Enriching Large-Scale Eventuality Knowledge Graph with Entailment Relations

Computational and cognitive studies suggest that the abstraction of\neventualities (activities, states, and events) is crucial for humans to\nunderstand daily eventualities. In this paper, we propose a scalable approach\nto model the entailment relations between eventualities ("eat an apple''\nentails ''eat fruit''). As a result, we construct a large-scale eventuality\nentailment graph (EEG), which has 10 million eventuality nodes and 103 million\nentailment edges. Detailed experiments and analysis demonstrate the\neffectiveness of the proposed approach and quality of the resulting knowledge\ngraph. Our datasets and code are available at\nhttps://github.com/HKUST-KnowComp/ASER-EEG.\n