Semantic Decision Internal-Attention Graph Convolutional Network for End-to-End Emotion-Cause Pair Extraction

Emotion-cause pair extraction is an emergent natural language processing task; the target to extract all pairs of emotion clauses and corresponding cause from unannotated text. Previous studies have employed two-step approaches. However, this research may lead error propagation across stages. In addition, previous did not correctly handle situation where are same clauses. To overcome these issues, authors first use a multitask learning model that based on graph perspective sorting, which can simultaneously clauses, emotion-cause via end-to-end strategy. Then propose convert text into structured data, process scenario through unique convolutional neural network. Finally, design semantic decision mechanism address in there multiple