基于Social Transformer的井下多人轨迹预测方法

目前煤矿井下人员轨迹预测方法中,Transformer与循环神经网络(RNN)、长短期记忆(LSTM)网络相比,在处理数据时不仅计算量小,同时还有效解决了梯度消失导致的长时依赖问题。但当环境中涉及多人同时运动时,Transformer对于场景中所有人员未来轨迹的预测会出现较大偏差。并且目前在井下多人轨迹预测领域尚未出现一种同时采用Transformer并考虑个体之间相互影响的模型。针对上述问题,提出一种基于Social Transformer的井下多人轨迹预测方法。首先对井下每一个人员独立建模,获取人员历史轨迹信息,通过Transformer编码器进行特征提取,接着由全连接层对特征进行表示,然后通过基于图卷积的交互层相互连接,该交互层允许空间上接近的网络彼此共享信息,计算预测对象在受到周围邻居影响时对周围邻居分配的注意力,从而提取其邻居的运动模式,继而更新特征矩阵,最后新的特征矩阵由Transformer解码器进行解码,输出对于未来时刻的人员位置信息预测。实验结果表明,Social Transformer的平均位移误差相较于Transformer降低了45.8%,且与其他主流轨迹预测方法LSTM,S−GAN,Trajectron++和Social−STGCNN相比分别降低了67.1%,35.9%,30.1%和10.9%,有效克服了煤矿井下多人场景中由于人员间互相影响导致预测轨迹失准的问题,提升了预测精度。

Rural Development Transformation and Social Governance from the Perspective of Specialization:A Case Study of Ruiling Village in Guangzhou City,China

The reorganization of the global production network and the increased mobility of urban-rural elements have accelerated the differentiation of rural spaces,resulting in a trend towards specialization in rural areas.This paper focuses on the concept of specialization,and takes a renowned bonsai specialized village in Guangzhou City,China as a typical case to explore its transformation and development process,action logic and coupling mechanism with the social governance.The results show that:1)the development of rural specialization in Ruiling Village is shaped by several key actors,such as local governments,grassroots organizations,villagers,social organizations,policies,markets and infrastructure.The transfer of these key actors and the reconstruction of actor-networks in different stages have an important influence on the transformation and realization of rural development goals.2)Through subjects-driven,organization management and resource connection,social governance supports the specialization of rural industrial organization,operation,technology and mode.The governance structure shifts from decentralization to networking,refinement and integration.A coupling community of social governance and specialization is created by the social governance system of‘local government-grassroots organizationseconomic cooperatives-villagers’,which links industries,industrial chains and cooperatives.3)Social governance integrates into the value chain extension process of rural specialized development with a central,multi-level organizational structure.It promotes the construction of the whole value chain through organizational planning,resource integration and image shaping,thus realizing the value,branding and organization of the industry,as well as the integration of rural sustainable development and governance mechanism.Social governance has a profound impact on the development of rural specialization in terms of subject,structure and mode,and participates in the construction of the whole value chain of rural areas.This paper provides a new perspective for understanding the coupling effect of social governance in the rural development transformation,and promotes the rural sustainable development and governance.

Actual Value of Rural Cultural Auditorium Construction from the Perspective of Social Transformation:A Case Study of Wenzhou City

From the perspective of social transformation,this paper analyzed the actual value of the construction of Wenzhou rural cultural auditorium,and it was found that it has important value in promoting the transformation of rural culture,economy,and grassroots governance.The construction of rural cultural auditorium is helpful for improving the present backward status of rural public cultural service supply,and promoting the integration and development of rural culture and urban culture;cultivating new rural economic entities,injecting action force for rural economic transformation and development;providing space guarantee for citizen participation in rural grassroots social governance,and is favorable for enhancing the awareness of public participation in rural grassroots governance.However,in the context of comprehensive rural revitalization,the value of cultural auditorium construction has not yet been fully realized.It is necessary to further enhance its value in rural social transformation from three aspects:auditorium construction,daily activity undertaking,and residents'participation in cultivation.

The Ascent of Money and Social Transformation Among the Badeng of Sarawak

This article examines the Badeng Kenyah,a people in Sarawak’s interior who are experiencing the shift from subsistence living to a money economy.Ethnographic description and oral history are used to examine the inside factors(such as the individual search for livelihood)and outside forces(government and the market)that brought about this shift.Some philosophical reflection addresses the impact of money on individuality and social life,with inspiration drawn from the classic works of Karl Polanyi and Georg Simmel.When the myth of money becomes accepted as a norm,human self-interest naturally guides individual behavior toward commodification,a path negotiated through those moral values that a community has maintained intact and those it has redefined to suit a monetary economy.For better and for worse,commodification works vast changes on social relations and the environment.What the Badeng experienced is what all modern peoples have gone through at one point or another.By looking at what happened to these particular people,we see again the experience that has shaped us all.

Impact of Chinese social transformation on the reform of social st）orts management system

China after the efforts of reform and opening up for 30 years, has established a relatively complete featuring system of sports administration, providing a system of protection for sports development in China. With the further improvement of the socialist market economy, the existing sports administrative functions are incompatible with the economic and social side of the market gradually. The current sports administrative functions can no longer adapt to the development and deeper reforms of sports system are imminent. Only by speeding up the reform of sports administration, can eliminate obstacles and promote rapid and healthy development of China＇s sports industry. In this paper, the object is the administrative functions of the sport in the social transformation period, through expert interviews, literature, and logical analysis, comparative study to discuss the problems of the administrative functions of sport in the social transformation period and propose appropriate countermeasures. Studies suggest that： sports administration departments should focus to implementing national policies, the development of the sports industry policy research and development planning, management and delivery of services. In accordance with the uniform principles of responsibility, rights and interests, to further clarify the responsibilities of all levels of sports administration, handling good working relationship between them. By Complete separation between rowing and steering, governance and administration, government and industry, to achieve ＂four transformations＂, and to break plan ＂all-round＂ type of government in the economic system, and establish ＂limited＂ government, effectively carry out the functions of guidance, policy regulation, market supervision, social management and public services.

Study on Social Service Transformation of New Established Undergraduate Institutions in China: Based on Text Analysis on News on Official Website

The transformation of social services is one of the key tasks for the undergraduate university changing to be an applied university. This article made a text analysis on the news about social service from the column of "university events”. It shows that the social service of newly-building undergraduate university has distinctive features and has an orientation of "localness, application, openness”, which is a natural advantage. It also has the following problems: too many plans lacking systematicness, core competence and sense of participation, lower service level, out-of-date philosophy and so on. It is suggested that newly-building undergraduate university make an inner-driven transformation, with direct and active service to others.

Legal thinking under legal globalization and social transformation

China now is in an important period of social transformation, it puts forward unique requirements for changes from the ethics community to the legal social, and also for law operation, legal developments and public awareness of the laws. Law globalization both is a result of globalization in other areas, but also a tool to promote other areas of globalization, and its essence is the performance of the laws of all countries in the world being more interdependent to further strengthen. From a legal point to analysis, to identify the root causes and find ways to solve and to provide a useful reference for the current socialist harmonious society is very necessary.

Architecture as a Means of Developing Social Exchange:Learning from a Danish Case Study of Refurbishment and Urban Transformation in a Disadvantaged Housing Area

Disadvantaged housing areas from the postwar period typically consist of relatively monotonous and monofunctional multi-storey blocks and have a separate traffic system rendering the areas isolated enclaves in the urban fabric.It is argued that this spatial isolation contributes to social isolation,and current refurbishments are therefore aimed at linking the housing areas with the surrounding neighbourhoods to enable social exchange between the residents and people from the surrounding areas.Based on a case study in a Danish social housing area,this paper discusses to what extent architectural development and urban strategic transformations can contribute to enabling social exchange in disadvantaged housing areas.Do neighbours enter the disadvantaged housing area as intended?If so,how does this influence their perception of the housing area?The analysis shows that architectural development and urban transformations are interdependent in establishing a social mix in disadvantaged housing areas.It is argued that refurbishments can reduce the stigma surrounding the housing area and its residents.However,the applicability of such strategies is highly dependent on the local context of the housing area.

基于多模态掩码Transformer网络的社会事件分类

多模态社会事件分类的关键是充分且准确地利用图像和文字2种模态的特征。然而,现有的大多数方法存在以下局限性:简单地将事件的图像特征和文本特征连接起来,不同模态之间存在不相关的上下文信息导致相互干扰。因此,仅仅考虑多模态数据模态间的关系是不够的,还要考虑模态之间不相关的上下文信息(即区域或单词)。为克服这些局限性,提出一种新颖的基于多模态掩码Transformer网络(MMTN)模型的社会事件分类方法。通过图-文编码网络来学习文本和图像的更好的表示。将获得的图像和文本表示输入多模态掩码Transformer网络来融合多模态信息,并通过计算多模态信息之间的相似性,对多模态信息的模态间的关系进行建模,掩盖模态之间的不相关上下文。在2个基准数据集上的大量实验表明:所提模型达到了最先进的性能。

Graph Transformer for Communities Detection in Social Networks

Graphs are used in various disciplines such as telecommunication,biological networks,as well as social networks.In large-scale networks,it is challenging to detect the communities by learning the distinct properties of the graph.As deep learning hasmade contributions in a variety of domains,we try to use deep learning techniques to mine the knowledge from large-scale graph networks.In this paper,we aim to provide a strategy for detecting communities using deep autoencoders and obtain generic neural attention to graphs.The advantages of neural attention are widely seen in the field of NLP and computer vision,which has low computational complexity for large-scale graphs.The contributions of the paper are summarized as follows.Firstly,a transformer is utilized to downsample the first-order proximities of the graph into a latent space,which can result in the structural properties and eventually assist in detecting the communities.Secondly,the fine-tuning task is conducted by tuning variant hyperparameters cautiously,which is applied to multiple social networks(Facebook and Twitch).Furthermore,the objective function(crossentropy)is tuned by L0 regularization.Lastly,the reconstructed model forms communities that present the relationship between the groups.The proposed robust model provides good generalization and is applicable to obtaining not only the community structures in social networks but also the node classification.The proposed graph-transformer shows advanced performance on the social networks with the average NMIs of 0.67±0.04,0.198±0.02,0.228±0.02,and 0.68±0.03 on Wikipedia crocodiles,Github Developers,Twitch England,and Facebook Page-Page networks,respectively.

The Transformation of Environmental Governance from the Point of the Interaction between Government and National People's Congress

Taking cadmium pollution in X River as an example, the interaction between National People＇s Congress and local government was de- scribed. How the local government transforms its role in environment protection with the help of several deputies to National People＇s Congress was expounded from the angle of social construction, which also enlightened the way to build an environment protection road with Chinese characteristics.

From Pariah to Prime Minister:Transformation in the Images of the Indian Community in the Caribbean

Between 1838 and 1917,close to half a million Indians arrived in the Caribbean as labourers under the British scheme of Indian indenture.Upon completion of their respective periods of indenture,most of these immigrants did not return to India,opting instead to make the Caribbean their home.Due to a range of very provocative social,economic,and political factors all indubitably tied to British colonial agenda,Indians were unceremoniously implanted into the various colonies where they continued to exist in a pariah like state for the duration of the system of Indian indenture.Upon termination of the system by 1920,Indians had begun a very checkered journey towards integration in the wider Caribbean societies.For three Caribbean countries,these journeys culminated in the election of Indian prime ministers and presidents;for many others,they were often marked by the emergence of profoundly eminent Indians who carved out niches for themselves in the social,political,economic,and cultural landscapes of the Caribbean community.As communities within nations,the experiences of Indians were marked by varying facets and degrees of growth and development in the locations with substantially large numbers of Indians.In those locations with substantially smaller numbers,the Indian communities were either assimilated as a cultural/ethnic group into the wider social fabrics or,have been struggling to retain their Indian identity as a minority group.The objective of this paper is to chart the many trajectories that have emerged with and for Indians in the Caribbean,and to explore the many images and imaginings of Indians in the Caribbean that have been framed along these trajectories;both of and among themselves and from the larger Caribbean framework.

基于多模态双协同Gather Transformer网络的虚假信息检测方法

社交媒体网站是人们在日常生活中分享信息、表达和交换意见的便捷平台。随着用户数量的不断增加,社交媒体网站上出现了大量的信息数据。然而,由于用户没有检查共享信息的可靠性,这些信息的真实性难以保证,从而导致大量虚假信息在社交媒体上广泛传播。然而,现有方法大多存在以下局限性:1)大多数方法通过简单提取文本与视觉特征,将其拼接后得到多模态特征来进行虚假信息判断,忽略了模态间和模态内细粒度内在联系,缺乏对关键信息的检索和筛选;2)多模态信息间缺乏指导性的特征提取,文本和视觉等特征之间缺乏交互增强,对多模态信息的理解不足。为了应对这些挑战,提出了一种新颖的基于多模态双协同Gather Transformer网络(Multimodal Dual-Collaborative Gather Transformer Network,MDCGTN)的虚假信息检测方法。在MDCGTN模型中,通过文本-视觉编码网络对文本和视觉信息的特征表示进行提取,将获得的视觉和文本特征表示输入多模态Gather Transformer网络进行多模态信息融合,使用Gather机制提取关键信息,充分捕捉和融合模态内和模态间细粒度关系。此外,设计了一个双协同机制对社交媒体帖子的多模态信息进行整合,以实现模态之间信息的交互和增强。在两个公开可用的基准数据集上进行了大量实验,结果表明,与现有的先进基准方法相比,所提方法准确率明显提升,证明了其对于虚假信息检测的优越性能。

基于时空Transformer的社交网络信息传播预测

随着社交网络的日益普及和广泛应用,信息传播预测逐渐成为了社交网络分析领域的一个热点研究问题.之前大部分研究要么只利用信息传播序列,要么只利用用户之间的社交网络来进行预测,难以对信息传播过程的复杂性进行有效建模.此外,常用于信息传播预测的循环神经网络(recurrent neural network,RNN)及其变体难以有效捕获信息之间的相关性.为解决上述问题,提出了一个新的基于时空Transformer的社交网络信息传播预测模型STT.该模型首先构建由社交网络图和动态传播图组成的异构图并使用图卷积网络(graph convolutional network,GCN)来学习用户的结构特征;然后将用户的时序特征和结构特征放入到Transformer中进行融合来获取时空特征;为有效融合用户的时序特征和结构特征,提出了一种新的残差融合方式来替代Transformer中原有的残差连接;最后利用Transformer来进行信息传播预测.真实数据集上的大量实验验证了模型STT的有效性.

基于改进Transformer的社交媒体谣言检测

随着互联网的快速发展,社交媒体日益广泛而深刻地融入人们日常生活的各个方面。社交媒体逐渐成为人们彼此之间用来分享意见、见解、经验和观点的工具和平台,是人们获取分享信息、表达交流观点的主要途径。社交媒体在互联网的沃土上蓬勃发展,爆发出令人眩目的能量。由于社交媒体的开放性,用户规模庞大且来源复杂众多,容易产生各种各样的谣言虚假信息。社交媒体谣言左右着网民对事件的认识、动摇着社会的稳定。因此,如何准确高效地检测谣言成为当下亟待解决的问题。现有基于Transformer的社交媒体谣言检测模型忽略了文本位置信息。为有效提取文本位置信息,充分利用文本潜在信息,提出了一种基于改进Transformer的社交媒体谣言检测模型。该模型从相对位置和绝对位置两方面对传统Transformer进行改进:一方面采用可学习的相对位置编码捕捉文本的方向信息和距离信息;另一方面采用绝对位置编码将不同位置词语映射到不同特征空间。实验结果表明,与其他基准模型相比,所提模型在Twitter15、Twitter16和Weibo 3种数据集上的准确率分别提高了0.9%、0.6%和1.4%。实验结果验证了所提的位置编码改进有效,基于位置编码改进的Transformer模型可显著提升社交媒体谣言检测效果。

Effective short text classification via the fusion of hybrid features for IoT social data

Nowadays short texts can be widely found in various social data in relation to the 5G-enabled Internet of Things (IoT). Short text classification is a challenging task due to its sparsity and the lack of context. Previous studies mainly tackle these problems by enhancing the semantic information or the statistical information individually. However, the improvement achieved by a single type of information is limited, while fusing various information may help to improve the classification accuracy more effectively. To fuse various information for short text classification, this article proposes a feature fusion method that integrates the statistical feature and the comprehensive semantic feature together by using the weighting mechanism and deep learning models. In the proposed method, we apply Bidirectional Encoder Representations from Transformers (BERT) to generate word vectors on the sentence level automatically, and then obtain the statistical feature, the local semantic feature and the overall semantic feature using Term Frequency-Inverse Document Frequency (TF-IDF) weighting approach, Convolutional Neural Network (CNN) and Bidirectional Gate Recurrent Unit (BiGRU). Then, the fusion feature is accordingly obtained for classification. Experiments are conducted on five popular short text classification datasets and a 5G-enabled IoT social dataset and the results show that our proposed method effectively improves the classification performance.

基于跨模态Transformer的多模态细粒度情感分析方法研究

多模态细粒度情感分析任务要求根据文本以及对应的图片信息,判断文本中每个观点实体的情感倾向。目前基于文本的细粒度情感分析模型无法对多模态数据进行建模,因此论文针对多模态细粒度情感分析任务,提出了一种基于跨模态Transformer的层次化神经网络模型。该模型利用从预训练模型中抽取出的文本以及图片特征,通过调整跨模态Transformer的输入来建模文本和图像之间的信息交互,从而达到利用图片信息来辅助文本信息进行情感分析的目的。论文在两个真实的多模态社交媒体数据集上开展了实验,并与其他方法展开对比,验证了论文提出的方法的有效性。

Pre-service Teachers’ Beliefs of Competence With Teaching Socially Sensitive Issues

This paper is part of a larger cross-sectional, phenomenological study that examined pre-service teachers’ belief of the competencies they possess to teach socially sensitive issues at secondary level in Trinidad and Tobago. Purposive sampling resulted in 39 pre-service social studies teachers (Years 2 to 4) who were exposed to a methodology course in a social studies specialisation programme and nine social studies lecturers, at the University of Trinidad and Tobago. The study is located in the mixed methods paradigm, with a quan/QUAL design. This paper represents discussions and findings from two questionnaires and four focus groups analysed through the IBM SPSS Statistics 22 software for the quantitative analysis and content and micro-interlocutor analyses for the qualitative analysis. The findings show that pre-service teachers believe they have the competencies to teach socially sensitive issues;yet they had not obtained these from the designated methodology course. Further that both contextual and personal factors affect the teaching of socially sensitive issues. Six themes, with accompanying competencies emerge from the data and frame affective teaching of socially sensitive issues. Arising out of the findings, the author developed a competency-based framework that has implications for curriculum development and transformational pedagogy.

“外察”与“内省”:环境威胁引发的绿色消费意愿研究——基于道德情绪的双路径模型

营销学者在探究绿色消费议题方面做了众多开创性工作,然而针对诱因不同的环境威胁是否及通过怎样的心理路径影响消费者绿色消费意愿的研究仍然匮乏。本文将环境威胁分为自然因素引发的环境威胁和人为因素引发的环境威胁两类,通过4组实验检验了两类环境威胁影响消费者绿色消费意愿的心理机制。研究结果证实了无论哪类环境威胁,都会对绿色产品购买行为产生积极影响,但不同类型的环境威胁会通过差异化认知过程和情绪反应影响绿色产品购买行为。当消费者面对自然因素引发的环境威胁时,其“外察”思维模式会起主导作用,此时敬畏感是促进绿色产品购买意愿的中介变量;当消费者面对人为因素引发的环境威胁时,其“内省”思维模式会起主导作用,此时内疚感是促进绿色产品购买意愿的中介变量。研究也证实了环境威胁对道德情绪的影响会受自然联结性和道德提升感的调节。

角色演变视角下辟谣信息对社交媒体意见领袖形成的影响——基于舆论领导法则

[目的/意义]探究辟谣信息对社交媒体用户的影响,特别是对意见领袖形成现象机制的挖掘。构建社交媒体用户角色演变行为模型,研究人员可以更好地了解意见领袖的形成原因,刻画辟谣信息对社交媒体用户的作用机制,有助于对参与辟谣的“准意见领袖”进行合理引导,以提高社交媒体舆情管理的效率和准确性。[方法/过程]首先,对微博社交媒体平台中的特定主题事件进行跟踪,识别并筛选出“新生意见领袖”群体,并利用质性访谈对此群体进行深度访谈。然后,基于程序化扎根理论方法,利用NVivo 12软件进行扎根编码构建模型,并通过理论饱和度检验验证模型的完整性。最后,对社交媒体用户角色演变的原因进行深入分析和探讨。[结果/结论]在辟谣信息作用下,个体主观因素、社会环境因素和辟谣信息因素对一般用户向新生意见领袖的演变行为产生重要影响,其中个体主观因素在辟谣信息的作用下对意见领袖形成的影响更加显著,自我认知清晰、高情绪需求且情绪稳定的个体更容易成为影响力持久的意见领袖;社会公众认可、实质性奖励和政策法律支持共同构成社会环境因素,是意见领袖形成的重要保障;辟谣信息因素中的辟谣信息热度、可信度和发布质量等方面都会直接影响“潜在意见领袖”用户群体对辟谣事件的把握,进而干预其参与辟谣的决定。