现有众包研究一般都是从应用层角度出发，通过调用众包平台的底层接口，逐个解决特定的众包应用问题(例如实体识别)，而缺少一个高层的管理系统来统一支持不同的众包应用。众包数据库系统(简称众包数据库)就是为了解决通用的众包问题而提出的研究，因此类似于提出数据库系统来替代传统文件系统的研究。众包数据库面临三个挑战: (1)缺乏众包计算的可度量模型来刻画众包工人的计算特点; (2)缺少众包计算的可控理论来控制众包计算的质量、代价和延迟; (3)缺少众包计算的可优化理论来指导众包算法的优化设计。为了解决这些挑战，项目凝练的关键科学问题是众包计算的可度量、可控制、可优化问题。研究内容包括: (1)众包数据库基础理论; (2)众包数据库算子设计; (3)众包数据库查询优化; (4)众包数据库查询接口。最终建立众包数据库的基础理论，突破关键技术，研制众包数据库系统，取得国际领先的研究成果，培养若干高水平人才。

Crowdsourcing is a new computing model, which utilizes the wisdom of the crowd to address the data management and analytics tasks that cannot be completely addressed by automated processes. However existing works studied the crowdsourcing problems cases by cases from the application level. There is lack of studies from the system level that devise efficient crowdsourced databases to help requesters manage their tasks. Thus it calls for a crowdsourced database to process the crowdsourced tasks, which is similar to the traditional database designed to manage files. Crowdsourced databases pose three new research challenges. First, there is lack of crowdsourcing metric to describe the characteristics of the crowd. Second, there is lack of crowdsourcing theory to control the cost, quality and latency. Third, there is lack of crowdsourcing optimization mechanism to optimize complex crowdsourced tasks. The key scientific problem summarized in this proposal is metric model, controllable theory and optimization mechanism. The research problems include crowdsourcing theory, crowdsourcing operator, crowdsourcing optimization, and crowdsourcing interface. The objective is to establish the crowdsourcing theory, address the key problems, construct a crowdsourced database, make breakthroughs, and train innovative talents.