A preliminary study on few-shot knowledge reasoning mechanism based on three-way partial order structure

Few-shot knowledge reasoning, as a novel theory that attempts to combine few-shot learning with knowledge reasoning, is a challenging subject. This article intends to explore few-shot knowledge reasoning mechanism based on three-way partial order structure. First, from the perspective of granular computing, some knowledge reasoning operators are defined: granule splitting, granule combination, and granule isomorphism; then, based on the knowledge reasoning operators, the data augmentation operation and knowledge reasoning algorithm are presented; finally, some specific experiments on five datasets are conducted to further illustrate the few-shot knowledge reasoning algorithm. Experiment results show that the method proposed in this article can effectively realize knowledge reasoning based on small sample size data. It can be concluded that the proposed few-shot knowledge reasoning method is effective at some extent, and is expected to bring inspiration to the development of other concept-based knowledge reasoning models.