在线社交网络中的内隐语境完整性
Implicit Contextual Integrity in Online Social Networks
摘要：
许多真实事件表明，在线社交网络的用户需要通过提高对在线社交网络中共存的不同上下文的认识来帮助他们管理他们的交互的机制，并防止他们在这些上下文中交换不适当的信息或将敏感信息从某些上下文传播到其他上下文。上下文完整性是一种隐私理论，它根据信息共享的上下文来概念化信息共享的适当性。情境完整性的计算模型假设存在定义明确的情境，在情境中，个体扮演预定义的角色，信息共享由一组明确的规范管理。然而，在线社交网络中的上下文是已知的是隐含的，未知的先验和不断变化的;用户关系是不断发展的;和信息共享规范是隐含的。这使得当前的上下文完整性模型不适合在线社交网络。本文提出了隐式上下文完整性的第一个计算模型，提出了一个信息模型和一个信息辅助Agent，该信息辅助Agent利用信息模型学习隐式上下文、关系和信息共享规范，以帮助用户避免不适当的信息交换和不期望的信息传播。通过实验评估，我们验证了信息辅助代理的特性，这些特性表现为：即使只有一小部分用户遵循信息共享规范，并且存在恶意用户，也能推断出信息共享规范;在只看到系统和用户收发信息的部分情况下，有助于减少不适当信息的交换和敏感信息的传播;并最大限度地减少用户发出不必要警报的负担。
英文摘要：
Many real incidents demonstrate that users of Online Social Networks need mechanisms that help them manage their interactions by increasing the awareness of the different contexts that coexist in Online Social Networks and preventing them from exchanging inappropriate information in those contexts or disseminating sensitive information from some contexts to others. Contextual integrity is a privacy theory that conceptualises the appropriateness of information sharing based on the contexts in which this information is to be shared. Computational models of Contextual Integrity assume the existence of well-defined contexts, in which individuals enact pre-defined roles and information sharing is governed by an explicit set of norms. However, contexts in Online Social Networks are known to be implicit, unknown a priori and ever changing; users relationships are constantly evolving; and the information sharing norms are implicit. This makes current Contextual Integrity models not suitable for Online Social Networks. In this paper, we propose the first computational model of Implicit Contextual Integrity, presenting an information model and an Information Assistant Agent that uses the information model to learn implicit contexts, relationships and the information sharing norms to help users avoid inappropriate information exchanges and undesired information disseminations. Through an experimental evaluation, we validate the properties of Information Assistant Agents, which are shown to: infer the information sharing norms even if a small proportion of the users follow the norms and in presence of malicious users; help reduce the exchange of inappropriate information and the dissemination of sensitive information with only a partial view of the system and the information received and sent by their users; and minimise the burden to the users in terms of raising unnecessary alerts.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
