In the age of big data,ensuring data privacy while enabling efficient encrypted data retrieval has become a critical challenge.Traditional searchable encryption schemes face difficulties in handling complex semantic q...In the age of big data,ensuring data privacy while enabling efficient encrypted data retrieval has become a critical challenge.Traditional searchable encryption schemes face difficulties in handling complex semantic queries.Additionally,they typically rely on honest but curious cloud servers,which introduces the risk of repudiation.Furthermore,the combined operations of search and verification increase system load,thereby reducing performance.Traditional verification mechanisms,which rely on complex hash constructions,suffer from low verification efficiency.To address these challenges,this paper proposes a blockchain-based contextual semantic-aware ciphertext retrieval scheme with efficient verification.Building on existing single and multi-keyword search methods,the scheme uses vector models to semantically train the dataset,enabling it to retain semantic information and achieve context-aware encrypted retrieval,significantly improving search accuracy.Additionally,a blockchain-based updatable master-slave chain storage model is designed,where the master chain stores encrypted keyword indexes and the slave chain stores verification information generated by zero-knowledge proofs,thus balancing system load while improving search and verification efficiency.Finally,an improved non-interactive zero-knowledge proof mechanism is introduced,reducing the computational complexity of verification and ensuring efficient validation of search results.Experimental results demonstrate that the proposed scheme offers stronger security,balanced overhead,and higher search verification efficiency.展开更多
High-resolution remote sensing images(HRSIs)are now an essential data source for gathering surface information due to advancements in remote sensing data capture technologies.However,their significant scale changes an...High-resolution remote sensing images(HRSIs)are now an essential data source for gathering surface information due to advancements in remote sensing data capture technologies.However,their significant scale changes and wealth of spatial details pose challenges for semantic segmentation.While convolutional neural networks(CNNs)excel at capturing local features,they are limited in modeling long-range dependencies.Conversely,transformers utilize multihead self-attention to integrate global context effectively,but this approach often incurs a high computational cost.This paper proposes a global-local multiscale context network(GLMCNet)to extract both global and local multiscale contextual information from HRSIs.A detail-enhanced filtering module(DEFM)is proposed at the end of the encoder to refine the encoder outputs further,thereby enhancing the key details extracted by the encoder and effectively suppressing redundant information.In addition,a global-local multiscale transformer block(GLMTB)is proposed in the decoding stage to enable the modeling of rich multiscale global and local information.We also design a stair fusion mechanism to transmit deep semantic information from deep to shallow layers progressively.Finally,we propose the semantic awareness enhancement module(SAEM),which further enhances the representation of multiscale semantic features through spatial attention and covariance channel attention.Extensive ablation analyses and comparative experiments were conducted to evaluate the performance of the proposed method.Specifically,our method achieved a mean Intersection over Union(mIoU)of 86.89%on the ISPRS Potsdam dataset and 84.34%on the ISPRS Vaihingen dataset,outperforming existing models such as ABCNet and BANet.展开更多
The rapid development of information technology and accelerated digitalization have led to an explosive growth of data across various fields.As a key technology for knowledge representation and sharing,knowledge graph...The rapid development of information technology and accelerated digitalization have led to an explosive growth of data across various fields.As a key technology for knowledge representation and sharing,knowledge graphs play a crucial role by constructing structured networks of relationships among entities.However,data sparsity and numerous unexplored implicit relations result in the widespread incompleteness of knowledge graphs.In static knowledge graph completion,most existing methods rely on linear operations or simple interaction mechanisms for triple encoding,making it difficult to fully capture the deep semantic associations between entities and relations.Moreover,many methods focus only on the local information of individual triples,ignoring the rich semantic dependencies embedded in the neighboring nodes of entities within the graph structure,which leads to incomplete embedding representations.To address these challenges,we propose Two-Stage Mixer Embedding(TSMixerE),a static knowledge graph completion method based on entity context.In the unit semantic extraction stage,TSMixerE leveragesmulti-scale circular convolution to capture local features atmultiple granularities,enhancing the flexibility and robustness of feature interactions.A channel attention mechanism amplifies key channel responses to suppress noise and irrelevant information,thereby improving the discriminative power and semantic depth of feature representations.For contextual information fusion,a multi-layer self-attentionmechanism enables deep interactions among contextual cues,effectively integrating local details with global context.Simultaneously,type embeddings clarify the semantic identities and roles of each component,enhancing the model’s sensitivity and fusion capabilities for diverse information sources.Furthermore,TSMixerE constructs contextual unit sequences for entities,fully exploring neighborhood information within the graph structure to model complex semantic dependencies,thus improving the completeness and generalization of embedding representations.展开更多
Perception of air pollution is subjective and context-dependent.Previous studies exploring the association between measured air pollution and perceived air quality mainly focused on air pollution levels measured in th...Perception of air pollution is subjective and context-dependent.Previous studies exploring the association between measured air pollution and perceived air quality mainly focused on air pollution levels measured in the residence-based(RB)or regional context,overlooking the mobility-based(MB)context in which people are exposed to air pollution.This study measures air pollution levels in MB,RB,and regional contexts and examines their relationships with perceived air quality across different neighborhoods and gender sub-groups of Hong Kong,China to investigate how people perceive air quality.The results indicate that particulate matter 2.5(PM_(2.5))measured in RB and the regional context significantly contributes to people’s perceived air quality compared to MB PM_(2.5).Individuals in Central and Western district of Hong Kong rely on RB,regional and MB PM_(2.5) to assess air pollution.In Sham Shui Po,RB PM_(2.5) exhibits the highest influence on people’s perceived air quality,followed by regional PM_(2.5).Women’s perceived air quality is strongly related to their RB PM_(2.5) exposure,while men’s perceived air quality is associated with both RB PM_(2.5) and regional PM_(2.5) levels.We conclude that neighborhood effects and mobility levels are the two most important factors influencing the association between meas-ured air pollution and perceived air quality.We reveal that the neighborhood effect averaging problem(NEAP)influences the associ-ation between perceived air quality and measured air pollution levels in a way that differs from health outcome-related studies.Effect-ive measures are needed to improve the public’s awareness of air pollution,and scientific control should be implemented to reduce pub-lic exposure.展开更多
Based on the contextual adaptation perspective of Verschueren’s Adaptation Theory,this paper explores the Chinese translation strategies of Japanese quotation sentences in the Yang translation of The Courage of One f...Based on the contextual adaptation perspective of Verschueren’s Adaptation Theory,this paper explores the Chinese translation strategies of Japanese quotation sentences in the Yang translation of The Courage of One from the perspectives of communicative context and linguistic context.The study finds that the Chinese translation of Japanese quotation sentences involves various strategies,including retaining direct quotations,converting direct quotations into statements,transforming direct quotations into attributive+noun forms,and alternating between direct and indirect quotations.This research provides a new perspective for the Chinese translation of Japanese quotation sentences and offers theoretical support for translation practices in cross-cultural communication.展开更多
The current study investigated how language context and word frequency influenced vowel perception of Chinese-Japanese cognates among Chinese learners of Japanese.Focusing on orthographic cognates,participants perform...The current study investigated how language context and word frequency influenced vowel perception of Chinese-Japanese cognates among Chinese learners of Japanese.Focusing on orthographic cognates,participants performed a vowel detection task on cognates,manipulating language context and target language(Chinese vs.Japanese),as well as word frequency(high vs.low).We measured reaction times,perceptual sensitivity,and response criterion.For high-frequency words,consistent language contexts facilitated faster vowel detection in both languages.However,in low-frequency conditions,participants showed higher perceptual sensitivity to Chinese targets and more conservative response criteria for Japanese targets,regardless of context.These findings revealed the complex interplay between word frequency,language dominance,and context in cross-language processing.Our study contributed to the understanding of vowel perception in languages with shared orthography but distinct phonological systems,offering insights for models of cross-language cognition and second language education.Furthermore,it highlighted the importance of considering both word frequency and language-specific features in cross-language studies.展开更多
In recent years,the rapid integration of artificial intelligence(AI)with various industries has led to an intelligent transformation in people’s learning and working patterns.In the field of higher vocational public ...In recent years,the rapid integration of artificial intelligence(AI)with various industries has led to an intelligent transformation in people’s learning and working patterns.In the field of higher vocational public music course teaching,AI provides intelligent teaching tools and learning platforms,while offering students timely and scientific support and companionship,enabling them to complete learning tasks more efficiently.How to explore the transformation path of teaching modes for higher vocational public music courses in the AI context has become a key consideration for frontline teachers.Based on this,this paper first analyzes the significance of teaching higher vocational public music courses in the AI context,and then proposes feasible transformation paths for teaching modes in combination with course characteristics for reference.展开更多
Background Three-dimensional(3D)shape representation using mesh data is essential in various applications,such as virtual reality and simulation technologies.Current methods for extracting features from mesh edges or ...Background Three-dimensional(3D)shape representation using mesh data is essential in various applications,such as virtual reality and simulation technologies.Current methods for extracting features from mesh edges or faces struggle with complex 3D models because edge-based approaches miss global contexts and face-based methods overlook variations in adjacent areas,which affects the overall precision.To address these issues,we propose the Feature Discrimination and Context Propagation Network(FDCPNet),which is a novel approach that synergistically integrates local and global features in mesh datasets.Methods FDCPNet is composed of two modules:(1)the Feature Discrimination Module,which employs an attention mechanism to enhance the identification of key local features,and(2)the Context Propagation Module,which enriches key local features by integrating global contextual information,thereby facilitating a more detailed and comprehensive representation of crucial areas within the mesh model.Results Experiments on popular datasets validated the effectiveness of FDCPNet,showing an improvement in the classification accuracy over the baseline MeshNet.Furthermore,even with reduced mesh face numbers and limited training data,FDCPNet achieved promising results,demonstrating its robustness in scenarios of variable complexity.展开更多
Conventional model transfer techniques,requiring the labelled source data,are not applicable in the privacy-protected medical fields.For the challenging scenarios,recent source data-free domain adaptation(SFDA)has bec...Conventional model transfer techniques,requiring the labelled source data,are not applicable in the privacy-protected medical fields.For the challenging scenarios,recent source data-free domain adaptation(SFDA)has become a mainstream solution but losing focus on the inter-sample class information.This paper proposes a new Credible Local Context Representation approach for SFDA.Our main idea is to exploit the credible local context for more discriminative representation.Specifically,we enhance the source model's discrimination by information regulating.To capture the context,a discovery method is developed that performs fixed steps walking in deep space and takes the credible features in this path as the context.In the epoch-wise adaptation,deep clustering-like training is conducted with two major updates.First,the context for all target data is constructed and then the context-fused pseudo-labels providing semantic guidance are generated.Second,for each target data,a weighting fusion on its context forms the anchored neighbourhood structure;thus,the deep clustering is switched from individual-based to coarse-grained.Also,a new regularisation building is developed on the anchored neighbourhood to drive the deep coarse-grained learning.Experiments on three benchmarks indicate that the proposed method can achieve stateof-the-art results.展开更多
Digital rural governance is a micro-level governance practice within the broader framework of building a Digital China.It involves the integration of digital technology into rural governance to drive the digital trans...Digital rural governance is a micro-level governance practice within the broader framework of building a Digital China.It involves the integration of digital technology into rural governance to drive the digital transformation of rural governance.In recognition of the varied development stages of digital rural governance,the concept of“digital context”provides an analytical lens for exploring the differences in practical models of digital rural governance.By examining the contextual characteristics and differential mechanisms of digital rural governance,this paper delves into its social foundation and technological adaptation.The research finds that the context of digital rural governance primarily encompasses three dimensions:contextual foundation,contextual logic,and contextual optimization.First,the contextual foundation of digital rural governance manifests as the social basis,comprising the social structure of villages,the type of village development,and the age structure of villagers,which constitute the social stratification forms underlying digital rural governance.Second,the contextual logic of digital rural governance focuses on the adaptation of digital technology to rural governance,promoting the adaptation of digital technology to the rural governance foundation,village governance scenarios,and villagers’digital capabilities.Third,the contextual optimization of digital rural governance emphasizes integrating digital technology with both administrative and livelihood-oriented governance affairs at the village level.This approach leverages the governance value and functional potential of digital technology to streamline digital governance processes and enhance digital governance capabilities.As a developmental direction for the transformation of rural governance,digital rural governance must not only highlight the governance advantages of digital technology but also prioritize the inherent context of rural governance.It aims to enhance the effectiveness of rural governance through digital technology and advance high-quality digital village development tailored to local conditions.展开更多
Ship detection in synthetic aperture radar(SAR)image is crucial for marine surveillance and navigation.The application of detection network based on deep learning has achieved a promising result in SAR ship detection....Ship detection in synthetic aperture radar(SAR)image is crucial for marine surveillance and navigation.The application of detection network based on deep learning has achieved a promising result in SAR ship detection.However,the existing networks encounters challenges due to the complex backgrounds,diverse scales and irregular distribution of ship targets.To address these issues,this article proposes a detection algorithm that integrates global context of the images(GCF-Net).First,we construct a global feature extraction module in the backbone network of GCF-Net,which encodes features along different spatial directions.Then,we incorporate bi-directional feature pyramid network(BiFPN)in the neck network to fuse the multi-scale features selectively.Finally,we design a convolution and transformer mixed(CTM)detection head to obtain contextual information of targets and concentrate network attention on the most informative regions of the images.Experimental results demonstrate that the proposed method achieves more accurate detection of ship targets in SAR images.展开更多
Segmentation of the retinal vessels in the fundus is crucial for diagnosing ocular diseases.Retinal vessel images often suffer from category imbalance and large scale variations.This ultimately results in incomplete v...Segmentation of the retinal vessels in the fundus is crucial for diagnosing ocular diseases.Retinal vessel images often suffer from category imbalance and large scale variations.This ultimately results in incomplete vessel segmentation and poor continuity.In this study,we propose CT-MFENet to address the aforementioned issues.First,the use of context transformer(CT)allows for the integration of contextual feature information,which helps establish the connection between pixels and solve the problem of incomplete vessel continuity.Second,multi-scale dense residual networks are used instead of traditional CNN to address the issue of inadequate local feature extraction when the model encounters vessels at multiple scales.In the decoding stage,we introduce a local-global fusion module.It enhances the localization of vascular information and reduces the semantic gap between high-and low-level features.To address the class imbalance in retinal images,we propose a hybrid loss function that enhances the segmentation ability of the model for topological structures.We conducted experiments on the publicly available DRIVE,CHASEDB1,STARE,and IOSTAR datasets.The experimental results show that our CT-MFENet performs better than most existing methods,including the baseline U-Net.展开更多
With the development of modern educational concepts and technologies,corporate financial audit is facing unprecedented challenges and opportunities.This paper first analyzes the new characteristics of corporate financ...With the development of modern educational concepts and technologies,corporate financial audit is facing unprecedented challenges and opportunities.This paper first analyzes the new characteristics of corporate financial audit in the context of modern education,including the widespread application of digital audit tools,the diversification of audit content,and the increased requirements for audit efficiency.Then,it explores the innovative practices in corporate financial audit,such as the introduction of big data analysis technology,the construction of intelligent audit platforms,and the implementation of continuous audit.The paper also conducts an in-depth study on the impact of these innovative practices on the processes,quality,and risk management of corporate financial audit.Finally,it summarizes the effectiveness of the innovation and practice of corporate financial audit in the context of modern education,and looks forward to future development trends,providing references for theoretical research and practical operations in related fields.展开更多
Unmanned aerial vehicle laser scanning(ULS)and terrestrial laser scanning(TLS)systems are effective ways to capture forest structures from top and side views,respectively.The registration of TLS and ULS data is a prer...Unmanned aerial vehicle laser scanning(ULS)and terrestrial laser scanning(TLS)systems are effective ways to capture forest structures from top and side views,respectively.The registration of TLS and ULS data is a prerequisite for a comprehensive forest structure representation.Conventional registration methods based on geometric features(e.g.,points,lines,and planes)are likely to fail due to the irregular natural point distributions of forest point clouds.Currently,automatic registration methods for forest point clouds typically rely on tree attributes(such as tree position and stem diameter).However,these methods are often unsuitable for forests with diverse compositions,complex terrains,irregular tree layouts,and insufficient common trees.In this study,an automated method is proposed to register ULS and TLS forest point clouds using ground points as registration primitives,which operates independently of tree attribute extraction and is estimated to reduce processing time by over 50%.A new evaluation method for registration accuracy evaluation is proposed,where transformation parameters from each TLS scan to the ULS obtained by the proposed registration algorithm are used to derive transformation parameters between TLS scans,which are then compared to reference parameters obtained using artificial spherical targets.Conventional ULS-TLS registration evaluation methods mostly rely on the manual corresponding points selection that is subject to inherent subjective errors,or control points in both TLS and ULS data that are difficult to collect.The proposed method presents an objective and accurate solution for ULS-TLS registration accuracy evaluation that effectively eliminates these limitations.The proposed method was tested on 12 plots with diverse stem densities,tree species,and altitudes located in a mountain forest.A total of 124 TLS scans were successfully registered to ULS data.The registration accuracy was assessed using both the conventional evaluation method and the proposed new evaluation method,with average rotation errors of 2.03 and 2.06 mrad,and average translation errors of 7.63 and 6.51 cm,respectively.The registration accuracies demonstrate that the proposed algorithm effectively and accurately registers TLS to ULS point clouds.展开更多
Background The Registrar Clinical Encounters in Training(ReCEnT)project is an Australian general practice vocational training programme with integrated and interdependent education and research functions.Trainees(regi...Background The Registrar Clinical Encounters in Training(ReCEnT)project is an Australian general practice vocational training programme with integrated and interdependent education and research functions.Trainees(registrars)contemporaneously document in-consultation clinical experience and actions.Objectives Using a realist lens,we elucidate the mechanisms underpinning project outcomes to answer questions around programme effectiveness,impacts,sustainability and the lessons and findings that are translatable to other primary care training programmes.Methods The context,input,process and product framework was used.As a means to understand the interactions between each of the interdependent components,it allows for inferences regarding causal mechanisms for specific outcomes.Results Context:ReCEnT occurs within an apprenticeship-like model of general practice vocational training entailing a central supervisor/apprentice relationship.ReCEnT has demystified the content and characteristics of registrar consultations.Input:multiple stakeholder involvement is both advantageous and a logistical challenge,with the programme’s success dependent on registrars,practices and training providers providing detailed and accurate data,with prompt subsequent processing.Process:contemporaneous consultation data collection in different stages of training constitutes a component of registrars’programmatic assessment.Product:individualised feedback provides educational benefit through reflection.Clinical and educational research questions can be addressed with resulting research translation feeding back into the programme model and government policy.Clinical behaviour change is also evaluated.Conclusion ReCEnT is unique,globally,in its scope and longevity(2010–present).Creation of meaningful,individualised feedback facilitates reflection and provides both immediate educational benefits and the substrate for further research,programme and policy design and targeted formal teaching and learning.展开更多
针对地图综合中建筑多边形化简方法依赖人工规则、自动化程度低且难以利用已有化简成果的问题,本文提出了一种基于Transformer机制的建筑多边形化简模型。该模型首先把建筑多边形映射至一定范围的网格空间,将建筑多边形的坐标串表达为...针对地图综合中建筑多边形化简方法依赖人工规则、自动化程度低且难以利用已有化简成果的问题,本文提出了一种基于Transformer机制的建筑多边形化简模型。该模型首先把建筑多边形映射至一定范围的网格空间,将建筑多边形的坐标串表达为网格序列,从而获取建筑多边形化简前后的Token序列,构建出建筑多边形化简样本对数据;随后采用Transformer架构建立模型,基于样本数据利用模型的掩码自注意力机制学习点序列之间的依赖关系,最终逐点生成新的简化多边形,从而实现建筑多边形的化简。在训练过程中,模型使用结构化的样本数据,设计了忽略特定索引的交叉熵损失函数以提升化简质量。试验设计包括主试验与泛化验证两部分。主试验基于洛杉矶1∶2000建筑数据集,分别采用0.2、0.3和0.5 mm 3种网格尺寸对多边形进行编码,实现了目标比例尺为1∶5000与1∶10000的化简。试验结果表明,在0.3 mm的网格尺寸下模型性能最优,验证集上的化简结果与人工标注的一致率超过92.0%,且针对北京部分区域的建筑多边形数据的泛化试验验证了模型的迁移能力;与LSTM模型的对比分析显示,在参数规模相近的条件下,LSTM模型无法形成有效收敛,并生成可用结果。本文证实了Transformer在处理空间几何序列任务中的潜力,且能够有效复用已有化简样本,为智能建筑多边形化简提供了具有工程实用价值的途径。展开更多
基金supported in part by the National Natural Science Foundation of China under Grant 62262073in part by the Yunnan Provincial Ten Thousand People Program for Young Top Talents under Grant YNWR-QNBJ-2019-237in part by the Yunnan Provincial Major Science and Technology Special Program under Grant 202402AD080002.
文摘In the age of big data,ensuring data privacy while enabling efficient encrypted data retrieval has become a critical challenge.Traditional searchable encryption schemes face difficulties in handling complex semantic queries.Additionally,they typically rely on honest but curious cloud servers,which introduces the risk of repudiation.Furthermore,the combined operations of search and verification increase system load,thereby reducing performance.Traditional verification mechanisms,which rely on complex hash constructions,suffer from low verification efficiency.To address these challenges,this paper proposes a blockchain-based contextual semantic-aware ciphertext retrieval scheme with efficient verification.Building on existing single and multi-keyword search methods,the scheme uses vector models to semantically train the dataset,enabling it to retain semantic information and achieve context-aware encrypted retrieval,significantly improving search accuracy.Additionally,a blockchain-based updatable master-slave chain storage model is designed,where the master chain stores encrypted keyword indexes and the slave chain stores verification information generated by zero-knowledge proofs,thus balancing system load while improving search and verification efficiency.Finally,an improved non-interactive zero-knowledge proof mechanism is introduced,reducing the computational complexity of verification and ensuring efficient validation of search results.Experimental results demonstrate that the proposed scheme offers stronger security,balanced overhead,and higher search verification efficiency.
基金provided by the Science Research Project of Hebei Education Department under grant No.BJK2024115.
文摘High-resolution remote sensing images(HRSIs)are now an essential data source for gathering surface information due to advancements in remote sensing data capture technologies.However,their significant scale changes and wealth of spatial details pose challenges for semantic segmentation.While convolutional neural networks(CNNs)excel at capturing local features,they are limited in modeling long-range dependencies.Conversely,transformers utilize multihead self-attention to integrate global context effectively,but this approach often incurs a high computational cost.This paper proposes a global-local multiscale context network(GLMCNet)to extract both global and local multiscale contextual information from HRSIs.A detail-enhanced filtering module(DEFM)is proposed at the end of the encoder to refine the encoder outputs further,thereby enhancing the key details extracted by the encoder and effectively suppressing redundant information.In addition,a global-local multiscale transformer block(GLMTB)is proposed in the decoding stage to enable the modeling of rich multiscale global and local information.We also design a stair fusion mechanism to transmit deep semantic information from deep to shallow layers progressively.Finally,we propose the semantic awareness enhancement module(SAEM),which further enhances the representation of multiscale semantic features through spatial attention and covariance channel attention.Extensive ablation analyses and comparative experiments were conducted to evaluate the performance of the proposed method.Specifically,our method achieved a mean Intersection over Union(mIoU)of 86.89%on the ISPRS Potsdam dataset and 84.34%on the ISPRS Vaihingen dataset,outperforming existing models such as ABCNet and BANet.
基金supported by the National Natural Science Foundation of China(No.62267005)the Chinese Guangxi Natural Science Foundation(No.2023GXNSFAA026493)+1 种基金Guangxi Collaborative Innovation Center ofMulti-Source Information Integration and Intelligent ProcessingGuangxi Academy of Artificial Intelligence.
文摘The rapid development of information technology and accelerated digitalization have led to an explosive growth of data across various fields.As a key technology for knowledge representation and sharing,knowledge graphs play a crucial role by constructing structured networks of relationships among entities.However,data sparsity and numerous unexplored implicit relations result in the widespread incompleteness of knowledge graphs.In static knowledge graph completion,most existing methods rely on linear operations or simple interaction mechanisms for triple encoding,making it difficult to fully capture the deep semantic associations between entities and relations.Moreover,many methods focus only on the local information of individual triples,ignoring the rich semantic dependencies embedded in the neighboring nodes of entities within the graph structure,which leads to incomplete embedding representations.To address these challenges,we propose Two-Stage Mixer Embedding(TSMixerE),a static knowledge graph completion method based on entity context.In the unit semantic extraction stage,TSMixerE leveragesmulti-scale circular convolution to capture local features atmultiple granularities,enhancing the flexibility and robustness of feature interactions.A channel attention mechanism amplifies key channel responses to suppress noise and irrelevant information,thereby improving the discriminative power and semantic depth of feature representations.For contextual information fusion,a multi-layer self-attentionmechanism enables deep interactions among contextual cues,effectively integrating local details with global context.Simultaneously,type embeddings clarify the semantic identities and roles of each component,enhancing the model’s sensitivity and fusion capabilities for diverse information sources.Furthermore,TSMixerE constructs contextual unit sequences for entities,fully exploring neighborhood information within the graph structure to model complex semantic dependencies,thus improving the completeness and generalization of embedding representations.
基金Under the auspices of the Hong Kong Research Grants Council(No.14605920,14606922,14603724,C4023-20GF,8601219,8601242,3110151)a Grant from the Research Committee on Research Sustainability of Major Research Grants Council Funding Schemes of the Chinese University of Hong Kong(CUHK,No.3133235)the Vice-Chancellor’s One-off Discretionary Fund of CUHK(No.4930787)。
文摘Perception of air pollution is subjective and context-dependent.Previous studies exploring the association between measured air pollution and perceived air quality mainly focused on air pollution levels measured in the residence-based(RB)or regional context,overlooking the mobility-based(MB)context in which people are exposed to air pollution.This study measures air pollution levels in MB,RB,and regional contexts and examines their relationships with perceived air quality across different neighborhoods and gender sub-groups of Hong Kong,China to investigate how people perceive air quality.The results indicate that particulate matter 2.5(PM_(2.5))measured in RB and the regional context significantly contributes to people’s perceived air quality compared to MB PM_(2.5).Individuals in Central and Western district of Hong Kong rely on RB,regional and MB PM_(2.5) to assess air pollution.In Sham Shui Po,RB PM_(2.5) exhibits the highest influence on people’s perceived air quality,followed by regional PM_(2.5).Women’s perceived air quality is strongly related to their RB PM_(2.5) exposure,while men’s perceived air quality is associated with both RB PM_(2.5) and regional PM_(2.5) levels.We conclude that neighborhood effects and mobility levels are the two most important factors influencing the association between meas-ured air pollution and perceived air quality.We reveal that the neighborhood effect averaging problem(NEAP)influences the associ-ation between perceived air quality and measured air pollution levels in a way that differs from health outcome-related studies.Effect-ive measures are needed to improve the public’s awareness of air pollution,and scientific control should be implemented to reduce pub-lic exposure.
文摘Based on the contextual adaptation perspective of Verschueren’s Adaptation Theory,this paper explores the Chinese translation strategies of Japanese quotation sentences in the Yang translation of The Courage of One from the perspectives of communicative context and linguistic context.The study finds that the Chinese translation of Japanese quotation sentences involves various strategies,including retaining direct quotations,converting direct quotations into statements,transforming direct quotations into attributive+noun forms,and alternating between direct and indirect quotations.This research provides a new perspective for the Chinese translation of Japanese quotation sentences and offers theoretical support for translation practices in cross-cultural communication.
文摘The current study investigated how language context and word frequency influenced vowel perception of Chinese-Japanese cognates among Chinese learners of Japanese.Focusing on orthographic cognates,participants performed a vowel detection task on cognates,manipulating language context and target language(Chinese vs.Japanese),as well as word frequency(high vs.low).We measured reaction times,perceptual sensitivity,and response criterion.For high-frequency words,consistent language contexts facilitated faster vowel detection in both languages.However,in low-frequency conditions,participants showed higher perceptual sensitivity to Chinese targets and more conservative response criteria for Japanese targets,regardless of context.These findings revealed the complex interplay between word frequency,language dominance,and context in cross-language processing.Our study contributed to the understanding of vowel perception in languages with shared orthography but distinct phonological systems,offering insights for models of cross-language cognition and second language education.Furthermore,it highlighted the importance of considering both word frequency and language-specific features in cross-language studies.
文摘In recent years,the rapid integration of artificial intelligence(AI)with various industries has led to an intelligent transformation in people’s learning and working patterns.In the field of higher vocational public music course teaching,AI provides intelligent teaching tools and learning platforms,while offering students timely and scientific support and companionship,enabling them to complete learning tasks more efficiently.How to explore the transformation path of teaching modes for higher vocational public music courses in the AI context has become a key consideration for frontline teachers.Based on this,this paper first analyzes the significance of teaching higher vocational public music courses in the AI context,and then proposes feasible transformation paths for teaching modes in combination with course characteristics for reference.
基金Supported by the National Key R&D Program of China(2022YFC3803600).
文摘Background Three-dimensional(3D)shape representation using mesh data is essential in various applications,such as virtual reality and simulation technologies.Current methods for extracting features from mesh edges or faces struggle with complex 3D models because edge-based approaches miss global contexts and face-based methods overlook variations in adjacent areas,which affects the overall precision.To address these issues,we propose the Feature Discrimination and Context Propagation Network(FDCPNet),which is a novel approach that synergistically integrates local and global features in mesh datasets.Methods FDCPNet is composed of two modules:(1)the Feature Discrimination Module,which employs an attention mechanism to enhance the identification of key local features,and(2)the Context Propagation Module,which enriches key local features by integrating global contextual information,thereby facilitating a more detailed and comprehensive representation of crucial areas within the mesh model.Results Experiments on popular datasets validated the effectiveness of FDCPNet,showing an improvement in the classification accuracy over the baseline MeshNet.Furthermore,even with reduced mesh face numbers and limited training data,FDCPNet achieved promising results,demonstrating its robustness in scenarios of variable complexity.
基金National Key R&D Program of China,Grant/Award Numbers:2018YFE0203900,2020YFB1313600German Research Foundation,Hamburg Landesforschungsförderungsprojekt Cross,Grant/Award Number:Sonderforschungsbereich Transregio 169+2 种基金Shanghai Artificial Intelligence Innovation Development Special Support Project,Grant/Award Number:3920365001Horizon2020 RISE project STEP2DYNA,Grant/Award Number:691154National Natural Science Foundation of China,Grant/Award Numbers:61773083,62206168,62276048,U1813202。
文摘Conventional model transfer techniques,requiring the labelled source data,are not applicable in the privacy-protected medical fields.For the challenging scenarios,recent source data-free domain adaptation(SFDA)has become a mainstream solution but losing focus on the inter-sample class information.This paper proposes a new Credible Local Context Representation approach for SFDA.Our main idea is to exploit the credible local context for more discriminative representation.Specifically,we enhance the source model's discrimination by information regulating.To capture the context,a discovery method is developed that performs fixed steps walking in deep space and takes the credible features in this path as the context.In the epoch-wise adaptation,deep clustering-like training is conducted with two major updates.First,the context for all target data is constructed and then the context-fused pseudo-labels providing semantic guidance are generated.Second,for each target data,a weighting fusion on its context forms the anchored neighbourhood structure;thus,the deep clustering is switched from individual-based to coarse-grained.Also,a new regularisation building is developed on the anchored neighbourhood to drive the deep coarse-grained learning.Experiments on three benchmarks indicate that the proposed method can achieve stateof-the-art results.
基金funded by the“Digital Governance Transformation in Rural Spatial Changes”(ID:22CSH015)a Youth Project under the National Social Science Fund of China.
文摘Digital rural governance is a micro-level governance practice within the broader framework of building a Digital China.It involves the integration of digital technology into rural governance to drive the digital transformation of rural governance.In recognition of the varied development stages of digital rural governance,the concept of“digital context”provides an analytical lens for exploring the differences in practical models of digital rural governance.By examining the contextual characteristics and differential mechanisms of digital rural governance,this paper delves into its social foundation and technological adaptation.The research finds that the context of digital rural governance primarily encompasses three dimensions:contextual foundation,contextual logic,and contextual optimization.First,the contextual foundation of digital rural governance manifests as the social basis,comprising the social structure of villages,the type of village development,and the age structure of villagers,which constitute the social stratification forms underlying digital rural governance.Second,the contextual logic of digital rural governance focuses on the adaptation of digital technology to rural governance,promoting the adaptation of digital technology to the rural governance foundation,village governance scenarios,and villagers’digital capabilities.Third,the contextual optimization of digital rural governance emphasizes integrating digital technology with both administrative and livelihood-oriented governance affairs at the village level.This approach leverages the governance value and functional potential of digital technology to streamline digital governance processes and enhance digital governance capabilities.As a developmental direction for the transformation of rural governance,digital rural governance must not only highlight the governance advantages of digital technology but also prioritize the inherent context of rural governance.It aims to enhance the effectiveness of rural governance through digital technology and advance high-quality digital village development tailored to local conditions.
基金supported by the National Science Fund for Distinguished Young Scholars(No.62325104).
文摘Ship detection in synthetic aperture radar(SAR)image is crucial for marine surveillance and navigation.The application of detection network based on deep learning has achieved a promising result in SAR ship detection.However,the existing networks encounters challenges due to the complex backgrounds,diverse scales and irregular distribution of ship targets.To address these issues,this article proposes a detection algorithm that integrates global context of the images(GCF-Net).First,we construct a global feature extraction module in the backbone network of GCF-Net,which encodes features along different spatial directions.Then,we incorporate bi-directional feature pyramid network(BiFPN)in the neck network to fuse the multi-scale features selectively.Finally,we design a convolution and transformer mixed(CTM)detection head to obtain contextual information of targets and concentrate network attention on the most informative regions of the images.Experimental results demonstrate that the proposed method achieves more accurate detection of ship targets in SAR images.
基金the National Natural Science Foundation of China(No.62266025)。
文摘Segmentation of the retinal vessels in the fundus is crucial for diagnosing ocular diseases.Retinal vessel images often suffer from category imbalance and large scale variations.This ultimately results in incomplete vessel segmentation and poor continuity.In this study,we propose CT-MFENet to address the aforementioned issues.First,the use of context transformer(CT)allows for the integration of contextual feature information,which helps establish the connection between pixels and solve the problem of incomplete vessel continuity.Second,multi-scale dense residual networks are used instead of traditional CNN to address the issue of inadequate local feature extraction when the model encounters vessels at multiple scales.In the decoding stage,we introduce a local-global fusion module.It enhances the localization of vascular information and reduces the semantic gap between high-and low-level features.To address the class imbalance in retinal images,we propose a hybrid loss function that enhances the segmentation ability of the model for topological structures.We conducted experiments on the publicly available DRIVE,CHASEDB1,STARE,and IOSTAR datasets.The experimental results show that our CT-MFENet performs better than most existing methods,including the baseline U-Net.
文摘With the development of modern educational concepts and technologies,corporate financial audit is facing unprecedented challenges and opportunities.This paper first analyzes the new characteristics of corporate financial audit in the context of modern education,including the widespread application of digital audit tools,the diversification of audit content,and the increased requirements for audit efficiency.Then,it explores the innovative practices in corporate financial audit,such as the introduction of big data analysis technology,the construction of intelligent audit platforms,and the implementation of continuous audit.The paper also conducts an in-depth study on the impact of these innovative practices on the processes,quality,and risk management of corporate financial audit.Finally,it summarizes the effectiveness of the innovation and practice of corporate financial audit in the context of modern education,and looks forward to future development trends,providing references for theoretical research and practical operations in related fields.
基金supported partially by the National Key Research and Development Program of China(No.2023YFF1303901)the National Natural Science Foundation of China(Nos.32171789,12411530088,and 32371654)the Joint Open Funded Project of State Key Laboratory of Geo-Information Engineering and Key Laboratory of the Ministry of Natural Resources for Surveying and Mapping Science and Geo-spatial Information Technology(No.2022-02-02).
文摘Unmanned aerial vehicle laser scanning(ULS)and terrestrial laser scanning(TLS)systems are effective ways to capture forest structures from top and side views,respectively.The registration of TLS and ULS data is a prerequisite for a comprehensive forest structure representation.Conventional registration methods based on geometric features(e.g.,points,lines,and planes)are likely to fail due to the irregular natural point distributions of forest point clouds.Currently,automatic registration methods for forest point clouds typically rely on tree attributes(such as tree position and stem diameter).However,these methods are often unsuitable for forests with diverse compositions,complex terrains,irregular tree layouts,and insufficient common trees.In this study,an automated method is proposed to register ULS and TLS forest point clouds using ground points as registration primitives,which operates independently of tree attribute extraction and is estimated to reduce processing time by over 50%.A new evaluation method for registration accuracy evaluation is proposed,where transformation parameters from each TLS scan to the ULS obtained by the proposed registration algorithm are used to derive transformation parameters between TLS scans,which are then compared to reference parameters obtained using artificial spherical targets.Conventional ULS-TLS registration evaluation methods mostly rely on the manual corresponding points selection that is subject to inherent subjective errors,or control points in both TLS and ULS data that are difficult to collect.The proposed method presents an objective and accurate solution for ULS-TLS registration accuracy evaluation that effectively eliminates these limitations.The proposed method was tested on 12 plots with diverse stem densities,tree species,and altitudes located in a mountain forest.A total of 124 TLS scans were successfully registered to ULS data.The registration accuracy was assessed using both the conventional evaluation method and the proposed new evaluation method,with average rotation errors of 2.03 and 2.06 mrad,and average translation errors of 7.63 and 6.51 cm,respectively.The registration accuracies demonstrate that the proposed algorithm effectively and accurately registers TLS to ULS point clouds.
文摘Background The Registrar Clinical Encounters in Training(ReCEnT)project is an Australian general practice vocational training programme with integrated and interdependent education and research functions.Trainees(registrars)contemporaneously document in-consultation clinical experience and actions.Objectives Using a realist lens,we elucidate the mechanisms underpinning project outcomes to answer questions around programme effectiveness,impacts,sustainability and the lessons and findings that are translatable to other primary care training programmes.Methods The context,input,process and product framework was used.As a means to understand the interactions between each of the interdependent components,it allows for inferences regarding causal mechanisms for specific outcomes.Results Context:ReCEnT occurs within an apprenticeship-like model of general practice vocational training entailing a central supervisor/apprentice relationship.ReCEnT has demystified the content and characteristics of registrar consultations.Input:multiple stakeholder involvement is both advantageous and a logistical challenge,with the programme’s success dependent on registrars,practices and training providers providing detailed and accurate data,with prompt subsequent processing.Process:contemporaneous consultation data collection in different stages of training constitutes a component of registrars’programmatic assessment.Product:individualised feedback provides educational benefit through reflection.Clinical and educational research questions can be addressed with resulting research translation feeding back into the programme model and government policy.Clinical behaviour change is also evaluated.Conclusion ReCEnT is unique,globally,in its scope and longevity(2010–present).Creation of meaningful,individualised feedback facilitates reflection and provides both immediate educational benefits and the substrate for further research,programme and policy design and targeted formal teaching and learning.
文摘针对地图综合中建筑多边形化简方法依赖人工规则、自动化程度低且难以利用已有化简成果的问题,本文提出了一种基于Transformer机制的建筑多边形化简模型。该模型首先把建筑多边形映射至一定范围的网格空间,将建筑多边形的坐标串表达为网格序列,从而获取建筑多边形化简前后的Token序列,构建出建筑多边形化简样本对数据;随后采用Transformer架构建立模型,基于样本数据利用模型的掩码自注意力机制学习点序列之间的依赖关系,最终逐点生成新的简化多边形,从而实现建筑多边形的化简。在训练过程中,模型使用结构化的样本数据,设计了忽略特定索引的交叉熵损失函数以提升化简质量。试验设计包括主试验与泛化验证两部分。主试验基于洛杉矶1∶2000建筑数据集,分别采用0.2、0.3和0.5 mm 3种网格尺寸对多边形进行编码,实现了目标比例尺为1∶5000与1∶10000的化简。试验结果表明,在0.3 mm的网格尺寸下模型性能最优,验证集上的化简结果与人工标注的一致率超过92.0%,且针对北京部分区域的建筑多边形数据的泛化试验验证了模型的迁移能力;与LSTM模型的对比分析显示,在参数规模相近的条件下,LSTM模型无法形成有效收敛,并生成可用结果。本文证实了Transformer在处理空间几何序列任务中的潜力,且能够有效复用已有化简样本,为智能建筑多边形化简提供了具有工程实用价值的途径。
