面向工业互联网复杂要素的多源特征融合表征与组织推理方法

工业互联网已成为国家创新发展战略，工业互联网复杂要素的多源异构数据表征深度融合与深层知识推理方法是工业互联网赖以发挥价值的核心支撑，然而当前正面临三类应用挑战：1)工业互联网异构复杂要素统一表征难；2)工业互联网多源数据特征协同学习难；3)工业互联网低质多维数据高效融合难。亟需突破三个瓶颈科学问题：1)工业异构要素统一表征问题；2)工业多源特征协同学习问题；3)工业低质多维高效融合问题。从三个科学问题出发，开展工业互联网异构复杂要素统一表征方法、工业生产多源数据特征协同学习方法、工业要素低质多维数据高效融合方法研究；在理论与方法研究基础上，针对工业互联网典型应用开展应用验证。通过本项目研究，对工业互联网这一新兴方向基础理论的完善与发展具有重要科学意义与应用价值。

Industrial Internet has become a national innovation and development strategy recently. The deep integration of multi-source heterogeneous data representation and deep knowledge reasoning methods of complex elements of Industrial Internet are playing an essential role. However, currently it is facing three challenges. It is difficult to uniformly represent heterogeneous complex elements of Industrial Internet, realize collaborative learning of multi-source data features of Industrial Internet, and efficiently integrate low-quality multidimensional data. It is urgent to break through three bottleneck scientific problems: 1) the unified representation of industrial heterogeneous elements; 2) collaborative learning of industrial multi-source features; 3) efficient integration of industrial low-quality multidimensional data. To address above three issues, we plan to carry out research on the unified representation method of heterogeneous and complex elements of Industrial Internet, the collaborative learning method of multi-source data characteristics of production, and the efficient fusion method of low-quality multi-dimensional data of industrial elements. On the basis of theoretical research, typical Industrial Internet application verification will be carried out. The research of this project has important scientific significance and application value for the improvement and development of the basic theory of the emerging area of Industrial Internet.