Affordances of social media in collective action: the case of Free Lunch for Children in China

This paper studies the socialised affordances of social media in the processes of collective action, exploring the role of technology in the under‐researched area of civil society. We examine the case of Free Lunch for Children (FL4C), a charitable programme in China based on the microblogging platform, Weibo. Adopting the perspective of affordances‐for‐practice, we draw upon the collective action model to better understand the sociomaterial practices and social processes involving social media, and seek to address the ‘when’ and ‘how’ questions of affordances. The study generates theoretical and practical implications for understanding the role of social media in social transformation.     

Asymptotic behaviours of a class of threshold models for collective action in social networks

A class of dynamic threshold models is proposed for describing the upset of collective actions in social networks. The agents of the network have to decide whether to undertake certain action or not. They make their decision by comparing the activity level of their neighbours with a time-varying threshold, evolving according to a time-invariant opinion dynamic model. Key features of the model are a parameter representing the degree of self-confidence of the agents and the mechanism adopted by the agents to evaluate the activity level of their neighbours. The case in which a radical agent, initially eager to undertake the action, interacts with a group of ordinary agents, is considered. The main contribution of the paper is the complete characterisation of the asymptotic behaviours of the network, for three different graph topologies. The asymptotic activity patterns are determined as a function of the self-confidence parameter and of the initial threshold of the ordinary agents. Numerical validation on a real ego network shows that the theoretical results obtained for simple graph structures provide useful insights on the network behaviour in more complex settings.     

Cyber-volunteering: Social media affordances in fulfilling NGO social missions

Cyber-volunteering behavior through social media is becoming popular among nongovernmental organization (NGO) members. Cyber-volunteering activities have become more aggressive with the availability of social media in recent years. The objective of this paper is to discuss voluntary behavior in fulfilling NGO social missions through social media usage based on affordance theory. There is substantial literature discussing social media affordances in the context of enterprise organization, but there has been less literature discussing the activities of NGOs in social media. Hence, seven key affordances derived from the enterprise organization were used in the context of NGO working behavior, visibility, editability, persistence, virtual collaboration, synthetic representation, individualized, and collective. This research applies a qualitative methodology using semi-structured interviews with 25 NGO members from seven selected Islamic NGOs in Malaysia that are active in different fields of Islamic education. The interviews focus on social media affordances that have influenced the execution of social missions in the NGOs. Data was analyzed using thematic coding based on the seven identified key social media affordances. The results suggest that the social media affordances related to cyber-volunteering are achieved through promoting, training, fundraising, knowledge sharing, and problem-solving activities. These affordances are highly influenced by cyber-volunteering behavior through work culture and personal privacy. The collective, individualized, and visibility affordances are most associated with cyber-volunteering behavior, followed by persistence, virtual collaboration, and editability, and synthetic representation is found to be the least.     

An affordance perspective of enterprise social media and organizational socialization

In response to the challenge of socializing new IT employees, some IT departments are exploring the incorporation of enterprise social media (hereinafter ESM) as an informal organizational socialization tool. Because this is a relatively new phenomenon, little is known about how ESM facilitate employee socialization. In order to contribute to our understanding of how ESM affects employee socialization, this paper invokes a case study to explore how one organization’s implementation of an ESM for its IT new hire program influenced the socialization process and outcomes. To delve deeply into how the ESM influences socialization, we draw upon technology affordance theory to uncover the various first and second-order affordances actualized by different actor groups and the various outcomes resulting from the affordances. We then identify five generative mechanisms – bureaucracy circumvention, executive perspective, personal development, name recognition, and morale booster – that explain how the actualization of different strands of affordances by various groups of users produces eight different outcomes. Our results provide insights into the different affordances made possible by ESM in the context of a new hire socialization program and how these affordances have repercussions beyond those experienced by the individuals using the ESM. The results have important implications for new hire socialization and technology affordance research.     

Generating timeline summaries with social media attention

Timeline generation is an important research task which can help users to have a quick understanding of the overall evolution of one given topic. Previous methods simply split the time span into fixed, equal time intervals without studying the role of the evolutionary patterns of the underlying topic in timeline generation. In addition, few of these methods take users’ collective interests into considerations to generate timelines.We consider utilizing social media attention to address these two problems due to the facts: 1) social media is an important pool of real users’ collective interests; 2) the information cascades generated in it might be good indicators for boundaries of topic phases. Employing Twitter as a basis, we propose to incorporate topic phases and user’s collective interests which are learnt from social media into a unified timeline generation algorithm.We construct both one informativeness-oriented and three interestingness-oriented evaluation sets over five topics.We demonstrate that it is very effective to generate both informative and interesting timelines. In addition, our idea naturally leads to a novel presentation of timelines, i.e., phase based timelines, which can potentially improve user experience.     

Towards an ecological account of media choice: a case study on pluralistic reasoning while choosing email

The discourse around media choice has generated a diverse array of media choice factors originating from both the media‐based and social interaction‐based approaches. The multitude of these factors hints at the adaptive nature of media choice. Alas, how a user engages with such factors and adaptively carries out media choice has remained understudied. We undertake a field study to explore the role of a plurality of choice factors and their interactions in shaping media choice processes and outcomes. In particular, we focus on how a user identifies relationships among plural choice factors while he or she works on his or her particular choice resulting in a similar outcome – email – given a large number of alternatives. Drawing upon a theory of affordances, we propose a systemic way of narrating the dynamics of media choice as a multi‐dimensional process where a user explores her or his surroundings – a niche – as to establish media affordances that will then help her or him achieve a communication goal. We identify five relational patterns of interactions among specific choice factors: reciprocity, emergence, complementarity, re‐exploration and actualisation. These patterns are shown to be emergent and highly interdependent. We conclude by reviewing future research avenues to formulate richer ‘ecological’ accounts of media choice.     

基于网络社交媒体的子话题检测技术综述

在当前多种平台崛起的互联网背景下,与传统媒体相比,网络社交媒体中的数据具有传递速度快、用户参与度高、内容覆盖全等特点,其中存在着人们关注并发布评论的众多话题,而一个话题的相关信息中可能存在更深层次、更细粒度的子话题,针对该问题进行基于网络社交媒体的子话题检测技术的研究,这是一个新兴且不断发展的研究领域。通过社交媒体获取话题及子话题信息并参与讨论,这一方式正全方位、深层次改变着人们的生活,但是该领域技术还不成熟,且相关研究在国内尚处于起步阶段。首先,简述网络社交媒体中子话题检测的发展背景和基本概念;其次,将子话题检测技术分为七大类,对每类方法均加以介绍、对比和总结;然后,将子话题检测方式分为在线检测和离线检测两种方式,并将这两种方式进行对比,列举通用技术及两种方式下的常用技术;最后,概括了该领域当前不足及未来发展趋势。     

面向社交媒体的直接因果网络发现算法

高维时序因果网络发现是社交媒体因果关系发现的重要问题。然而,现有的时序因果关系发现方法不能发现直接因果以致因果网络推断结果不准确。针对此问题提出了一种直接因果网络发现方法。该方法考虑了时序因果模型的因果延迟、滞后期数量和条件节点集等因素,更准确地发现直接因果关系;另外,采用结合置换检验的因果关系检验方法,解决传递熵阈值难以设定的问题。实验结果表明,该方法在因果网络推断中优于现有方法,有效提升时序上直接因果网络推断的准确率,适用于发现潜在社交媒体因果关系网络。     

面向社交媒体的分级注意力表情符预测模型

将预测社交媒体表情符的任务作为文本分类问题,将输入文本映射到最有可能的伴随表情符号。首先,通过研究帖子中出现的表情符与标签之间的关系,提出一个基于标签、发帖用户、发帖时间、发帖地点的注意力机制;其次,添加表情符位置特征;最后,探讨注意力机制、分级模型对于表情符预测任务的作用,训练多种模型并比较其预测效果。实验结果表明,模型对于不同使用频率的表情符的预测效果均有显著提升,模型是可行的、高效的。     

Impact of the usage of social media in the workplace on team and employee performance

How does the usage of social media in the workplace affect team and employee performance? To address this cutting edge and up-to-date research question, we ran a quasinatural field experiment, collecting data of two matched-sample groups within a large financial service firm in China. We find that work-oriented social media (DingTalk) and socialization-oriented social media (WeChat) are complementary resources that generate synergies to improve team and employee performance. The instrumental value provided by work-oriented social media is reinforced by the expressive value provided by socialization-oriented social media, which help firms to create business value from information technology investments.     

Advanced social media sentiment analysis for short-term cryptocurrency price prediction

In recent years, the scrutiny of bitcoin and other cryptocurrencies as legal and regulated components of financial systems has been increasing. Bitcoin is currently one of the largest cryptocurrencies in terms of capital market share. Therefore, this study proposes that sentiment analysis can be used as a computational tool to predict the prices of bitcoin and other cryptocurrencies for different time intervals. A key characteristic of the cryptocurrency market is that the fluctuation of currency prices depends on people's perceptions and opinions, not institutional money regulation. Therefore, analysing the relationship between social media and web search is crucial for cryptocurrency price prediction. This study uses Twitter and Google Trends to forecast the short-term prices of the primary cryptocurrencies, as these social media platforms are used to influence purchasing decisions. The study adopts and interpolates a unique multimodel approach to analyse the impact of social media on cryptocurrency prices. Our results prove that people's psychological and behavioural attitudes have a significant impact on the highly speculative cryptocurrency prices.     

Key success factors for the launch of government social media platform: Identifying the formation mechanism of continuance intention

Social media has flourished recently, and government agencies across different levels are experimenting with launching government social media (GSM) to socialize government services, processes, and data. Both researchers and practitioners are focusing on understanding the key success factors related to the launch of GSM. This study aims to identify the key success factors by exploring the formation mechanism of individuals' continuous usage intention. Through the theoretical lens of the uses and gratifications theory (UGT), we identify the gratification factors (i.e., information seeking, social activity, content consumption, collective intelligence, and network externality) that stimulate users' continuance intention toward GSM. Furthermore, we draw upon the stimulus–organism–response (SOR) framework to develop a model for exploring the effects of gratification factors (as stimulus in the SOR framework) on individuals' online experiential states (i.e., flow experience and sense of belonging) and, subsequently, their continuance intention. The results of our survey of 336 government microblogging residents indicate that usage continuance intention is in part determined by gratification factors. Furthermore, the associations among gratification factors and users’ continuance intention can be mediated by the sense of belonging and flow experience.     

面向重大突发社会安全事件的新闻媒体国际影响力分析

大数据时代下重大突发社会安全事件的舆情主要通过媒体快速传播,但现有研究大都没有考虑新闻媒体这一特殊群体以及在某一类特定事件中新闻媒体的影响力。为了研究上述问题,提出一种综合用户间的网络结构与行为关系来评价影响力的方法,并以新疆暴恐和巴黎暴恐事件为例,得出在Twitter平台中各国新闻媒体在此类事件中的国际影响力。该评价方法可以更好地得出各新闻媒体在事件层面上的影响力。通过使用该评价方法对新疆暴恐事件和巴黎暴恐事件中新闻媒体影响力进行计算,实验结果显示,各国新闻媒体在新疆暴恐事件和巴黎暴恐事件中的影响力存在差异,说明这两起同类型事件的影响范围不同,同时也从侧面反映了各国政治立场的差异。     

User requirements of social media for the elderly: a case study in Taiwan

Numerous researchers have proposed website design norms that are suitable for the elderly. However, in the design of community platforms, elderly users were not considered in social media usage. Young and middle-aged people are the main targets in several social media platforms. To co-ordinate the digital lives of elderly users, the emphasis in this study was to determine the problems and to search for solutions. The real requirements and proper solutions for the elderly were integrated by analysing possible factors. This study is an anthropological user centred approach, to explore the verbal behaviour of senior citizens while they accessed Facebook. Facebook, which is a social media platform with a multilingual language interface, is currently used worldwide and served the purpose of an experimental base for this research. By determining the user environments that are suitable for the elderly, including web page accessibility, interface design and real social life transformation, this article proposes the factors for a social media website, the factors for the elderly to use social media platforms, a social media platform design that can be easily used by the elderly and design factors suitable for the elderly.     

A conceptual framework for developing solutions that organise social media information for emergency response teams

ABSTRACT

Social media have great power to spread information, and this is particularly noticeable when an emergency occurs. The extraction of accurate information from social media can offer an important resource for emergency management, both in terms of decision-making and increasing situational awareness. This paper describes a conceptual framework for the development of applications to treat messages from social media. It is designed to select, classify and prioritise, using parameters, messages containing information that is relevant to the emergency context. It allows a team to act on this information and to generate rescue actions that contribute to the emergency solution. It has a collaborative bias, providing perceptual, coordination and communication mechanisms. We also present an instantiation and the simulation of its use in the treatment of tweets (Twitter messages) about two emergencies: an earthquake in Mexico City (19/09/2017) and a California fire (December, 2017). The volume of messages is enormous, but most of them do not present significant value to the emergency response. We categorised those that contained relevant information. With only 2% of the tweets, it was possible to identify and prioritise messages with potential to aid in response and rescue operations.     

一种面向分布式深度学习的轻量级聚合通信库

聚合通信操作在分布式训练中应用广泛,特别是AllReduce操作被用于同步每个节点上模型的参数。为了获得更高的精度,数据集和神经网络模型的规模越来越大,节点间的通信开销在训练过程中占比很大且已成为训练加速的瓶颈。目前已有许多针对这一场景下聚合操作的优化工作,但都聚焦于操作的合理使用而不是其本身,例如通信调度和梯度量化。事实上,聚合操作与分布式训练应用之间存在许多不相匹配的地方,比如后者不要求所有节点同时同步梯度,而前者却需要。这使得针对分布式训练中聚合通信的研究是有必要的。然而发现目前分布式训练中的通信框架结构复杂、代码量大,对开展相关工作来说是不合适的。为了解决这一问题,设计并实现了一个轻量级的聚合通信库,以方便分析和改进分布式训练中的聚合操作。它支持主流框架和网络,并且架构简洁。这便于研究人员实现自定义通信操作,并能应用到主流的实验环境中以产生较广的影响。在多种情况下分别通过纯聚合操作和分布式深度学习应用来评估所设计的聚合通信库。实验结果显示,该库可以实现与MPI相近的性能,可以作为分析和研究分布式训练中梯度同步的聚合通信库。     

Contextualizing a professional social network for health care: Experiences from an action design research study

Professional social networks (PSN) are online environments where practitioners can present themselves, get in contact and socialize with coworkers, share and discuss ideas, or exchange business‐related knowledge. Despite the fact that collaboration and information sharing are becoming more relevant for delivering high‐quality services, PSN are not yet widely adopted in complex domains such as health care. While most of the literature is still focusing on the exposition of the unbound potential of PSN, this paper seeks to clarify the question of how to capture and manage the professional identity of an industry such that PSN can be purposefully anchored in the working context. Following an Action Design Research approach, we describe practical design propositions and possible tensions along the contextualization of a PSN, which was specifically catered for improving interprofessional and interorganizational collaboration in and between hospitals. We identify several implications for future research. In particular, we explain intended and unintended uses of PSN in hospitals and provide metaphors for explaining possible alternative understandings of domain engineering.     

基于社交媒体文本挖掘的个人事件检测方法

用户的社交媒体中蕴含着他们过去的个人经历和潜在的生活规律,研究其规律对预测用户未来的行为以及对用户进行个性化推荐有很大的价值。通过收集微博数据,定义了11种类型的事件,并提出了一个三阶段的Pipeline的系统,利用BERT预训练模型,分别在三个阶段使用BERT+BiLSTM+Attention、BERT+FullConnect、BERT+BiLSTM+CRF方法进行个人事件检测。从微博文本中抽取出该文本是否包含定义的事件、包含的事件类型、每种事件包含的元素等信息,具体元素为Subject（事件主语）、Object（事件元素）、Time（事件发生时间）、Place（事件发生的地点）和Tense（事件发生的时态）,从而探究用户个人时间轴上的事件变化规律来预测个人事件。在收集的真实用户微博数据集上进行实验,并与逻辑回归、朴素贝叶斯、随机森林、决策树等分类算法进行对比分析。实验结果表明,三个阶段中的BERT+BiLSTM+Attention、BERT+FullConnect和BERT+BiLSTM+CRF方法均取得了最高的F1值,验证了所提方法的有效性。最后根据所提方法抽取出的事件和其中的时间信息可视化地构建了用户的个人事件时间轴。     

时域候选优化的时序动作检测

目的 时序动作检测（temporal action detection）作为计算机视觉领域的一个热点课题,其目的是检测视频中动作发生的具体区间,并确定动作的类别。这一课题在现实生活中具有深远的实际意义。如何在长视频中快速定位且实现时序动作检测仍然面临挑战。为此,本文致力于定位并优化动作发生时域的候选集,提出了时域候选区域优化的时序动作检测方法TPO （temporal proposal optimization）。方法 采用卷积神经网络（convolutional neural network,CNN）和双向长短期记忆网络（bidirectional long short term memory,BLSTM）来捕捉视频的局部时序关联性和全局时序信息;并引入联级时序分类优化（connectionist temporal classification,CTC）方法,评估每个时序位置的边界概率和动作概率得分;最后,融合两者的概率得分曲线,优化时域候选区域候选并排序,最终实现时序上的动作检测。结果 在ActivityNet v1.3数据集上进行实验验证,TPO在各评价指标,如一定时域候选数量下的平均召回率AR@100（average recall@100）,曲线下的面积AUC （area under a curve）和平均均值平均精度mAP （mean average precision）上分别达到74.66、66.32、30.5,而各阈值下的均值平均精度mAP@IoU （mAP@intersection over union）在阈值为0.75和0.95时也分别达到了30.73和8.22,与SSN （structured segment network）、TCN （temporal context network）、Prop-SSAD （single shot action detector for proposal）、CTAP （complementary temporal action proposal）和BSN （boundary sensitive network）等方法相比,TPO的所有性能指标均有提高。结论 本文提出的模型兼顾了视频的全局时序信息和局部时序信息,使得预测的动作候选区域边界更为准确和灵活,同时也验证了候选区域的准确性能够有效提高时序动作检测的精确度。     

类别敏感的全局时序关联视频动作检测

目的 视频动作检测是视频理解领域的重要问题,该任务旨在定位视频中动作片段的起止时刻并预测动作类别。动作检测的关键环节包括动作模式的识别和视频内部时序关联的建立。目前主流方法往往试图设计一种普适的检测算法以定位所有类别的动作,忽略了不同类别间动作模式的巨大差异,限制了检测精度。此外,视频内部时序关联的建立对于检测精度至关重要,图卷积常用于全局时序建模,但其计算量较大。针对当前方法的不足,本文提出动作片段的逐类检测方法,并借助门控循环单元以较低的计算代价有效建立了视频内部的全局时序关联。方法 动作模式识别方面,首先对视频动作进行粗略分类,然后借助多分支的逐类检测机制对每类动作进行针对性检测,通过识别视频局部特征的边界模式来定位动作边界,通过识别动作模式来评估锚框包含完整动作的概率;时序建模方面,构建了一个简洁有效的时序关联模块,利用门控循环单元建立了当前时刻与过去、未来时刻间的全局时序关联。上述创新点整合为类别敏感的全局时序关联视频动作检测方法。结果 为验证本文方法的有效性,使用多种视频特征在两个公开数据集上进行实验,并与其他先进方法进行比较。在ActivityNet-1.3数据集中,该方...