Emotion-Cause Pair Extraction in Conversations Based on Multi-Turn MRC with Position-Aware GCN

Emotion Cause Pair Extraction in Conversations (ECPEC) is a challenging new task in the field of sentiment analysis. Its objective is to extract emotion utterances and their corresponding cause utterances from a conversation. Most recent studies have adopted end-to-end approaches to handle this task. However, it is difficult for these approaches to fully address the issue of label sparsity in the data. Thus, we introduce Machine Reading Comprehension (MRC) framework with Position-aware Graph Convolutional Network (GCN) and leverage dialogue characteristics to model conversations. Furthermore, we also explore the impact of data input methods on the results. Experiments show that our approach is competitive with existing methods. To the best of our knowledge, this is the first attempt to use multi-turn MRC for ECPEC and it brings new insights for this task.