Open-Domain Hierarchical Event Schema Induction by Incremental Prompting and Verification

Event schemas are a form of world knowledge about the typical progression of events. Recent methods for event schema induction use information extraction systems to construct a large number of event graph instances from documents, and then learn to generalize the schema from such instances. In contrast, we propose to treat event schemas as a form of commonsense knowledge that can be derived from large language models (LLMs). This new paradigm greatly simplifies the schema induction process and allows us to handle both hierarchical relations and temporal relations between events in a straightforward way. Since event schemas have complex graph structures, we design an incremental prompting and verification method INCSCHEMA to break down the construction of a complex event graph into three stages: event skeleton construction, event expansion, and event-event relation verification. Compared to directly using LLMs to generate a linearized graph, INCSCHEMA can generate large and complex schemas with 7.2% F1 improvement in temporal relations and 31.0% F1 improvement in hierarchical relations. In addition, compared to the previous state-of-the-art closed-domain schema induction model, human assessors were able to cover ∼10% more events when translating the schemas into coherent stories and rated our schemas 1.3 points higher (on a 5-point scale) in terms of readability. 1 Relation Allen's base relations e1 starts before e2? e1 ends before e2? e1 duration longer than e2? e1 ≺ e2 e1 precedes e2, e1 meets e2 Yes Yes -e1 ≻ e2 e1 is preceded by e2, e1 is met by e2 No No -e1 ⊂ e2 e1 starts e2, e1 during e2, e1 finishes e2 No Yes No e1 ⊃ e2 e1 is started by e2, e1 contains e2, e1 is finished by e2 Yes No Yes e1 ∥ e2 e1 overlaps with e2, e1 is equal to e2 Yes No No e1 ∥ e2 e1 is overlapped by e2 No Yes Yes