Event Extraction Using Trigger Word Vectors for Overlapping Text
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing event extraction technologies face challenges in effectively extracting overlapping events from text due to difficulties in identifying trigger words and resolving element relationships, resulting in unsatisfactory extraction results.
Innovation Solution
The method involves identifying trigger words in a target text, obtaining trigger word vectors, determining element word information associated with event types, and generating event extraction results based on location information and element relationships, using a deep learning-based neural network model to improve event extraction accuracy for overlapping events.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional event extraction methods are used, then simple events can be extracted, but overlapping events cannot be effectively extracted
Solution Approach 1:
The patent segments the event extraction process into distinct components: trigger word identification, element word identification, and relationship determination. By processing each trigger word independently and identifying its associated element words separately, the system can handle multiple overlapping events without confusion, as each event is processed through the same segmentation framework
Solution Approach 2:
The patent introduces a new dimension by incorporating relative location vectors that capture the positional relationships between trigger words and element words. This dimensional enhancement allows the model to distinguish between overlapping events by considering not only semantic information but also spatial positioning, thereby resolving ambiguities in overlapping event scenarios
2Productivity
If trigger word identification is simplified, then processing speed increases, but identification accuracy decreases
Solution Approach 1:
The patent transforms trigger word identification from a discrete classification task into a continuous vector representation task. By using pre-trained word vectors and transforming them through neural network layers to produce trigger word vectors, the system maintains high identification accuracy while enabling efficient batch processing, thus resolving the trade-off between speed and precision
3Manufacturing precision
If element relationship determination is made complex, then extraction precision improves, but computational complexity increases
Solution Approach 1:
The patent performs preliminary action by pre-computing and storing word vectors for all words in the corpus before actual event extraction. These pre-computed vectors serve as the foundation for rapid trigger and element word identification, reducing the computational burden during the actual extraction process while maintaining high precision in determining element relationships
Data Source
AI summary
The present application discloses an event extraction method and apparatus, a computer program product, a storage medium, and a device. The method includes: identifying at least one trigger word in a target text, and obtaining a trigger word vector corresponding to each of the at least one trigger word; determining, in the target text, element word information associated with an event type corresponding to each trigger word based on the trigger word vector corresponding to each trigger word, an event type vector corresponding to each trigger word, and a relative location vector corresponding to each trigger word, where the element word information includes location information corresponding to each of at least one element word and an element relationship between the element words; and generating an event extraction result corresponding to the target text based on the location information of each element word and the element relationship between the element words, where the event type vector corresponding to each trigger word indicates the event type corresponding to the trigger word, and the relative location vector corresponding to each trigger word indicates a relative location relationship between a word and the trigger word in the target text.