Related Experiment Video
Updated: Sep 20, 2025
06:37
Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention
Published on: December 15, 2023
4.2K
Hierarchical Attention Neural Network for Event Types to Improve Event Detection
Yanliang Jin1, Jinjin Ye1, Liquan Shen1
1Key Laboratory of Specialty Fiber Optics and Optical Access Networks, Joint International Research Laboratory of Specialty Fiber Optics and Advanced Communication, Shanghai University, Shanghai 201620, China.
Sensors (Basel, Switzerland)
|June 10, 2022
Summary
This study introduces a Hierarchical Attention Neural Network for Event Types (HANN-ET) to address data sparsity in event detection. The model effectively uses event type correlations, achieving competitive F1 scores on benchmark datasets.
Area of Science:
- Natural Language Processing
- Machine Learning
Background:
- Event detection is crucial in NLP for identifying and classifying trigger words.
- Small-scale datasets present challenges like data sparsity and imbalanced event instances.
Purpose of the Study:
- To propose a novel Hierarchical Attention Neural Network for Event Types (HANN-ET).
- To leverage event type correlations to mitigate data sparsity and imbalance issues in event detection.
Main Methods:
- Utilized Long Short-Term Memory (LSTM) as a semantic encoder.
- Employed dynamic multi-pooling and Graph Attention Network (GAT) for feature enrichment.
- Developed upper-level event type modules using Neural Module Networks (NMNs) with weighted attention aggregation.
Main Results:
- Achieved an F1 score of 78.9% on the ACE2005 dataset.
- Obtained an F1 score of 68.8% on the MAVEN dataset.
- Demonstrated superior performance compared to previous state-of-the-art methods.
Conclusions:
- The HANN-ET model effectively addresses data sparsity and imbalance in event detection.
- The proposed architecture successfully integrates event type correlations for improved performance.
- The approach offers a promising solution for enhancing event detection accuracy in NLP.
