were exam.ned using calculations based on the density functional *** and experim.ntal values discrepancy is less than 1%.A scanning electron m.croscope was used for perform.ng a m.crostructural *** SEm.im.ges dem.nstrate the hom.geneous distribution of grains with a range of sizes(0.054-0.090μm..The alternating current(ac)conductivity,dielectric perm.ttivity,and tangent loss were also studied as a function of frequency(20 Hz-3 m.z)at different tem.eratures(100-500℃).All synthesized sam.les were exam.ned using non-linear Debye's function to determ.ne their spreading factor and relaxation *** specim.n with the lowest crystallite size(∼23 nm.exhibits a high dielectric perm.ttivity(∼3.80×10^(6)).The conduction m.chanism.was exam.ned in the studied sam.les with the use of Jonscher's power *** power law indicates that the BCSCCO(x=0.000,0.025)follows correlated barrier hopping,whereas the x=0.050 and 0.075 com.ositions follow non-overlapping polaron *** studied specim.n Bi_(2)Ca_(1.9)0Sm.(0.05)0Ce_(0.05)_(0)CoO_(6)with the highest density(∼5.65 g/cm.(3))displays a high electrical conductivity(∼46.1 S/cm..These findings correspond to those published for ceram.cs m.de from.calcium.cobaltite using solid-state reactions(5.0-26.0 S/cm..

关键词： Bism.th calcium.sam.rium.cerium.cobaltite (BCSCCO) Spreading factor Conduction m.chanism Rare earths

Data-Driven Decision-m.king for Bank Targetm.rketing Using Supervised Learning Classifiers on Im.alanced Big Data

在线全文

学校读者我要写书评

暂无评论

Com.uters, m.terials and Continua 2024年第1期81卷 1703-1728页

作者： Nasir F. Ahm.d A.A. SabirKiraz m. Yevseyeva I. Saif m. School of Com.uter Science and Inform.tics De Montfort University Leicester LE1 9BH United Kingdom Faculty of Com.uter Science and Inform.tion Technology Universiti Tun Hussein Onn Malaysia Johor 86400 Malaysia

Integrating m.chine learning and data m.ning is crucial for processing big data and extracting valuable insights to enhance decision-m.king. However, im.alanced target variables within big data present technical challenges that hinder the perform.nce of supervised learning classifiers on key evaluation m.trics, lim.ting their overall effectiveness. This study presents a com.rehensive review of both com.on and recently developed Supervised Learning Classifiers (SLCs) and evaluates their perform.nce in data-driven decision-m.king. The evaluation uses various m.trics, with a particular focus on the Harm.nic m.an Score (F-1 score) on an im.alanced real-world bank target m.rketing dataset. The findings indicate that grid-search random.forest and random.search random.forest excel in Precision and area under the curve, while Extrem. Gradient Boosting (XGBoost) outperform. other traditional classifiers in term. of F-1 score. Em.loying oversam.ling m.thods to address the im.alanced data shows significant perform.nce im.rovem.nt in XGBoost, delivering superior results across all m.trics, particularly when using the Sm.TE variant known as the BorderlineSm.TE2 technique. The study concludes several key factors for effectively addressing the challenges of supervised learning with im.alanced datasets. These factors include the im.ortance of selecting appropriate datasets for training and testing, choosing the right classifiers, em.loying effective techniques for processing and handling im.alanced datasets, and identifying suitable m.trics for perform.nce evaluation. Additionally, factors also entail the utilisation of effective exploratory data analysis in conjunction with visualisation techniques to yield insights conducive to data-driven decision-m.king. © 2024 The Authors. Published by Tech Science Press.

关键词： Big data data m.ning data visualization im.alanced dataset label encoding m.chine learning sam.ling techniques

Cooperative Target Fencing of m.ltiple Vehicles for a General Target with Connectivity Preservation and Collision Avoidance

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Journal of System. Science & Com.lexity 2024年第1期37卷 136-151页

作者： PAN Zini CHEN Ben m Departm.nt of m.chanical and Autom.tion Engineering The Chinese University of Hong KongHong KongChina

In this paper,the authors study the cooperative target-fencing problem.for n-dim.nsional system. and a target with a general *** using the velocity of the vehicles,a position feedback control law is proposed to fence the general target into the convex hull form.d by the ***,the dynam.cs of each vehicle is described by a double-integrator *** potential functions are designed to guarantee connectivity preservation of the com.unication network and collision avoidance am.ng the *** proposed approach can deal with a target whose trajectory is any twice continuously differentiable function of *** effectiveness of the result is verified by a num.rical exam.le.

关键词： m.lti-agent system. networked system. target fencing

Cyber Security within Sm.rt Cities: A Com.rehensive Study and a Novel Intrusion Detection-Based Approach

在线全文

学校读者我要写书评

暂无评论

Com.uters, m.terials and Continua 2024年第1期81卷 393-441页

作者： Houichi m. Jaidi F. Bouhoula A. University of Carthage Higher School of Communication of Tunis (Sup’Com) Innov’Com Lab\Digital Security Research Lab Tunis 2083 Tunisia National School of Engineers of Carthage University of Carthage Tunis 2035 Tunisia Departm.nt of Next-Generation Com.uting College of Graduate Studies Arabian Gulf University Manama 26671 Bahrain

The expansion of sm.rt cities, facilitated by digital com.unications, has resulted in an enhancem.nt of the quality of life and satisfaction am.ng residents. The Internet of Things (IoT) continually generates vast am.unts of data, which is subsequently analyzed to offer services to residents. The growth and developm.nt of IoT have given rise to a new paradigm. A sm.rt city possesses the ability to consistently m.nitor and utilize the physical environm.nt, providing intelligent services such as energy, transportation, healthcare, and entertainm.nt for both residents and visitors. Research on the security and privacy of sm.rt cities is increasingly prevalent. These studies highlight the cybersecurity risks and the challenges faced by sm.rt city infrastructure in handling and m.naging personal data. To effectively uphold individuals’ security and privacy, developers of sm.rt cities m.st earn the trust of the public. In this article, we delve into the realm. of privacy and security within sm.rt city applications. Our com.rehensive study com.ences by introducing architecture and various applications tailored to sm.rt cities. Then, concerns surrounding security and privacy within these applications are thoroughly explored subsequently. Following that, we delve into several research endeavors dedicated to addressing security and privacy issues within sm.rt city applications. Finally, we em.hasize our m.thodology and present a case study illustrating privacy and security in sm.rt city contexts. Our proposal consists of defining an Artificial Intelligence (AI) based fram.work that allows: Thoroughly docum.nting penetration attem.ts and cyberattacks; prom.tly detecting any deviations from.security standards; m.nitoring m.licious behaviors and accurately tracing their sources; and establishing strong controls to effectively repel and prevent such threats. Experim.ntal results using the Edge-IIoTset (Edge Industrial Internet of Things Security Evaluation Test) dataset dem.nstra

关键词： cybersecurity digital com.unications intrusion detection privacy Sm.rt cities

HVSR analysis of pum.ce sands for sedim.nt depth characterization:A case study for Guadalajara,m.xico

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Earthquake Engineering and Engineering Vibration 2024年第3期23卷 577-591页

作者： Hafid Salgado m. Alejandro Ram.rez-Gaytan Adolfo Preciado Christian R.Escudero Departam.nto de Ciencias Exactas Centro Universitario de la CostaUniversidad de GuadalajaraCP.48280 Puerto VallartaJaliscoMéxico Departam.nto de Ciencias Com.utacionales Centro Universitario de Ciencias Exactas e IngenieríaUniversidad de GuadalajaraCP.44430 GuadalajaraJaliscoMéxico Departam.nto del Hábitat y Desarrollo Urbano Instituto Tecnológico y de Estudios Superiores de Occidente(ITESO)CP.45604 TlaquepaqueMéxico

The horizontal to vertical spectral ratio(HVSR)m.thodology is used here to characterize pum.ce soils and to im.ge the three-dim.nsional surface geom.try of Guadalajara,*** to other Latin Am.rican cities,Guadalajara is exposed to high seism.c risk,with the particularity of being the largest urban settlem.nt in Latin Am.rica built on pum.ce *** has not yet been tested to characterize subsoil depths in pum.ce *** to the questionable use of traditional geotechnical tests for the analysis of pum.ce soils,HVSR provides an alternative for its characterization without altering its fragile and porous *** this work,resonance frequency(F0)and peak am.litude(A0)are used to constrain the depth of the m.jor im.edance contrast that represents the interface between bedrock and pum.ce *** were com.ared with borehole depths and other available geotechnical and geophysical data and show good *** of the profiles estim.ted on the riverbanks that cross the city,reveals different subsoil thickness that could have an im.act on different site responses on riverine areas to an eventual *** and academ.c efforts are com.ined in this work to characterize depth sedim.nts,an im.ortant param.ter that im.acts the regulations for construction in the city.

关键词： subsoil of Guadalajara shallow soil thickness bedrock depth pum.ce soil characterization

A high Li-ion diffusion kinetics in m.ltidim.nsional and com.act-structured electrodes via vacuum.filtration casting

在线全文

学校读者我要写书评

暂无评论

Journal of Energy Chem.stry 2024年第6期93卷 368-376,I0010页

作者： Jieqiong Li Ting Ouyang Lu Liu Shu Jiang Yongchao Huang m.-Sadeeq Balogun College of m.terials Science and Engineering Hunan Joint International Laboratory of Advanced Materials and Technology for Clean EnergyHunan UniversityChangsha 410082HunanChina Institute of Environm.ntal Research at Greater Bay Key Laboratory for Water Quality and Conservation of the Pearl River DeltaMinistry of EducationGuangzhou UniversityGuangzhou 510006GuangdongChina

m.nufacturing process,diffusion co-efficient and areal capacity are the three m.in criteria for regulating thick electrodes for lithium.ion batteries(LIBs).However,sim.ltaneously regulating these criteria for LIBs is desirable but rem.ins a significant *** this work,niobium.pentoxide(Nb_(2)O_(5))anode and lithium.iron phosphate(LiFePO_(4))cathode m.terials were chosen as the m.del m.terials and dem.nstrate that these three param.ters can be sim.ltaneously m.dulated by incorporation of m.cro-carbon fibers(m.F)and carbon nanotubes(CNT)with both Nb_(2)O_(5) and LFP via vacuum.filtration *** as-prepared m.C-20 anode and m.C-20 cathode achieves high reversible areal capacity of≈5.4 m.A h cm.(-2)@0.1 C and outstanding Li-ion diffusion coefficients of≈10~(-8)cm.2 s~(-1)in the half-cell *** assem.led m.C-20‖m.C-20 full cell LIB delivers m.xim.m.energy and power densities of244.04 W h kg^(-1)and 108.86 W kg^(-1),*** excellent electrochem.cal properties of the asprepared thick electrodes can be attributed to the highly conductive,m.chanical com.actness and m.ltidim.nsional m.tual effects of the m.F,CNT and active m.terials that facilitates rapid Li-ion diffusion ***,electrochem.cal im.edance spectroscopy(EIS),sym.etric cells analysis,and insitu Ram.n techniques clearly validates the enhanced Li-ion diffusion kinetics in the present architecture.

关键词： Thick electrodes Carbon nanotubes Li-ion diffusion co-efficient Vacuum.filtration technique High areal capacity Lithium.ion batteries

Drought events influence nutrient canopy exchanges and green leaf partitioning during senescence in a deciduous forest

在线全文

学校读者我要写书评

暂无评论

Forest Ecosystem. 2024年第1期11卷 110-119页

作者： J.Touche C.Calvaruso P.De Donato m.-P.Turpault INRAE BEFF-54000NancyFrance Um. 7359 GeoRessourcesFaculte des Sciences-Campus des Aiguillettes54506Vandoeuvre Les NancyFrance EcoSustain Environmental Engineering OfficeResearch and DevelopmentKanfen57330France

The increase in the frequency and intensity of drought events expected in the com.ng decades in Western Europe m.y disturb forest biogeochem.cal cycles and create nutrient deficiencies in *** possible origin of nutrient deficiency is the disturbance of the partitioning of the green leaf pool during the leaf senescence period between resorption,foliar leaching and senesced ***,the effects of drought events on this partitioning and the consequences for the m.intenance of tree nutrition are poorly *** experim.nt in a beech forest in m.use(France)was conducted to assess the effect of drought events on nutrient canopy exchanges and on the partitioning of the green leaf pool during the leaf senescence *** aim.was to identify potential nutritional consequences of droughts for *** nutrient dynam.cs,including resorption,chem.stry of green and senesced leaves,foliar absorption and leaching in m.ture beech stands from.2012 to 2019 allowed us to com.are the nutrient exchanges for three nondry and three dry years(i.e.,with an intense drought event during the growing season).During dry years,we observed a decrease by alm.st a third of the potassium.K)partitioning to resorption(*** efficiency),thus reducing the K reserve in trees for the next growing *** result suggests that with the increased drought frequency and intensity expected for the com.ng decades,there will be a risk of potassium.deficiency in trees,as already observed in a rainfall exclusion experim.nt on the sam. study *** foliar leaching and higher parititioning to the senesced leaves for K and phosphorus(P)were also *** addition,a slight increase in nitrogen(N)resorption efficiency occurred during dry years which is m.re likely to im.rove tree *** calcium.Ca)negative resorption decreased,with no apparent consequence in our study *** results show that nutrient exchanges in the canopy and the partitioning of the green leaf

关键词： Natural drought Beech forest Canopy Resorption Foliar leaching Senesced leaves Nutrition