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
高中英语词汇教学面临输入碎片化、语境割裂化与输出单一化的困境,制约学生词汇深度习得与核心素养发展。文章以北师大版高一英语必修一“Unit 2 Sports and Fitnes”为例,提出三维融合破解路径:其一,通过跨单元主题语义场重构实现词汇...高中英语词汇教学面临输入碎片化、语境割裂化与输出单一化的困境,制约学生词汇深度习得与核心素养发展。文章以北师大版高一英语必修一“Unit 2 Sports and Fitnes”为例,提出三维融合破解路径:其一,通过跨单元主题语义场重构实现词汇聚合输入;其二,运用隐喻识别、跨域映射、文化调适组成的三阶策略促进认知加工;其三,设计控制性、指导性、自主性三级任务链驱动语言产出。展开更多
Through the centurial research upon the thoughts of Marx,there is a turning from historical restoration to fusion of horizons,which is closely related to the research methodology of the thoughts of Marx as well as the...Through the centurial research upon the thoughts of Marx,there is a turning from historical restoration to fusion of horizons,which is closely related to the research methodology of the thoughts of Marx as well as the big change of the academic study,economy and society.To have a better understanding of Marxism,communication between readers and authors,integration of the theory and practice are necessary.Besides,we should distinguish academic study from politics,definition from significance,and genuine knowledge from misunderstanding.Different researchers study Marxism from different aspects and in different historical contexts,thus forming different images of Marx and different understandings of Marxism.Therefore,research upon the thoughts of Marx and Marxism cannot be completed once and for all.展开更多
深度学习算法在虚假新闻检测关键特征提取方面具有优势,然而,现有的基于深度学习的多模态虚假新闻检测方法仍存在不足之处,例如,从输入的图像与文本中提取特征并进行特征融合时存在融合不充分的问题。针对这一问题,该文提出了一种基于...深度学习算法在虚假新闻检测关键特征提取方面具有优势,然而,现有的基于深度学习的多模态虚假新闻检测方法仍存在不足之处,例如,从输入的图像与文本中提取特征并进行特征融合时存在融合不充分的问题。针对这一问题,该文提出了一种基于多模态上下文融合及语义增强的虚假新闻检测模型MCEFSE(Multimodal Context based Early Fusion and Semantic Enhancement)。首先,该文利用预训练语言模型BERT对句子进行编码。同时,以Swin Transformer模型作为主要框架,在早期视觉特征编码时引入文本特征,增强语义交互。此外,我们还使用InceptionNetV3作为图像模式分析器。最后,对文本语义、视觉语义和图像模式特征进行细化和融合,得到最终的多模态特征表示。结果显示,MCEFSE模型在微博数据集和微博-21数据集上的准确率分别为0.921和0.932,验证了该方法的有效性。展开更多
People with different cultural backgrounds would inevitably encounter contradictions and conflicts in their communication due to their different values and lifestyles.This paper draws on the research findings of high-...People with different cultural backgrounds would inevitably encounter contradictions and conflicts in their communication due to their different values and lifestyles.This paper draws on the research findings of high-low context and high-low power distance theories to analyze the cultural contradictions and conflicts in The Joy Luck Club.The result shows that mutual respect and understanding are needed in intercultural communication and to achieve cultural fusion,people should absorb new culture on the basis of inheriting their traditional culture.展开更多
针对现有的事件因果关系识别方法未考虑导入外部知识后产生的噪声干扰,导致事件表示歧义增加,从而影响识别效果的问题,提出了基于外部词库和超图降噪的事件因果关系识别模型(event causality identification model based on external vo...针对现有的事件因果关系识别方法未考虑导入外部知识后产生的噪声干扰,导致事件表示歧义增加,从而影响识别效果的问题,提出了基于外部词库和超图降噪的事件因果关系识别模型(event causality identification model based on external vocabulary and hypergraph denoising,EHDM)。首先,从外部词库中检索事件的背景知识来丰富事件的语义信息,并对带有背景知识的事件描述进行编码。然后,根据事件背景知识中多个关系对应的知识特征构建超图,通过超图卷积神经网络和多头注意力机制进一步处理特征,得到降噪后的事件特征表示。接着,对事件及其上下文进行编码得到基于上下文的特征表示,并与降噪后的事件特征表示一起通过门单元进行特征融合。最后,将融合的特征表示输入多层感知器得到预测值,实现因果关系识别。结果表明,EHDM在因果时间库(causal-timebank,CTB)数据集句内方面的F1分数相比关系图卷积网络(relation graph convolutional networks,RGCN)模型提高了1.5个百分点,在事件情节链(event story line,ESL)数据集句内方面的F1分数相比RGCN模型提高了2.4个百分点,跨句、总体方面的F1分数相比事件关系图变换器模型分别提高了2.1、3.0个百分点。该研究证实了EHDM能有效应用于事件因果关系识别领域。展开更多
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.
文摘现有的基于深度学习的医学图像分割方法,大多是利用大量的训练数据拟合检测网络,以获得优异的检测性能。这些方法往往需要较大的模型参数,导致检测实时性较差。为此,提出了基于局部上下文引导特征深度融合轻量级医学分割网络(local context guided feature deep fusion lightweight medical segmentation network,LCGML-net)。LCGML-net通过精确的特征选择与特征融合来减少模型拟合所需的参数数量,从而在保证检测精度的同时实现更小的模型参数。在特征提取阶段和映射阶段,分别通过提取和融合目标的多层次多尺度局部上下文特征来丰富特征表达和精准分割。在STARE、CHASEDB1和KITS19等多个基准数据集上开展的实验证明,与其他方法相比,所提出的LCGML-net具有最佳的检测性能和最小的模型参数。
文摘高中英语词汇教学面临输入碎片化、语境割裂化与输出单一化的困境,制约学生词汇深度习得与核心素养发展。文章以北师大版高一英语必修一“Unit 2 Sports and Fitnes”为例,提出三维融合破解路径:其一,通过跨单元主题语义场重构实现词汇聚合输入;其二,运用隐喻识别、跨域映射、文化调适组成的三阶策略促进认知加工;其三,设计控制性、指导性、自主性三级任务链驱动语言产出。
基金the staged achievement of the national fund program of social sciences for the western regions-The Historical Evolution and Contemporary Development of the Research Methodology of the Formation History of the Marxist Philosophy in China(12XZX003)
文摘Through the centurial research upon the thoughts of Marx,there is a turning from historical restoration to fusion of horizons,which is closely related to the research methodology of the thoughts of Marx as well as the big change of the academic study,economy and society.To have a better understanding of Marxism,communication between readers and authors,integration of the theory and practice are necessary.Besides,we should distinguish academic study from politics,definition from significance,and genuine knowledge from misunderstanding.Different researchers study Marxism from different aspects and in different historical contexts,thus forming different images of Marx and different understandings of Marxism.Therefore,research upon the thoughts of Marx and Marxism cannot be completed once and for all.
文摘深度学习算法在虚假新闻检测关键特征提取方面具有优势,然而,现有的基于深度学习的多模态虚假新闻检测方法仍存在不足之处,例如,从输入的图像与文本中提取特征并进行特征融合时存在融合不充分的问题。针对这一问题,该文提出了一种基于多模态上下文融合及语义增强的虚假新闻检测模型MCEFSE(Multimodal Context based Early Fusion and Semantic Enhancement)。首先,该文利用预训练语言模型BERT对句子进行编码。同时,以Swin Transformer模型作为主要框架,在早期视觉特征编码时引入文本特征,增强语义交互。此外,我们还使用InceptionNetV3作为图像模式分析器。最后,对文本语义、视觉语义和图像模式特征进行细化和融合,得到最终的多模态特征表示。结果显示,MCEFSE模型在微博数据集和微博-21数据集上的准确率分别为0.921和0.932,验证了该方法的有效性。
文摘People with different cultural backgrounds would inevitably encounter contradictions and conflicts in their communication due to their different values and lifestyles.This paper draws on the research findings of high-low context and high-low power distance theories to analyze the cultural contradictions and conflicts in The Joy Luck Club.The result shows that mutual respect and understanding are needed in intercultural communication and to achieve cultural fusion,people should absorb new culture on the basis of inheriting their traditional culture.
文摘针对现有的事件因果关系识别方法未考虑导入外部知识后产生的噪声干扰,导致事件表示歧义增加,从而影响识别效果的问题,提出了基于外部词库和超图降噪的事件因果关系识别模型(event causality identification model based on external vocabulary and hypergraph denoising,EHDM)。首先,从外部词库中检索事件的背景知识来丰富事件的语义信息,并对带有背景知识的事件描述进行编码。然后,根据事件背景知识中多个关系对应的知识特征构建超图,通过超图卷积神经网络和多头注意力机制进一步处理特征,得到降噪后的事件特征表示。接着,对事件及其上下文进行编码得到基于上下文的特征表示,并与降噪后的事件特征表示一起通过门单元进行特征融合。最后,将融合的特征表示输入多层感知器得到预测值,实现因果关系识别。结果表明,EHDM在因果时间库(causal-timebank,CTB)数据集句内方面的F1分数相比关系图卷积网络(relation graph convolutional networks,RGCN)模型提高了1.5个百分点,在事件情节链(event story line,ESL)数据集句内方面的F1分数相比RGCN模型提高了2.4个百分点,跨句、总体方面的F1分数相比事件关系图变换器模型分别提高了2.1、3.0个百分点。该研究证实了EHDM能有效应用于事件因果关系识别领域。
文摘目的 近年来,Transformer跟踪器取得突破性的进展,其中自注意力机制发挥了重要作用。当前,自注意力机制中独立关联计算易导致权重不明显现象,限制了跟踪方法性能。为此,提出了一种融合上下文感知注意力的Transformer目标跟踪方法。方法 首先,引入SwinTransformer(hierarchical vision Transformer using shifted windows)以提取视觉特征,利用跨尺度策略整合深层与浅层的特征信息,提高网络对复杂场景中目标表征能力。其次,构建了基于上下文感知注意力的编解码器,充分融合模板特征和搜索特征。上下文感知注意力使用嵌套注意计算,加入分配权重的目标掩码,可有效抑制由相关性计算不准确导致的噪声。最后,使用角点预测头估计目标边界框,通过相似度分数结果更新模板图像。结果 在TrackingNet(large-scale object tracking dataset)、LaSOT(large-scale single object tracking)和GOT-10K(generic object tracking benchmark)等多个公开数据集上开展大量测试,本文方法均取得了优异性能。在GOT-10K上平均重叠率达到73.9%,在所有对比方法中排在第1位;在LaSOT上的AUC(area under curve)得分和精准度为0.687、0.749,与性能第2的ToMP(transforming model prediction for tracking)相比分别提高了1.1%和2.7%;在TrackingNet上的AUC得分和精准度为0.831、0.807,较第2名分别高出0.8%和0.3%。结论 所提方法利用上下文感知注意力聚焦特征序列中的目标信息,提高了向量交互的精确性,可有效应对快速运动、相似物干扰等问题,提升了跟踪性能。
