随着信息产业的飞速发展,围绕大数据搜索展开的服务已渗透到人们生活的方方面面。相关技术领域也正在经历深刻变革,如数据融合的隐私保护、场景感知的搜索意图理解、统计概率式的搜索模式等。结合国内外最新研究进展,对大数据安全高效搜索与隐私保护问题进行了研究展望:首先,从多源数据发布、用户搜索需求感知及隐私感知的智慧解答3个视角凝练了大数据安全搜索与隐私保护的科学问题;其次,提出了面向大数据的信息融合与知识萃取技术、粒度化的知识表示与推演技术、支持平台与用户互动的搜索任务表示模型、基于用户体验驱动的任务管理技术、效用与代价平衡的粒度化搜索技术和基于差分隐私的安全搜索机制等研究内容;最后,对相关的技术路线进行了展望。
With rapid development of information industry, services related to big data search have permeated into almost every aspect of human lives. Relevant technologies are undergoing profound changes, such as privacy preservation for data fusion, context-aware search intention comprehending and statistic search pattern. The issues of secure-efficient search and privacy-preserving for big data were studied and looked ahead. Firstly, three chief scientific problems were refined from the aspects of multi-source data publication, awareness of users' search requirements and privacy-aware wise solutions, respectively. Secondly, main research contents were proposed, including big data oriented information fusion and knowledge extraction, granular knowledge representation and inference, interactive representation of search task, user experience driven task management, granular search pattern balancing utility with cost, and secure search mechanism based on differential privacy. Lastly, several promising techniques in future were discussed.