Emotion-cause pair extraction method based on multi-granularity information and multi-module interaction

TL;DR

This work tackles the emotion-cause pair extraction (ECPE) problem by proposing an end-to-end multitask framework (MM-ECPE) that explicitly models interactions among emotion extraction, cause extraction, and emotion-cause pair prediction. A multi-level shared module enables cross-task information flow, while a knowledge-graph-based filtering strategy selects balanced samples to mitigate label imbalance. A transformer-based ECPE predictor and a position-aware encoding module (PAIM), including a BERT-enhanced variant (MM-ECPE(BERT)) with lexicon augmentation, further improve performance on position-imbalanced samples. Empirical results on the ECPE benchmark show state-of-the-art F1 scores across tasks, with notable gains for imbalanced samples, confirming the effectiveness of joint learning, KG filtering, and PAIM. The approach offers a practical path toward robust ECPE in real-world text by leveraging mutual task benefits and external knowledge.

Abstract

The purpose of emotion-cause pair extraction is to extract the pair of emotion clauses and cause clauses. On the one hand, the existing methods do not take fully into account the relationship between the emotion extraction of two auxiliary tasks. On the other hand, the existing two-stage model has the problem of error propagation. In addition, existing models do not adequately address the emotion and cause-induced locational imbalance of samples. To solve these problems, an end-to-end multitasking model (MM-ECPE) based on shared interaction between GRU, knowledge graph and transformer modules is proposed. Furthermore, based on MM-ECPE, in order to use the encoder layer to better solve the problem of imbalanced distribution of clause distances between clauses and emotion clauses, we propose a novel encoding based on BERT, sentiment lexicon, and position-aware interaction module layer of emotion motif pair retrieval model (MM-ECPE(BERT)). The model first fully models the interaction between different tasks through the multi-level sharing module, and mines the shared information between emotion-cause pair extraction and the emotion extraction and cause extraction. Second, to solve the imbalanced distribution of emotion clauses and cause clauses problem, suitable labels are screened out according to the knowledge graph path length and task-specific features are constructed so that the model can focus on extracting pairs with corresponding emotion-cause relationships. Experimental results on the ECPE benchmark dataset show that the proposed model achieves good performance, especially on position-imbalanced samples.

Figures (7)

• Figure 1: Extraction examples of emotion-cause pairs.
• Figure 2: Relative position of emotion-causes.
• Figure 3: MM-ECPE model.
• Figure 4: Examples of knowledge graph filtering.
• Figure 5: Emotion cause pair prediction.