Data-driven Connectionist AI and knowledge-driven Symbolic AI are expected to solve four important challenges of today's image understanding algorithms: long-tailed distribution of categories, unseen scene generalization, entity/relationship description, and complex cognitive tasks. It should be noted that this "Symbolism + Connectionism" concept was very highly evaluated by Zhang Cymbal, an academician of the Chinese Academy of Sciences, at the 2019 Beijing Zhiyuan Conference, as representing the mainstream trend of future third-generation AI algorithms.

Data-driven is a data-driven organization that acquires, processes, and uses data in a timely manner to create benefits, iterates and develops new products, and explores within the data. There are many ways to assess whether an organization is data-driven, such as: the amount of data generated, the extent to which data is used, and the process of internalizing data.