为科学服务的大数据

Big Data for Better Science

        郭毅可(伦敦帝国理工学院数据科学研究所 伦敦 SW7 2AZ)
        潘为(伦敦帝国理工学院数据科学研究所 伦敦 SW7 2AZ)
        于思淼(伦敦帝国理工学院数据科学研究所 伦敦 SW7 2AZ)
        吴超(伦敦帝国理工学院数据科学研究所 伦敦 SW7 2AZ)
        王世才(伦敦帝国理工学院数据科学研究所 伦敦 SW7 2AZ)

         大数据;数据科学;数据驱动的科学研究

        big data;data science;data-driven scientific research

        数据驱动的科研活动已蔚为大观,然而厘清关于数据研究的基本问题仍是数据科学的首要任务。文章根据伦敦帝国理工学院建设数据科学研究院的经验,将数据科学聚焦于交叉研究上,讨论从数据整合与理解,到数据感知与交互,再到数据学习与认知,最后到数据交换与经济的完整链条,并结合开展的科研实践工作,分析了其中的基本研究问题。

        Data driven scientific research has now gain great prosperity. However, we believe that the principle task of data science is to understand the basic problems within data research. In this paper, based on our experience in building the Data Science Institute in Imperial College London, we consider data science as the core of interdisciplinary research, and discuss the whole pipeline of data science research, including data integration and understanding, data sensing and interaction, data learning and cognition, and data exchange and economy. We discuss these basic scientific problems based on our practices in practice. We hope the work presented in this paper can bring thinking and discussion in a larger scale.