The application belongs to the field of
artificial intelligence and
natural language processing, and relates to an efficient event extraction method based on controllable
pattern generation, which comprises the following steps: obtaining a text sequence to be tested, performing entity extraction on the text sequence to be tested to obtain an event entity set, a trigger word set and a position vector sequence; constructing multiple events according to the event entity set and the trigger word set; the event contains the relationship between the entity and the trigger word; inputting the event and the position vector sequence into a trained event
pattern generation model to obtain an event pattern; merging and filtering the event pattern to obtain an
event structure; the application learns the definition process of the event pattern based on the candidate mentioned entity by using the generation model, and then automatically generates the event pattern, thereby avoiding modeling of all pre-defined event types and improving the generation efficiency; and the application designs a decoding
constraint algorithm to finely control the
pattern generation process, prevents obtaining non-standard results and ensures that the number of generated roles corresponds to the number of entities.