TLAG: An Informative Trigger and Label-Aware Knowledge Guided Model for Dialogue-based Relation Extraction
This work addresses a specific bottleneck in dialogue-based relation extraction for natural language processing applications, representing an incremental improvement over existing trigger-enhanced methods.
The paper tackled the problem of noise and underutilization of trigger information in dialogue-based relation extraction by proposing TLAG, which adaptively fuses triggers and incorporates label-aware knowledge, achieving improved performance over baseline models on the DialogRE dataset.
Dialogue-based Relation Extraction (DRE) aims to predict the relation type of argument pairs that are mentioned in dialogue. The latest trigger-enhanced methods propose trigger prediction tasks to promote DRE. However, these methods are not able to fully leverage the trigger information and even bring noise to relation extraction. To solve these problems, we propose TLAG, which fully leverages the trigger and label-aware knowledge to guide the relation extraction. First, we design an adaptive trigger fusion module to fully leverage the trigger information. Then, we introduce label-aware knowledge to further promote our model's performance. Experimental results on the DialogRE dataset show that our TLAG outperforms the baseline models, and detailed analyses demonstrate the effectiveness of our approach.