Incentive Mechanisms for Crowdblocking Rumors in Mobile Social Networks
Incentive Mechanisms for Crowdblocking Rumors in Mobile Social Networks
复制标题
移动社交网络围堵谣言的激励机制
DOI：
10.1109/tvt.2019.2930667
复制
发表时间：
2019-09-01
影响因子：
6.8
通讯作者：
Hao, Fei
中科院分区：
文献类型：
--
作者：
Lin, Yaguang;Cai, Zhipeng;Hao, Fei
移动的社交网络（MSNs）已经成为人们获取信息、表达情感和彼此交流的不可或缺的方式。但是，MSN的出现和广泛使用，也为谣言的滋生和快速传播创造了肥沃的土壤。因此，阻断谣言在MSN中的传播一直是该领域的热门话题。基于众包的思想，提出了一种新的谣言控制框架Crowdblocking，使用户能够以协作和分布式的方式实现控制方案，从而更有效地控制谣言。在所提出的框架中，出现的主要问题是如何激励更多的用户积极参与谣言阻止活动。为此，本文设计了两种有效的激励机制。首先，我们提出了一种基于Stackelberg博弈的同质控制任务的激励机制。我们从理论上分析了Stackelberg均衡，以最大化网络管理者和参与阻止谣言任务的用户的效用，并确保没有用户可以通过单方面改变当前策略来提高自己的效用。其次，针对异构控制任务，设计了实时反向拍卖激励机制，使用户有更多的自主权，自由定制自己的计划参与控制任务。此外，我们证明了该机制具有所需的性能的任务的及时性，计算效率，用户的理性，经理的盈利能力，和价格的真实性。最后，通过在真实的数据集上的仿真实验，验证了所提机制的有效性.
Mobile social networks (MSNs) have become an indispensable way for people to access information, express emotions, and communicate with each other. However, the advent and extensive use of MSNs has also created fertile soil for the breeding and rapid spread of rumors. Therefore, blocking the spread of rumors in MSNs has always been a hot topic in this field. With the idea of crowdsourcing, we propose a novel rumor control framework, called Crowdblocking, in which users can implement control schemes in a collaborative and distributed way, so that the rumors can be controlled more effectively. With the proposed framework, the main problem that arises is how to motivate more users to actively participate in rumor blocking activities. To this end, we design two effective incentive mechanisms in this paper. First, we propose an incentive mechanism based on the Stackelberg game for homogeneous control tasks. We theoretically analyze the Stackelberg equilibrium to maximize the utility of the network manager and users involved in blocking rumor tasks, and ensure that no user can improve its own utility by unilaterally changing the current strategy. Second, for heterogeneous control tasks, we design a real-time reverse auction incentive mechanism, which allows users to have more autonomy and freely customize their own plans to participate in control tasks. Also, we prove that the mechanism possesses the desired properties of task timeliness, computational efficiency, user rationality, manager profitability, and price truthfulness. Finally, we validate the efficiency of the proposed mechanisms through extensive simulation experiments on the real datasets.