Method and electronic device for extracting event arguments from text

Abstract

The present disclosure relates to a method and an electronic device for extracting event arguments in a text. The method includes: detecting the event type contained in the text to be processed to obtain a target event type; according to the target event type and the corresponding target theory The a priori correlation information of the meta role is constructed to obtain the argument extraction problem; the text to be processed and the argument extraction problem are spliced ​​to obtain the target text; the representation vector of the target text is input into the machine reading comprehension model, which includes two The classifiers are set in sequence. The first layer of classifiers synchronizes whether there is an answer to the target text, identifies and predicts the start position and end position of the answer, and obtains the position prediction result with the answer indicating label; the second layer classifier Predict whether the predicted entity after the pairing combination of the start position and the end position is an event argument, and obtain the answer prediction result; output the event argument according to the answer prediction result and the corresponding label.

Description

technical field [0001] The present disclosure relates to the field of text data processing and artificial intelligence technology, and in particular, to a method and electronic device for extracting event arguments in text. Background technique [0002] In the face of massive information, how to mine useful information from massive text data becomes an important issue. A lot of information generally exists in the form of events, and event extraction is to extract specific types of event trigger words and corresponding event arguments from a large number of unstructured texts and present them in a structured form. This is the field of information extraction. an important task. [0003] In the process of realizing the concept of the present disclosure, it is found that there are at least the following technical problems in the related art: in the event argument extraction, there is a scene where one argument role corresponds to multiple argument entities and the arguments are...

AI Technical Summary

Problems solved by technology

[0003] In the process of realizing the concept of the present disclosure, it is found that there are at least the following technical problems in related technologies: when extracting event arguments, there is a scene where one argument role corresponds to multiple argument entities and arguments are nested entities. The current method cannot Accurate extraction of arguments for the above two scenarios; cascaded extraction first extracts candidate arguments, and then classifies candidate arguments, there is a problem of error transmission

Images