个性化位置隐私信息的协同保护及评估

The protection of privacy information in Location-Based Services (LBSs) has formed some typical characteristics and development trends, including multiple types on privacy information, diversity on privacy preference, personalized demand on privacy protection, complex situation of the risk of information leakage and systematization of effectiveness evaluation. All of these aggravate the difficulty of screening, protection and evaluation work on the privacy information. Existing solutions are only valid for specific types of scenarios, procedures and information, but they are difficult to cope with the innovative scenarios and the explosive diversified demands. Therefore, we conclude two scientific problems, including model construction and collaborative protection of personalized privacy information, evaluation of the risk of privacy disclosure and the protective effects. To address these problems, we identity five main entry points, privacy information screening and description, precise modeling of privacy preferences, collaborative protection of location privacy, risk prediction of privacy exposure associated with scenarios, evaluation of privacy protective effects associated with spatial and temporal multidimensional information. Then we focus on breaking through a series of key technologies, such as efficient methods of similarity-based privacy information discovery, intelligent quantification for privacy preferences, privacy protection under the circumstances of the coexistence of trusted/semi-trusted/malicious participants, privacy information value analysis based on subjective and objective background knowledge, effectiveness evaluation of privacy, availability, cost, and their correlation, in order to achieve collaborative protection and systematic evaluation of personalized privacy information in LBSs. The results of project will enrich the connotations of protection and evaluation of privacy information in LBSs, and offer the beneficial exploration on leading the theoretical research to the reality.

位置服务中的隐私信息保护已形成隐私信息多元化、隐私偏好多样化、保护需求个性化、泄露风险复杂化、效能评估系统化等特征和发展趋势，加剧了隐私信息的甄别、保护、评估等工作的难度。传统的面向特定场合、环节或信息的隐私保护方法已难以应对应用场景的革新与多样化需求的激增。为此，凝练个性化隐私信息建模与协同保护、隐私泄露风险预判与保护效果评估两个科学问题，以隐私信息快速甄别与描述、隐私偏好精准建模、多方协同的位置隐私信息保护、场景关联的隐私泄露风险预判、时空多维信息关联的保护效能评估为切入点，突破基于相似度的隐私信息快速发现、场景适应的隐私偏好智能量化、可信／半可信／恶意参与方并存的隐私信息保护、基于主客观背景知识的隐私信息价值分析、隐私性／可用性／代价关联的效能评估等关键技术，实现对个性化位置隐私信息的协同保护和系统化评估。项目研究成果将丰富位置隐私信息保护及评估的内涵，并为其实用落地进行有益探索。

本项目从位置服务中隐私信息保护信息多元化、隐私偏好多样化、保护需求个性化、泄露风险复杂化、效能评估系统化等特征和需求入手，重点研究了个性化隐私信息建模与协同保护、隐私泄露风险预判与保护效果评估两个科学问题，突破了基于相似度的隐私信息快速发现、场景适应的隐私偏好智能量化、可信／半可信／恶意参与方并存的隐私信息保护、基于主客观背景知识的隐私信息价值分析、隐私性／可用性／代价关联的效能评估等关键技术，一定程度上实现了对个性化位置隐私信息的协同保护和系统化评估。项目组在项目执行期间出版了学术专著1部，发表和录用了21篇论文，其中大部分发表在国际著名的学术期刊或重要的国际会议上，被SCI检索和待检索论文6篇，EI检索和待检索论文21篇，同时申请了专利6项，均已授权。