检索结果-内蒙古大学图书馆

Centroid-based Contrastive Consistency Learning for transferable deepfake detection

NEUrOCOMPUTING 2025年 637卷

作者： Zha, ruiqi Lian, Zhichao Li, Qianmu Nanjing Univ Sci & Technol Sch Cyber Sci & Engn Nanjing 210094 Peoples R China

Previous research efforts in deepfake detection mainly concentrated on identifying and differentiating artifacts discernible to humans. Those methods have left a bias in learned models, as they tend to concentrate on the disparities between forged and natural regions from the perspective of a single sample while overlooking consistency within categories from the perspective of the entire sample set, which remains crucial across various real-world applications. Therefore, inspired by contrastive learning, we tackle the deepfake detection problem by learning the invariant representations of both categories. Our proposed method, termed Centroidbased Contrastive Consistency Learning (C3L) method, integrates constraints on representations at both the data preprocessing and feature extraction stages. Specifically, during data preprocessing, we consider both temporal relationships within videos and the latent relationships within synthesis data. We introduce a novel Positive Enhancement Module (PEM) designed to characterize natural and forged samples in a facial semantically irrelevant way, thereby guiding a task-oriented positive pair contrasting strategy. In addition, at the feature extraction stage, we introduce the Margin Feature Simulation Module (MFSM), which leverages the centroid of the natural category to simulate marginal features for both categories. Subsequently, we employ the Supervised Contrastive Margin Loss (SCML), utilizing simulated features to emphasize differences at decision boundaries and optimize the learning process. The effectiveness and robustness of the proposed method have been demonstrated through extensive experiments.

关键词： deep learning deepfake detection Contrastive learning Outlier detection

来源：评论

学校读者我要写书评

暂无评论

VUV imaging of type-ⅠELM filamentary structures and their temporal characteristics on EAST

引用

Plasma Science and Technology 2024年第11期26卷 22-31页

作者： rongjing dENG Tingfeng MING Bang LI Qiqi SHI Shanwei HOU Shuqi YANG Xiaoju LIU Shaocheng LIU Guoqiang LI Xiang GAO Yasuhiro SUZUKI Yunfeng LIANG Institute of Plasma Physics Hefei Institutes of Physical ScienceChinese Academy of SciencesHefei 230031People’s Republic of China University of Science and Technology of China Hefei 230026People’s Republic of China Key Laboratory of Optoelectronic Devices and Systems of Ministry of Education and Guangdong Province College of Optoelectronic EngineeringShenzhen UniversityShenzhen 518060People’s Republic of China Graduate School of Advanced Science and Engineering Hiroshima UniversityHigashi-Hiroshima 739-8527Japan Forschungszentrum Jülich GmbH Institut für Energie-und Klimaforschung-PlasmaphysikJülich D-52425Germany

In the H-mode experiments conducted on the Experimental Advanced Superconducting Tokamak(EAST),fluctuations induced by the so-called edge localized modes(ELMs)are captured by a high-speed vacuum ultraviolet(VUV)imaging *** field line-aligned filamentary structures are analyzed in this *** transport induced by ELM filaments in the scrape-off layer(SOL)under different discharge conditions is analyzed by comparing the VUV signals with the divertor probe *** is found that convective transport along open field lines towards the divertor target dominates the parallel ion particle transport mechanism during *** toroidal mode number of the filamentary structure derived from the VUV images increases with the electron density pedestal *** analysis of the toroidal distribution characteristics during ELM bursts reveals toroidal *** influence of resonance magnetic perturbation(rMP)on the ELM size is also analyzed using VUV imaging *** the phase difference of the coil changes periodically,the widths of the filaments change as ***,the temporal evolution of the ELMs on the VUV signals provides rise time and decay time for each single ELM event,and the results indicate a negative correlation trend between these two times.

关键词： EAST tokamak VUV imaging filamentary structure temporal characteristics rMP

来源：评论

学校读者我要写书评

暂无评论

Classification-oriented dawid skene model for transferring intelligence from crowds to machines

引用

Frontiers of Computer Science 2023年第5期17卷 53-66页

作者： Jiaran LI richong ZHANG Samuel MENSAH Wenyi QIN Chunming HU Department of Computer Science and Engineering Beihang UniversityBeijing 100191China Department of Computer Science University of SheffieldSheffieldS102TNUK

When a crowdsourcing approach is used to assist the classification of a set of items,the main objective is to classify this set of items by aggregating the worker-provided labels.A secondary objective is to assess the workers’skill levels in this process.A classical model that achieves both objectives is the famous dawid-Skene *** this paper,we consider a third objective in this context,namely,to learn a classifier that is capable of labelling future items without further assistance of crowd *** extending the dawidSkene model to include the item features into consideration,we develop a Classification-Oriented dawid Skene(COdS)model,which achieves the three objectives *** effectiveness of COdS on this three dimensions of the problem space is demonstrated experimentally.

关键词： crowdsourcing information aggregation lear-ning from crowds

来源：评论

学校读者我要写书评

暂无评论

Selective recovery of Th(Ⅳ) from radioactive rare earth waste residue by utilizing MoS_(2)-modified ion-absorbed type rare earth tailings

引用

Journal of rare Earths 2024年第9期42卷 1782-1791,I0005页

作者： Xirong Chen Xinyu Fu Yun Gao Zhiyuan Zeng Bin Huang Xiaoqi Sun Faculty of Materials Metallurgy and Chemistry Jiangxi University of Science and TechnologyGanzhou341000China CAS Key Laboratory of Design and Assembly of Functional Nanostructures Fujian Key Laboratory of NanomaterialsFujian Institute of Research on the Structure of MatterChinese Academy of SciencesFuzhou350002China Fujian Research Center for Rare Earth Engineering Technology Xiamen Institute of Rare Earth MaterialsHaixi InstituteChinese Academy of SciencesXiamen361021China Fujian College University of Chinese Academy of SciencesFuzhou350002China Jiangxi Province Key Laboratory of Cleaner Production of Rare Earths Ganjiang Innovation AcademyChinese Academy of SciencesGanzhouJiangxi341000China

In the process of extracting ion-absorbed rare earth ore(IrEO),the production of radioactive waste is a major environmental *** address this issue,MoS_(2) was used to modify ion-absorbed rare earth tailings(rET) to synthesize a novel MoS_(2)@rET composite material for the effective handling of radioactive waste generated in IrEO separation *** composite material was thoroughly characterized using various analytical techniques,including X-ray diffraction(Xrd),X-ray photoelectron spectroscopy(XPS),thermogravimetry(TG),Fourier-transform infrared(FTIr),scanning electron microscopy(SEM),Brunaue r-Emmett-Teller(BET) and energy dispersive spectroscopy(EdS).By optimizing the operating parameters,the optimal experimental conditions were determined to be pH=3,contact time=60 min,liquid-solid ratio=6 g/L,and initial concentration=150 mg/*** adsorption data fitted well with the pseudo second-order rate *** thermodynamic parameters concerning the adsorption of Th(Ⅳ) were analyzed and *** isotherm model is a more fitting choice for the adsorption process compared to the Freundlich isotherm ***_(2)@rET was used in the acid leachate of IrEO waste residue,achieving the separation of Th and rare earth *** mechanism of Th(Ⅳ) adsorption by MoS_(2)@rET was investigated,revealing that the adsorption process involves electrostatic interactions,chemical bonding,and redox *** above research results indicate that MoS_(2)@rET composite materials have application potential in the sustainable treatment of IrEO radioactive waste.

关键词： rare earths Thorium radioactive waste residue Adsorption Separation

来源：评论

学校读者我要写书评

暂无评论

Construction of a novel six-gene signature to predict tumour response to induction chemotherapy and overall survival in locoregionally advanced laryngeal and hypopharyngeal carcinoma

引用

Genes & diseases 2024年第3期11卷 148-150页

作者： Chen Tan Lingwa Wang Yifan Yang Shizhi He George G.Chen Jason YK.Chan Michael CF.Tong C.A.van Hasselt Wenbin Xu Ling Feng ru Wang Jugao Fang Department of Otolaryngology Head and Neck Surgery Beijing Tongren HospitalCapital Medical UniversityBeijing 100730China Key Laboratory of Otolaryngology Head and Neck Surgery(Ministry of Education of China) Beijing Institute of OtolaryngologyBeijing 100730China Department of Otorhinolaryngology Head and Neck SurgeryThe Chinese University of Hong KongPrince of Wales HospitalHongKong SAR999077China Department of Medical Genetics Institute of Basic Medical ScienceChinese Academy of Medical Science&Peking Union Medical CollageBeijing 100730China

Locoregionally advanced laryngeal and hypopharyngeal cancers(LA-LHCs)are traditionally treated with surgery followed by postoperative radiotherapy,which impairs speech and swallowing functions and reduces the quality of life.^(1,2)The use of induction chemotherapy(IC)as a larynxpreserving approach for LA-LHCs has been verified and refined.^(3)However,the short-term tumor response to IC varies,non-responders usually show poor survival and little benefit.

关键词： chemotherapy laryngeal surgery

来源：评论

学校读者我要写书评

暂无评论

In Situ Study of Surface Morphology Formation Mechanism during Laser Powder Bed Fusion

引用

APPLIEd SCIENCES-BASEL 2025年第5期15卷 2550-2550页

作者： Zhang, Yuhui ren, Hang Yan, Hualin Long, Yu Guangxi Univ State Key Lab Featured Met Mat & Life Cycle Safety Nanning 530004 Peoples R China Guangxi Univ Inst Laser Intelligent Mfg & Precis Proc Sch Mech Engn Nanning 530004 Peoples R China

In the laser powder bed fusion (LPBF) process, the surface quality of intermediate layers impacts interlayer bonding and part forming quality. due to the complex dynamic process inherent in LPBF, current monitoring methods struggle to achieve high-quality in situ online monitoring, which limits the in-depth understanding of the evolution mechanisms of the surface morphology of LPBF intermediate layers. This paper employs an optimized coaxial optical imaging method to monitor key LPBF processes and analyzes the intermediate layer surface morphology evolution mechanism considering heat, force, and mass transfer. results indicate that LPBF intermediate layer surfaces are influenced by energy density, melt pool behavior, and previous layer morphology, forming complex topological structures. At a low energy density, insufficient powder melting causes balling, extended by subsequent melt pools to form a reticulated structure and local large-scale protrusions. Heat accumulation at a high energy density promotes melt pool expansion, reduces melt track overlap, and effectively eliminates defects from previous layers via remelting, with spatter becoming the main defect. Additionally, the melt pool wettability on the part contours captures external powder, forming unique, overhanging contour protrusions. This paper enhances understanding of LPBF intermediate layer surface morphology formation mechanisms and provides a theoretical basis for optimizing surface quality.

关键词： laser powder bed fusion in situ co-axial monitoring melt pool surface topography forming mechanism build quality

来源：评论

学校读者我要写书评

暂无评论

The impact of combined thymol and rosmarinic acid on the intestinal microbiota and barrier function of the piglets challenged by Escherichia coli K88

引用

Animal Nutrition 2025年第1期20卷 131-144页

作者： runlin Li Xuedong ding Mingkang Lei Panpan Li Ilias Giannenas Jing Wang Weiyun Zhu National Center for International Research on Animal Gut Nutrition Jiangsu Key Laboratory of Gastrointestinal Nutrition and Animal HealthLaboratory of Gastrointestinal MicrobiologyNational Experimental Teaching Demonstration Center of Animal ScienceCollege of Animal Science and TechnologyNanjing Agricultural UniversityNanjing 210095China Aristotle University of Thessaloniki University CampusThessaloniki 54124Greece

It has been found that thymol(Thy)and rosmarinic acid(ros-A)improve the growth performance of piglets and relieve intestinal inflammation in *** effects of Thy and ros-A separately or in combination(Thyros-A)on the intestinal function and health of piglets challenged with Escherichia coli K88(*** K88)were investigated.A total of 30 piglets aged 21 d were assigned to 5 groups(n=6).The control(Con)and K88 groups piglets received a basal diet,while the Thy,ros-A,and Thyros-A groups were fed a basal diet supplemented with 500 mg/kg Thy,500 mg/kg ros-A,and 250 mg/kg Thy t 250 mg/kg ros-A,*** the 19th and 20th day,piglets in the K88,Thy,ros-A,and Thyros-A groups were orally administered 10 mL of phosphate-buffered saline(PBS)containing approximately 1109 CFU/mL of *** K88,while the Con group received an equal volume of *** results showed that the Thyros-A treatment reduced the damage to ileal villi induced by the *** K88 challenge,leading to longer villi in the ileum(P<0.05).Thy and ros-A modulated the composition of the ileal *** to the K88 group,the Thyros-A group had a higher abundance of Lactobacillus and romboutsia,while Escherichia-Shigella and desulforvibrio were lower(P<0.05).Additionally,the Thyros-A group showed elevated levels of gene and protein expressions for zonula occludens-1,occludin,and claudin-1 compared to the K88 group(P<0.05).In conclusion,combining Thy and ros-A reduced ileal damage and relieved the inflammation in weaned piglets challenged with *** K88 by regulating intestinal microflora and improving barrier function.

关键词： Thymol rosmarinic acid Intestinal microbiota Weaned piglet

来源：评论

学校读者我要写书评

暂无评论

HdN-ddI: a novel framework for predicting drug-drug interactions using hierarchical molecular graphs and enhanced dual-view representation learning

引用

BMC BIOINFOrMATICS 2025年第1期26卷 1-20页

作者： Sun, Jinchen Zheng, Haoran Univ Sci & Technol China Sch Comp Sci & Technol 443 Huangshan Rd Hefei 230027 Peoples R China Univ Sci & Technol China Anhui Key Lab Software Engn Comp & Commun 443 Huangshan Rd Hefei 230027 Peoples R China

Backgrounddrug-drug interactions (ddIs) especially antagonistic ones present significant risks to patient safety, underscoring the urgent need for reliable prediction methods. recently, substructure-based ddI prediction has garnered much attention due to the dominant influence of functional groups and substructures on drug properties. However, existing approaches face challenges regarding the insufficient interpretability of identified substructures and the isolation of chemical *** study introduces a novel framework for ddI prediction termed HdN-ddI. HdN-ddI integrates an explainable substructure extraction module to decompose drug molecules and represents them using innovative hierarchical molecular graphs, which effectively incorporates information from real chemical substructures and improves molecules encoding efficiency. Furthermore, the enhanced dual-view learning method inspired by the underlying mechanisms of ddIs enables HdN-ddI to comprehensively capture both hierarchical structure and interaction information. Experimental results demonstrate that HdN-ddI has achieved state-of-the-art performance with accuracies of 97.90% and 99.38% on the two widely-used datasets in the warm-start setting. Moreover, HdN-ddI exhibits substantial improvements in the cold-start setting with boosts of 4.96% in accuracy and 7.08% in F1 score on previously unseen drugs. real-world applications further highlight HdN-ddI's robust generalization capabilities towards newly approved *** its accurate predictions and robust generalization across different settings, HdN-ddI shows promise for enhancing drug safety and efficacy. Future research will focus on refining decomposition rules as well as integrating external knowledge while preserving the model's generalization capabilities.

关键词： drug-drug interactions Chemical substructures Hierarchical molecular graphs Graph attention network dual-view representation learning

来源：评论

学校读者我要写书评

暂无评论

Helicity-Controlled Spin Hall Angle in 2d Altermagnets with rashba Spin-Orbit Coupling

引用

Chinese Physics Letters 2025年第1期42卷 134-140页

作者： Weiwei Chen Longhai Zeng W.Zhu Key Laboratory of Intelligent Manufacturing Quality Big Data Tracing and Analysis of Zhejiang Province College of SciencesChina Jiliang UniversityHangzhou 310018China Key Laboratory of Intelligent Manufacturing Quality Big Data Tracing and Analysis of Zhejiang Province College of ScienceChina Jiliang UniversityHangzhou 310018China School of Science Westlake UniversityHangzhou 310024China

We investigated the efficiency of charge-to-spin conversion in two-dimensional rashba altermagnets,a class of materials that combines the characteristics of both ferromagnets and *** quantum linear response theory,we quantified the longitudinal and spin Hall conductivities in this system and demonstrated a substantial enhancement in the spin Hall angle below the band crossing point through the dual effects of relativistic spin–orbit interaction and nonrelativistic altermagnetic exchange ***,the results showed that the skew scattering and intrinsic mechanisms arising from Fermi sea states are almost negligible in this system,in contrast to conventional ferromagnetic rashba *** findings not only elucidate the spin dynamics in rashba altermagnets but also pave the way for developing novel strategies for manipulating charge-to-spin conversion via sophisticated control of noncollinear and collinear out-of-plane spin textures.

关键词： interaction magnets rashba

来源：评论

学校读者我要写书评

暂无评论

Unraveling the dynamics of atmospheric methane: the impact of anthropogenic and natural emissions

引用

ENVIrONMENTAL rESEArCH LETTErS 2024年第6期19卷 064001页

作者： Fu, Bo Jiang, Yongye Chen, Guolong Lu, Mengmeng Lai, Yuqin Suo, Xinhao Li, Bengang Peking Univ Inst Carbon Neutral Coll Urban & Environm Sci Beijing 100871 Peoples R China

The reduction in methane concentration is crucial for achieving the goals of the Paris agreement. However, its annual growth rate is unstable, and understanding the reasons for changes in methane growth is essential for climate policy-making. Currently, there is considerable uncertainty regarding its attribution. Here, we utilize multi-source data and optimal fingerprinting methods to detect the contributions of several key drivers to the methane trend and interannual variability. We find that the methane growth trend is primarily influenced by anthropogenic emissions, while interannual variability is predominantly determined by wetland and biomass burning emissions. This result underscores the central role of anthropogenic emissions in methane dynamics, providing confidence in the effectiveness of human efforts to control methane atmospheric concentrations through emission reductions. It also helps alleviate concerns about the recent surge in atmospheric methane concentration, as it may be a short-term peak caused by increased wetland emissions rather than a long-term change.

关键词： methane climate change anthropogenic wetland

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：