Emotion-Cause Pair Extraction in Conversations via Semantic Decoupling and Graph Alignment

2026-04-21

AI Summary

This work addresses the limitations of existing approaches to emotion-cause pair extraction in dialogues, which often overlook the semantic distinction between emotion propagation and causal explanation, thereby struggling to model globally consistent many-to-many causal relationships. To tackle this issue, the paper introduces a novel framework that, for the first time, reformulates the task from a semantic disentanglement perspective by mapping emotions and causes into complementary representation spaces. The proposed method integrates unified dialogue structure modeling, a graph alignment mechanism, and optimal transport theory to establish a globally coherent many-to-many matching paradigm. Extensive experiments demonstrate that this approach significantly improves both accuracy and consistency in emotion-cause pair extraction across multiple benchmark datasets, achieving state-of-the-art performance.

Technology Category

• Knowledge Representation and Reasoning: Action, Change, and Causality
• Natural Language Processing: Discourse, Pragmatics & Argument Mining
• Machine Learning: Causal Learning

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

Emotion-Cause Pair Extraction in Conversations (ECPEC) aims to identify the set of causal relations between emotion utterances and their triggering causes within a dialogue. Most existing approaches formulate ECPEC as an independent pairwise classification task, overlooking the distinct semantics of emotion diffusion and cause explanation, and failing to capture globally consistent many-to-many conversational causality. To address these limitations, we revisit ECPEC from a semantic perspective and seek to disentangle emotion-oriented semantics from cause-oriented semantics, mapping them into two complementary representation spaces to better capture their distinct conversational roles. Building on this semantic decoupling, we naturally formulate ECPEC as a global alignment problem between the emotion-side and cause-side representations, and employ optimal transport to enable many-to-many and globally consistent emotion-cause matching. Based on this perspective, we propose a unified framework SCALE that instantiates the above semantic decoupling and alignment principle within a shared conversational structure. Extensive experiments on several benchmark datasets demonstrate that SCALE consistently achieves state-of-the-art performance. Our codes are released at https://github.com/CoCoSphere/SCALE.

Problem

• Emotion-Cause Pair Extraction
• Conversational Causality
• Semantic Decoupling
• Many-to-Many Alignment
• Dialogue Understanding

Innovation

• Semantic Decoupling
• Graph Alignment
• Optimal Transport
• Emotion-Cause Pair Extraction
• Conversational Causality