A General Paradigm of Knowledge-driven and Data-driven Fusion
Fei Hu, Wei Zhong, Long Ye, Dan-Ting Duan, Qin Zhang · 2023
Knowledge and data fusion is a vital research hotspot in current artificial intelligence. The fusion of data-driven and knowledge-driven would be able to organically combine implicit intuition of machines and common-sense judgment of people. Yet, fusion models have mainly been applied to specific problems. Because of the complex mechanisms underlying knowledge formation, the creation of a novel fusion model requires tedious and time-consuming experiments. Here, we approach the problem of general data-driven and knowledge-driven fusion paradigm. We first propose a general framework of the fusion mode. Next, we introduce and show the specific types of the three basic models under this paradigm. We hope that this paradigm will play a role in the development of machine learning.